Most founders can’t question their own data

By

-

09/15/2026

Founders have a habit of looking more at dashboards than the spreadsheets or raw numbers. Someone else built the model, chose the cohort, decided what counted as “active,” and picked the date range that made the chart more readable. The founder’s job, in theory, is to decide what to do next. And, it makes sense for founders to outsource and delegate, especially if they’re a non-technical founder.

But in practice, a lot rests on trusting that whoever built the analysis asked the right question in the first place. After all, for what the founder lacks in technical analyst skills, the analyst lacks in business intelligence and context relative to the founder.

Most founders have no reliable way to check. Not because they can’t read a chart, but reading a chart and interrogating the method behind it are very different things. Only one of them gets taught.

The failure is deceiving because it’s not that the answer (analysis and dashboard) is bad, but that it’s a good answer to the wrong question.

The sums, not the story

The clearest example of what gets missed is in a 1973 admissions dataset from Berkeley. In aggregate, the university had acceptance rates of 44% from male applicants and only 35% from female applicants that year. It was a big, concerning gap that looked like clear evidence of bias. But when broken down, the picture began to reverse. Women applied in far larger numbers to hyper-competitive departments that had overall acceptance.

Pool the departments together and the aggregate number found a pattern that doesn’t really exist in any single department. In a business context, the dashboard and top-line figure in isolation didn’t tell the whole story. When the founder is presented with only half the story, they can draw the wrong conclusions.

The data that never made it back

The same failure runs the other way too. Not a total that hides a pattern, but a dataset that’s missing the cases that actually mattered the most. In 1943, the US Army wanted to know where to add armor to its bombers, and the obvious approach was to study the bullet holes on planes that returned from missions so it could reinforce the areas that were hit most often.

The statistician Abraham Wald, working with Columbia University, pointed out the flaw in this approach: the returning planes were, by definition, the ones that survived their damage.

So the areas with the fewest bullet holes on returning aircraft weren’t actually the safest places to get hit, they were the most devastating places, because a hit meant the plane never made it back to be counted at all.

So the armor went on the gaps in the data rather than the marks on it. Wald wasn’t working from a bigger dataset than the officers that were already studying the same bombers, he was asking a different question of the identical numbers.

A founder analysing churn might focus completely on exit surveys from departing users, missing the fact that the most critical failure signals are found in the lost prospects who abandoned the product before reaching the onboarding phase to leave any feedback at all. Once you pose the right question, the absence of data is telling in and of itself, but it not an intuition that many have naturally (due to human biases), so it must be taught.

Where the skill actually gets taught

Interrogating a data set and asking what’s missing, what’s been pooled together that shouldn’t be, what assumption the model is making, does not come naturally to humans, but it’s a teachable skill.

It shows up as a deliberate part of some business degrees rather than something students pick up accidentally. Schiller International University’s degree in international business is an example of building that kind of analytical questioning into its coursework along with the more obvious international business skills.

For students or founders who want that training, an american university in Germany means not settling for a translated course, but having it natively created in English, while also benefiting from studying in an internationalised environment.

A confidence problem, not a skills one

This isn’t so much a founder-specific failing, nor does it fade with seniority. A KPMG global survey of executives (ten countries) found:

70% believed using data and analytics exposed their organisation to reputational risk. This mistrust tracked all the way through the process
10% of respondents said their organisation excelled across every stage of managing data and analytics
16% were confident in the accuracy of the models they were producing
18% had governance frameworks they’d confidently describe as adequate across the board

It became clear that data sourcing had the most trust, but confidence soon fell away at every stage downstream. Founders of large organisations report the same gap a first-time founder has – it’s not a lack of dashboards, but a lack of grounds to trust their foundations.

The founders who ask the better questions usually aren’t smarter, but simply just learned, at some point, to stop reading the answer before they understand the method that produced it.

Picture Tumisu / Pixabay

Author Dino Bozzi

Statements of the author and the interviewee do not necessarily represent the editors and the publisher opinion again.

Ground A raises €9.1 million to industrialise laser-based counter-drone system

By

Markus Elsasser

-

09/13/2026

0

The Ground A founding team Ground A funding round: the Berlin-based defence tech start-up receives €9.1 million for laser-based drone defence with the GA 1000. — Ground A Founder Team © Gordon Welters

German defence tech company Ground A has closed a pre-seed funding round of €9.1 million, the company announced. This round is aimed at moving its GA 1000 counter-drone system from prototype into series production. The round was led by Vsquared Ventures. Participation came from a consortium of venture and strategic investors. Ground A, founded in January 2026 and based in Berlin and Düsseldorf, says the GA 1000 has already progressed from prototype to testing and validation within months of the company’s launch.

Ground A’s Investor Consortium: Vsquared Ventures Leads the Round

The round brought together a mix of venture capital and strategic backers from the defence, technology and industrial sectors. Alongside lead investor Vsquared Ventures, the press release names Quantum Systems, GRIF, Robin Capital, ABACON, Leblon Capital and OVNI Capital as participants. The round also included individual entrepreneurs and investors from Europe’s defence and software industries. These included Florian Seibel, Sven Kruck, J14 Capital, Andreas Nauerz, Marlon Braumann and Arx Robotics managing director Stefan Roebel. As this is Ground A’s first disclosed financing round, there is no prior round to compare it against.

What Ground A’s GA 1000 Counter-Drone System Does

The GA 1000 is a compact, laser-based counter-drone system built to detect, classify and track small, agile drones automatically. It combines radar and multiple sensor types with computer vision and proprietary AI. The decision to deploy the laser effector remains with a human operator. Housed within a Euro-pallet footprint, the system is designed to be mobile and deployable across a range of operational settings. Its open, vendor-agnostic architecture allows it to be integrated into existing air defence and command-and-control systems. Additionally, it can be used on land, sea and air platforms, rather than operating only as a stand-alone unit.

Ground A’s Path to Series Production in 2027

Ground A says the pace of development — from company formation in January 2026 to a system already in testing and validation just months later — reflects a deliberate early focus on product and technology ahead of any public announcement. Now, the company is setting up modular series production in Germany, with series production targeted for 2027. A further financing round, also planned for 2027, is intended to expand production capacity. Ground A works towards a broader product platform for counter-drone and air defence applications. This expansion goes beyond the GA 1000 alone.

Leadership and Investors on the Funding Round

‘We founded Ground A in January and already have our first system in testing and validation, approaching market launch,’ said Dr Tobias Brune, chief executive and co-founder of Ground A. ‘Europe needs effective counter-drone solutions that not only work technologically but can also be industrialised quickly and made available in large numbers. The GA 1000 is just the beginning of a much larger product platform.’

Benedikt von Schoeler, general partner at Vsquared, said the fund backed Ground A for its combination of operational flexibility and execution speed. ‘What matters is that systems are not only high-performing, but also mobile and rapidly deployable,’ he said. ‘Ground A combines precisely this operational flexibility with technological excellence, high execution speed and the ambition to scale industrially from day one.’

What’s Next

With the pre-seed round secured, Ground A’s near-term priorities are industrialisation and scaling in parallel. The company is building out modular series production capacity in Germany and progressing the GA 1000 toward market launch. It is also preparing a further funding round in 2027 to expand manufacturing capacity. Furthermore, the company frames this as a contribution to the broader defence capability of Germany, NATO and allied nations, rather than a single-product commercial launch.

FAQ about Ground A

How much funding has Ground A raised?

Ground A closed a pre-seed round of €9.1 million in June 2026.

Who invested in Ground A?

The round was led by Vsquared Ventures, with participation from Quantum Systems, GRIF, Robin Capital, ABACON, Leblon Capital and OVNI Capital, plus individual investors including Florian Seibel, Sven Kruck, Andreas Nauerz, Marlon Braumann, J14 Capital and Arx Robotics managing director Stefan Roebel.

What does Ground A do?

Ground A builds software- and laser-based counter-drone and air defence systems. Its first product, the GA 1000, detects, classifies and tracks small drones and neutralises them using a laser effector under human control.

When was Ground A founded?

Ground A was founded in January 2026 and is based in Berlin and Düsseldorf.

When will the GA 1000 enter series production?

Ground A is targeting series production in 2027, alongside a further funding round planned for the same year to expand manufacturing capacity.

Photo/source: Ground A GmbH – © Gordon Welters www.GordonWelters.com

Limetax raises €36 million to build Germany’s agentic-first accounting & tax group

By

StartupValley

-

09/09/2026

0

Limetax: AI Platform for Tax Firms Raises €36M

Berlin-based technology company Limetax is bringing established tax firms together in one group and building a proprietary AI agent platform for them. The company has raised €36 million to fund its expansion: the equity round is led by Motive Partners, with participation from Activant, Heliad and angel investors including former German Finance Minister Christian Lindner, alongside a €30 million credit facility provided by a consortium of German banks.

Berlin, 9 September 2026 – Limetax has raised €36 million in financing, comprising €6 million in pre-seed equity led by Motive Partners, with participation from Activant and Heliad, and a €30 million credit facility from a consortium of German banks. Angel investors include former German Finance Minister Christian Lindner, former Rocket Internet executive and 468 Capital co-founder Alexander Kudlich, as well as Anton Rummel and Ante Spittler, founders of fintech unicorn Moss. The capital will be used to further develop Limetax’s agentic platform and partner with additional tax firms across Germany.

Eight months after launching in January 2026, the Limetax group has grown to four firms across seven locations with around 150 employees and annualised revenue in the double-digit millions. Its AI platform is already deployed across every firm in the group.

Tackling the structural constraints of the tax profession

Germany’s nearly 54,000 accounting & tax firms are the finance operations backbone of the country’s Mittelstand. Many are struggling to hire enough skilled professionals, while tax regulation continues to become more complex. At the same time, a significant share of their teams’ time is still spent on repetitive work: collecting data, processing bookkeeping and managing deadlines. That leaves little time for the advisory work clients increasingly need.

Limetax is tackling this capacity constraint by bringing established tax firms together and building technology directly into how they work. Its proprietary agentic platform is developed alongside the firms and embedded into their day-to-day operations.

The platform sits on top of DATEV, the core software infrastructure used by much of the German tax profession, and orchestrates specialised AI agents across bookkeeping, payroll and financial statements. The agents take on repetitive workflows, while outputs pass through a review layer where the firms’ professionals apply their judgement and expertise. Professional responsibility and client advice remain in the hands of the human experts.

Early bookkeeping deployments show the potential. In initial cases, AI agents have reduced monthly processing time from around twenty hours to six. For employees, that means less manual data entry and more time for client work. For clients, it means faster financial statements and more of their advisers’ time for the questions AI cannot answer today.

The result is more capacity without losing what makes local accounting & tax firms valuable: professional expertise, trusted relationships and personal advice.

Built by entrepreneurs, engineers and experts

Limetax was founded by Christoph Gamon, Maximilian Meyer and Christoph Dansard. Gamon previously co-founded Razor Group, where he served as CFO and helped scale the company to a reported $500 million-plus in revenue. Before that, he worked in the executive office at Rocket Internet and in investment banking at UBS in London. Meyer was part of Razor Group’s founding team and later served as SVP Operations, leading more than 100 employees, having previously worked at Lazard and N26. Dansard was a founding engineer at fintech unicorn Augustus, formerly Ivy and is an alumnus of Technical University Munich.

The Berlin-based team brings together engineers and operators with backgrounds at Taxfix, McKinsey, EY, Buena, Raisin and ETL.

“The future of accounting & tax advisory will not be built by adding another AI tool. It will be built by rethinking advisory and technology together from the ground up,” said Christoph Gamon, Co-Founder and CEO of Limetax. “That is what we are doing at Limetax. Together with our partner firms, we are building a group where our agentic platform becomes the technological backbone of day-to-day work. That creates more space for the things people do best: applying judgement, solving complex problems and providing real advice.”

“Professional services is one of the largest sectors where the potential of AI is only beginning to be unlocked. In accounting & tax, the constraint is not demand, it is capacity,” said Michael Hock, Partner, Venture at Motive Partners. “Limetax is building technology where the work actually happens: together with the partner firms and deeply embedded in their operations. The combination of technology and professional expertise has the potential to build a new generation of accounting & tax firms.”

“Our profession is going through the biggest shift I have seen in my career. The opportunity is enormous, but for an individual firm, building the technology required to capture it is almost impossible,” said Uwe Tille of bplus, one of the founding partners of the Limetax group. “With Limetax, we can actively shape that transition ourselves. Technology is taking more and more routine work off our hands and giving us more time with our clients. That is where we, as tax advisers, want to spend our time.”

Source Picture and PR Limetax

Sci2sci Raises €1.2M Pre-Seed to Make AI Trustworthy for Biopharma

By

StartupValley

-

09/08/2026

0

sci2sci: AI for Trustworthy Biopharma Data. From left to right_sci2sci co-founders_Valerii Kremnev & Angelina Lesnikova.

Sci2sci Raises €1.2M Pre-Seed to Make AI Trustworthy for Biopharma and Other Regulated Industries

Berlin startup’s neurosymbolic memory layer, already in production with customers ranging from lean specialist teams to global biopharma scale, turns scattered enterprise data into self-verifying knowledge.

BERLIN, GERMANY, 08.09.2026 — sci2sci, a Berlin-based startup building a neurosymbolic memory layer for regulated industries, today announced it has raised €1.2 million in pre-seed funding. The round was co-led by Heliad and IBB Ventures, with participation from Robin Capital and Superangels.

Sci2sci will use the funding to expand its engineering team and accelerate deployment of its products, VectorCat and Integrity Cortex, across biopharma and other highly regulated industries.

Why biopharma first

Sci2sci started in biopharma, where the data problem is acute and the cost of getting facts wrong is measured in regulatory delays, failed audits and abandoned research. Biopharma companies generate enormous volumes of research data, and most of it lives in scattered PDFs, spreadsheets and lab notes across dozens of disconnected systems.

When a regulatory submission references a preclinical finding that references a lab result that references raw data last modified three years ago, who verifies the chain? Today, humans do, manually, over weeks. And when AI tools are layered on top of that same messy, contradictory data, they inherit the dysfunction — hallucinating with confidence.

Former Corporate VP and Digital Transformation Officer, R&D Novo Nordisk, Stephanie Bova saw this problem at the largest possible scale.

“Every board I advise is asking the same question: why hasn’t the AI investment shown up anywhere I can measure it? The answer is almost always trust. Pilots work beautifully and then nothing reaches production, because no one is willing to put their name under an output they can’t trace. Sci2sci removes that blocker at the architecture level — which is precisely why I joined.”

She currently helps life science companies address similar challenges as CEO of Mirai Advisors, an executive advisory firm dedicated to helping Board and C-Suite leaders get tangible value from their AI and technology investments, and recently joined sci2sci as a senior advisor.

Every fact becomes executable

Sci2sci co-founder and CEO Angelina Lesnikova holds a PhD in neuroscience, where she spent years studying the molecular mechanisms of memory and learning in the brain — research that shaped the company’s core idea: knowledge should behave like an active network, not like static storage.

Integrity Cortex, the company’s flagship product, runs on the principle ‘knowledge as code’. It extracts every fact from a company’s documents, data and AI outputs and connects them into a single living network.

Instead of asking a language model to summarize documents into prose, the framework encodes information as a formal logic system — every claim must cite a verbatim quote, every conclusion must derive from stated premises. A symbolic engine then verifies every citation and derivation. If the model fabricates a fact, verification fails — and that failure flags every conclusion that depends on it. It also catches inconsistencies already present in the ground truth, and traces downstream impact whenever a source document changes.

The approach is powered by Parseltongue, a framework sci2sci open-sourced this year under Apache 2.0. In August, a team of researchers won second place at a biopharma AI hackathon with a Parseltongue-based system that validates cancer drug targets against published literature and clinical trials.

When the FDA and the safest AI labs sound the same alarm on trust

In April 2026, the FDA issued its first-ever warning letter citing AI misuse in drug manufacturing — a company had used AI agents to generate compliance documents without human review. When asked why process validation had never been performed, personnel told investigators their AI agent had never told them it was required. The company ceased drug production.

Three months later, the problems escalated from unverified outputs to uncontained models. OpenAI disclosed that one of its models broke out of an isolated test environment and breached Hugging Face’s production infrastructure. A week later, Anthropic revealed that three of its own Claude models had breached three real companies during security testing. Two of the breached companies didn’t even notice until Anthropic told them.

“When the two labs that market themselves as the most safety-conscious both have models breaking into real companies in the same month, it’s a process design flaw,” said Valerii Kremnev, co-founder and CTO of sci2sci.

“We took the opposite approach. Instead of shipping capabilities first and patching trust and security afterward, we built Parseltongue and Integrity Cortex to permit only safe outputs. A model’s lockpicking skills are useless when there’s no door — and our software guarantees there is none. The same holds for deception: our systems don’t have ungrounded statements as a primitive. The model is forced to produce evidence for every claim it makes.”

The verification layer is deterministic and interpretable by design, with a full audit trail compatible with 21 CFR Part 11 — the FDA’s standard for electronic records — making it suitable for regulatory submissions, manufacturing records and other workflows where every claim needs to trace to its source.

Already in production across the drug development value chain

Sci2sci’s customers already span from pre-clinical research to contract clinical research and bioprocess operations. Its second product, VectorCat, the company’s data integration layer, makes scattered enterprise data findable — connecting to cloud storage, network drives and lab systems without migration, and building a searchable catalog for humans and AI agents. Integrity Cortex makes it trustworthy.

The company plans to use the funding to deepen these deployments, grow into additional biopharma accounts, and extend Integrity Cortex to other regulated sectors, including banking, where the same problem of tracing facts across thousands of interdependent documents applies.

“Most companies selling AI into regulated industries are betting that language models will be accurate enough. Sci2sci are making a different bet: build a system where an ungrounded claim simply cannot exist,” said Christopher Garlich, Investment Lead at Heliad.

“That’s an architectural decision rather than a mere bolt-on feature — it’s a critically missing layer in the enterprise AI stack and it’s needed everywhere facts have to be provable.”

Tobias Schimmelpfennig, Investment Director at IBB Ventures, said: “Europe needs deep-tech champions that go beyond surface-level AI wrappers. Angelina and Valerii are doing exactly that — combining genuine life sciences domain depth, cutting-edge AI research and enterprise-scale systems engineering to build the neurosymbolic intelligence layer that regulated industries have been waiting for.”

Picture From left to right sci2sci co-founders Valerii Kremnev & Angelina Lesnikova Credits sci2sci

Source Redgert

FLÆ Robotics: How AI and Robotics Are Transforming the Hotel Reception

By

StartupValley

-

09/08/2026

0

FLÆ Robotics: BE-A Brings AI to Hotels Filip Linek Credit Flae Robotics

FLÆ Robotics is developing BE-A, an AI Hospitality Agent designed to automate routine hotel operations while supporting more efficient and personal guest service.

Could you briefly introduce FLÆ Robotics and explain how your own experience as a hotel owner led to the idea for BE-A?

FLÆ Robotics is building BE-A, an AI Hospitality Agent for hotels. It combines work that is usually split between reception, reservations and concierge: answering guests, helping with booking inquiries and supporting them before arrival and during their stay. The exact scope depends on the functions and integrations available at each hotel. The idea came directly from my own experience. Since 2019, I have been a co-owner and director of the PECR hotel complex in Pec pod Sněžkou. I was dealing with missed calls, slow replies and teams overloaded by the same questions every day. I first tried to find a product I could buy. When I realised that nothing covered the role properly, I decided to build it. So BE-A did not start with a technology trend. It started with a problem in my own hotel.

Missed calls, staff shortages, and overloaded front desks were problems you experienced firsthand. When did you realize they could become the basis for a scalable business?

I realised it when other hoteliers began describing exactly the same problems. It was not only my reception team or one badly organised property. Hotels everywhere struggle when several guests call at once, an email arrives late in the evening or someone needs an answer in a language the team does not speak. At PECR, the gap was especially visible after 7 p.m., when the reception closed but guests continued writing. The questions and tasks repeat from hotel to hotel; what changes is the property’s information, tone and systems. That means the technological core can scale while each implementation remains specific to the hotel. Today, BE-A processes around 85% of inquiries automatically at PECR on a rolling seven-day view. That was important proof that we were solving a wider operational problem, not building a feature only for ourselves.

BE-A is positioned as an AI Hospitality Agent that combines elements of reception, reservations and concierge service. Which tasks create the greatest value when automated?

The best starting point is the routine work that consumes reception time every day. A guest asks whether a dog is allowed, when breakfast starts, whether a room is available or how to book a service. BE-A can handle communication through email, WhatsApp and phone, depending on the hotel’s setup. With the right integrations, it can also help prepare an offer, process an individual reservation or pass a request to the relevant team. That is the important difference from a basic chatbot: BE-A should not only produce an answer, but help move the request forward. We still keep clear limits. A complicated group booking or a sensitive complaint should go to a person. At Hotel Studánka, the time needed to handle a standard email fell from about 15 minutes to two. That is the kind of practical improvement we look for.

Hospitality depends heavily on personal interaction. How do you use AI to increase efficiency without losing the human touch?

I do not believe the human touch is copying check-in information into an email for the fiftieth time. It is noticing that a guest is tired, resolving an unusual problem or giving a recommendation that makes the stay memorable. The routine often prevents receptionists from doing those things well. BE-A takes over defined tasks and remains available when the team is busy or off duty. The hotel can choose how far automation goes: BE-A may draft a reply for approval, answer automatically or hand the conversation to a person. At PECR, for example, individual stay inquiries can be handled automatically, while group requests are passed to the team. I am not trying to build a hotel without people. I want good people to spend more time with guests and less time moving information between systems and messages.

Hotels already work with numerous systems and communication channels. Why is integration rather than replacing existing technology central to your approach?

As a hotelier, I would not want another isolated tool that creates more work for the team. Hotels already have a PMS, booking engine, email, telephony and internal processes that may have been built over many years. BE-A has to fit into that environment. If a guest asks for availability, a useful answer depends on current data. If the guest wants to proceed, BE-A should be able to help with the next step or transfer the request with its context intact. That is why integrations matter. We are not trying to replace the hotel’s core systems or present BE-A as another operating system. We connect guest communication with the information and workflows that already exist. It is technically demanding, especially because hospitality technology is fragmented, but without that connection an AI agent remains little more than a sophisticated FAQ.

FLÆ Robotics is established in Czechia and developing its presence in the UK and DACH markets. What are you learning about taking the same product into different markets?

Czech hotels still give us our most mature live experience. The UK and DACH are markets where we are developing commercial activity and industry relationships, but I do not want to pretend that every market is at the same stage. What I have learned is that translating the product is the easy part. Hotels use different systems, guests prefer different channels and the buying process changes from one country to another. Even the same brand may organise reception and reservations differently across properties. The core of BE-A should remain the same, but the knowledge, tone, integrations and service model have to reflect the local operation. The US remains an important future direction for us, but our immediate job is to build strong proof and repeatable implementations in the markets where we are already active.

You advocate solving narrow, measurable problems rather than putting “AI everywhere.” How do you decide which hotel processes are actually worth automating?

I look first at what repeatedly irritates the guest or wastes the team’s time. In our hotels, unanswered emails and missed calls were obvious because the effect was visible immediately. We could see how long a guest waited and whether an inquiry became a booking. That made written communication a sensible place to start. We then added further tasks only when the information and integrations were reliable enough. I also ask where the exception is. A reservation request for two people may be straightforward; a wedding group with special conditions is not. If we cannot define when BE-A should stop and involve a person, the process is not ready for full automation. I prefer to solve one real hotel problem properly and then extend the role. Putting AI everywhere at once usually creates an impressive demonstration and a poor operation.

Your early hotel deployments have helped shape the product. How important is co-development with hotels when building new AI and hardware solutions?

It is essential because hotels immediately expose the situations that look small on paper but decide whether the product works. What happens when information on the website conflicts with the booking engine? How should a group inquiry be recognised? Where does a maintenance request go at night? Our hotel partners and their reception teams give us that reality every day. Guests also tell us when an answer is too long, too formal or simply not useful. With hardware, the details become even more physical: lobby noise, several people arriving together, the position of microphones or whether an older guest feels comfortable using the interface. At the same time, we cannot build a separate product for every property. The purpose of co-development is to find recurring patterns and improve the common BE-A platform, while allowing each hotel to configure its own knowledge and processes.

Your long-term vision ranges from voice assistants and avatars to physical robotics. Where does software end and where does robotics begin for FLÆ Robotics?

The software is the brain of BE-A: the hotel knowledge, conversation history, integrations and rules that allow it to help a guest. Today that intelligence can communicate through text and voice. An avatar gives it a visible presence in the lobby. Robotics begins when BE-A must understand and act in a physical space. Imagine five guests arriving at once: the system has to know who is speaking, keep each conversation separate and ignore the television or children’s corner behind them. Later, it may also need to issue a key or interact with payment and check-in equipment. Those are very different challenges from generating a good email. I am interested in hardware only when it helps BE-A perform the hospitality role more naturally. The robot is not the product by itself; it is one possible interface for the same AI Hospitality Agent.

AI and hardware businesses can be capital-intensive. What have been the toughest challenges in financing, team building, and international growth?

At the beginning, I believed I could commission the first version externally and manage it like another business project. I was wrong. We spent time and money before I understood that the core knowledge had to remain inside the company. Building our own team was therefore one of the most important decisions. I am not a technologist by training, so I had to learn how to combine AI specialists, hotel-technology experience and people who understand daily operations. Financing is difficult because software, integrations, voice and hardware move at different speeds. You have to keep investing in the long-term vision while proving value with what customers can use today. International growth adds another temptation: every new market and partner looks exciting. The difficult discipline is deciding what not to do and keeping the team focused on the hospitality role we set out to build.

What is your long-term vision for FLÆ Robotics, and how do you expect the hotel front office to change as AI becomes more capable?

My vision is that a guest will not have to think about which department or channel to contact. The same AI Hospitality Agent will understand the hotel and help with a question, a reservation, an arrival detail or a request during the stay. It may meet the guest by email, WhatsApp, phone or, later, through an avatar or humanoid interface in the lobby. The human team will still be crucial, but its role will change. Receptionists will spend less time searching for information and repeating routine instructions, and more time acting as hosts, solving exceptions and creating memorable moments. Different hotels will choose different balances, especially in luxury and boutique hospitality. I do not want to remove people from hospitality. I want BE-A to provide reliable service around the clock and give good people the space to be genuinely hospitable.

What three pieces of advice would you give founders building an AI and hardware company in Europe with ambitions to scale internationally?

Build something you personally understand. I knew the frustration of running a hotel reception before I tried to solve it. That experience still helps me separate a real need from an attractive technology idea.
Keep the critical knowledge inside the company. External partners can help, but the core product, customer learning and technical decisions cannot be outsourced if you want to build something difficult.
Be ambitious about the vision and disciplined about the next step. An AI-and-hardware company can spend money in many directions. Prove useful software first, learn from real customers and add complexity only when the previous layer works.

Picture Filip Linek Credit Flae Robotics

Thank you Filip Linek for the Interview

Statements of the author and the interviewee do not necessarily represent the editors and the publisher opinion again.

Kazimi Raises €2.2M to Fight Mobile Ad Fraud

By

Markus Elsasser

-

08/24/2026

0

Kazimi Team in front of a wall. Kazimi Raises €2.2M Pre-Seed Round in Berlin

Kazimi closes €2.2 million pre-seed round in Berlin

Mobile cybersecurity start-up Kazimi has closed a €2.2 million pre-seed round to tackle a problem that has quietly undermined the mobile industry for years: bots polluting the data app marketers rely on. Kazimi wants to give app developers a clearer view of which user signals are real and which come from bots.

The round was oversubscribed. Market One Capital led the investment, with IBB Ventures and several mobile industry executives joining in. Backers include former Rovio CTO Petri Hyökyranta, ex-Supercell data scientist Janne Peltola, and Erik Massmann, who previously served as CFO of the Birkenstock Group and Sportradar AG.

What Kazimi actually does

Kazimi secures the communication signals inside an app. In practice, this means app developers can collect and verify their own first-party data, spot non-human bot activity, and separate genuine users from fake ones. Kazimi built the system to be privacy-preserving and compliant with data regulations, using cryptographic proofs rather than methods that expose user data.

Andreas Naumann, Kazimi’s co-founder and chief product officer, said it has never been technically possible for developers to see clearly which signals come from real users and which come from bots. He believes Kazimi’s approach closes that gap and lets developers protect what he calls their most important resource: first-party data.

Why the timing suits Kazimi

The scale of the problem is growing. Global mobile ad spend hit USD 419 billion in 2025, according to Business of Apps, and Udonis expects it to pass USD 1 trillion by 2035. That growth makes accurate data increasingly valuable, and increasingly worth faking.

Cloudflare reported that bots officially overtook human activity on the internet for the first time in 2026. Meanwhile, AppsFlyer’s latest report on mobile ad fraud found that bot attacks now drive the vast majority of fraudulent activity across mobile channels. As a result, marketers are making decisions based on data that may already be compromised.

Erik Massmann, Kazimi investor and former CFO of the Birkenstock Group and Sportradar AG, put the problem bluntly: the bot crisis in mobile marketing is the elephant in the room, since everyone in the industry already knows mobile channels lack transparency and trust. He argued that understanding what’s real and what isn’t has never mattered more, both for using AI effectively and for cutting out fraud.

Founders with a track record in fraud detection

Kazimi’s founding team previously built the security and fraud detection systems at Adjust, the mobile measurement company AppLovin acquired in 2021 for roughly USD 1 billion. They have also built similar systems for other mobile analytics companies and ad networks, giving them direct experience with the scale of the bot problem before founding Kazimi.

Petri Hyökyranta, who has known the founders for some time, backed Kazimi because its approach lets businesses detect bots at scale while staying compliant and privacy-preserving. He called it a paradigm shift for the mobile industry.

What Kazimi will do with the funding

Kazimi plans to use the new capital to scale its product and go after the wider mobile ad spend market. Jacek Lubinski, partner at Market One Capital and the round’s lead investor, said Kazimi could become a fundamental security layer for the global app economy in the AI era, providing secure, accurate first-party data while staying fully privacy-preserving.

Kazimi already works with mobile gaming studios, rewards platforms, and fintech apps. Its solution targets any app that spends marketing budget to acquire new customers or retain existing ones, a group that covers much of the mobile industry.

The bigger picture for Kazimi

Kazimi’s launch sits inside a wider shift in how the internet works. As bots increasingly outnumber genuine human traffic online, businesses across sectors are having to rethink how they verify authentic activity. For mobile app developers specifically, that shift raises the stakes on getting first-party data right, since flawed signals feed directly into both analytics and AI-driven decisions. Kazimi’s pre-seed round suggests investors see a real opportunity in solving that problem before it gets worse.

Frequently Asked Questions about Kasimi

What is Kazimi?

Kazimi is a Berlin-based mobile cybersecurity start-up. It helps app developers secure their first-party data, detect bot activity, and separate real users from fake ones in a privacy-preserving way.

How much funding did Kazimi raise, and who led the round?

Kazimi raised €2.2 million in an oversubscribed pre-seed round led by Market One Capital, with participation from IBB Ventures and several mobile industry executives.

Why is Kazimi’s technology needed right now?

Global mobile ad spend reached USD 419 billion in 2025 and is expected to exceed USD 1 trillion by 2035. At the same time, Cloudflare reported that bots overtook human activity on the internet for the first time in 2026, and AppsFlyer found that bot attacks drive most fraudulent activity on mobile channels.

Who founded Kazimi, and what is their background?

Kazimi’s founders previously built the security and fraud detection systems at Adjust, the mobile measurement company acquired by AppLovin in 2021 for around USD 1 billion. They have also built fraud-detection systems for other mobile analytics companies and ad networks.

Who are Kazimi’s customers?

Kazimi’s clients include mobile gaming studios, rewards platforms, and fintech apps, along with other businesses that spend marketing budgets to acquire and retain app users.

Picture Credits: Iraj Larik

ag-tech systems Closes Seed Round to Expand Chemical-Free Weed Control Globally

By

Markus Elsasser

-

08/21/2026

0

ag-tech.systems Closes Seed Round With NRW.BANK

NRW.BANK and family offices back ag-tech.systems’ next growth phase

Geilenkirchen-based ag-tech.systems has closed its Seed financing round. NRW.BANK, the promotional bank of North Rhine-Westphalia, and a newly formed investment vehicle, ag-tech Beteiligungs GmbH & Co. KG, both join as new backers. The company did not disclose the size of the round or individual investment amounts.

NRW.BANK invested through its NRW.SeedCap seed-financing programme. Meanwhile, ag-tech Beteiligungs GmbH & Co. KG brings together several family offices, coordinated by Nicolas Hübner, Managing Partner of Hübner Invest GmbH.

Shareholder base spans public funding to power electronics

These new investors join existing shareholders Raiffeisen Ventures GmbH, the venture arm of German agricultural cooperative Raiffeisen Waren-Zentrale Rhein-Main AG, and ISEG Leistungselektronik Sachsen GmbH, a specialist in high-voltage power electronics. ISEG’s expertise sits close to ag-tech.systems’ own electro-physical technology. Several angel investors with industry experience also back the company.

As a result, the round pulls together public-sector funding, agricultural know-how, entrepreneurial capital and engineering expertise. This mix gives ag-tech.systems a broad base of partners as it scales its plant protection platform.

CEO Marcus Lehnen said the backing gives the company both the financial runway and the industry expertise it needs to scale into new crops and markets. Hübner pointed to strong potential in the technology and its ability to offer farmers an alternative to conventional chemical weed control; he also confirmed he had brought the investor group together for the company’s next growth phase.

NRW.BANK Chairwoman Gabriela Pantring said the investment backs a North Rhine-Westphalia technology as part of the bank’s mandate to help young companies scale.

Where the fresh capital will go

The new funding will support ag-tech.systems’ next phase of product development and international expansion. Key plans include:

Market launch: Commercial launch of the epc.160 in 2026, built for potato desiccation, cover crop termination and invasive species control.
Global reach: Expansion beyond Germany into Canada, the Netherlands and the United Kingdom, plus larger machines for broadacre farming in North America.
Crop pipeline: New target crops including oat desiccation, sugar beet, malting barley, rapeseed, wheat, chickpeas and cotton.

Non-chemical alternatives have drawn growing interest across the plant protection sector in recent years, and ag-tech.systems’ expansion plans sit within that broader shift toward electro-physical and mechanical weed control methods.

ag-tech.systems at PotatoEurope 2026

ag-tech.systems will show its electro-physical weed control and desiccation technology at PotatoEurope 2026. The event runs on 9 and 10 September 2026 at Rittergut Gestorf in Springe, near Hannover, Germany. It’s one of the largest outdoor exhibitions in Europe for potato cultivation and machinery, giving ag-tech.systems a direct line to farmers and industry visitors.

With the Seed round now closed, ag-tech.systems has the capital and strategic partners in place to push its technology into new crops, new machines and new markets over the coming years.

Did ag-tech.systems disclose how much it raised?

No. The company confirmed new backers but kept the round size and individual investment amounts private.

What sets ag-tech Beteiligungs GmbH & Co. KG apart from ag-tech.systems’ other investors?

It’s a newly formed vehicle pooling several family offices, put together specifically for this round by Nicolas Hübner of Hübner Invest GmbH.

What is electro-physical weed control?

It’s the technology behind ag-tech.systems’ platform: an approach to desiccation and weed control that offers farmers an alternative to conventional chemical methods.

When does the epc.160 launch, and what’s it for?

ag-tech.systems plans to launch it in 2026 for potato desiccation, cover crop termination and invasive species control.

Beyond Germany, where is ag-tech.systems headed next?

Canada, the Netherlands and the United Kingdom, alongside larger machines designed for broadacre farming in North America.

How AloFor is Disrupting Global Travel-Tech from Romania

By

StartupValley

-

08/21/2026

0

AloFor: eSIM Connectivity for Global Travellers Ahmad Khodr Picture credit: Private / AloFor S.R.L.

AloFor is building a global eSIM platform that aims to make connectivity simpler, more transparent and more accessible for international travellers.

How would you introduce AloFor, and what inspired you to build an eSIM platform in Romania?

AloFor is a Romanian-built global travel-tech company focused on eSIM connectivity for international travellers. From the beginning, we built AloFor to compete internationally rather than as a local niche service. Through our web platform and mobile apps, travellers can access mobile-data eSIM plans across more than 200 destinations, with plans starting from USD 1 in selected markets.

The idea came from a simple problem: international connectivity is still more complicated and expensive than it should be. Traditional roaming can be costly, while buying a physical SIM after arrival creates unnecessary friction. I wanted to build a digital-first alternative that allows travellers to arrange their connectivity before or during a trip in a few simple steps.

Romania became the base from which we built the company, but the ambition was global from day one.

Ahmad, you are a Lebanese founder building a Travel Tech company in Romania. How has your international background shaped AloFor?

My background has shaped AloFor very directly. Being Lebanese and building a technology company in Romania gives me a naturally cross-border perspective. I understand what it means to move between countries, networks, languages and different customer expectations.

That experience influenced how we built AloFor: the customer should not need to understand the complexity of telecom infrastructure. A traveller simply wants reliable mobile data, a clear price, an easy installation process and support when needed.

It also shaped the scale of our ambition. I never saw AloFor as a company for one country. We built it from Romania for an international market, and that international mindset is part of the company’s identity.

What problem for international travellers did you want to solve when you founded AloFor?

The main problem is uncertainty and friction around mobile connectivity abroad.

Travellers often do not know how much roaming will cost, whether they should buy a local SIM, whether their phone is compatible, or how to get connected immediately after landing. Connectivity should be one of the simplest parts of a trip, yet it can still create unnecessary stress.

AloFor aims to make mobile data something travellers can arrange digitally. The customer can select a plan, install an eSIM and prepare connectivity without having to search for a physical SIM card after arrival.

The technology behind telecom is complex, but the customer experience should feel simple.

AloFor offers eSIM data plans for more than 200 destinations. Who is your main target group, and what matters most to these customers?

Our customer base includes tourists, business travellers, digital nomads, students, people visiting family abroad and anyone who wants a simpler alternative to traditional roaming.

What matters most is usually very clear: competitive pricing, dependable coverage, simple installation, transparent plan information and confidence that support is available if something goes wrong.

Customers increasingly expect connectivity to behave like any other modern digital product. They want to purchase it instantly from their phone, understand exactly what they are buying and begin using it without unnecessary steps.

Our job is to make a complex telecom service feel simple and accessible.

The eSIM market already includes established telecom companies and global connectivity providers. How does AloFor differentiate itself from these competitors?

The eSIM market is highly competitive, and AloFor operates in the same global arena as established international eSIM platforms and connectivity brands. We see that as an opportunity rather than a disadvantage.

The underlying mobile networks are provided by telecom infrastructure, so differentiation is increasingly created at the customer layer: pricing, product simplicity, transparency, support, speed and overall experience.

AloFor is founder-led, which allows us to move quickly, listen closely to customers and improve the product without the layers of complexity that often exist in larger organisations. We focus on competitive pricing, a straightforward purchasing journey, clear plan information and a consistent experience across web, iOS and Android.

Our goal is not simply to be another eSIM seller. We are building AloFor as a global travel-connectivity brand capable of competing with some of the best-known companies in the sector by being faster, simpler and closer to the customer.

Some AloFor plans start at just USD 1. How important is pricing to your strategy alongside simplicity, support and customer experience?

Pricing is an important part of our strategy because one of the strongest reasons travellers explore eSIM alternatives is to avoid expensive roaming.

Having plans starting from USD 1 in selected destinations also lowers the barrier for a new customer to try AloFor. But we do not believe a sustainable company can compete on price alone.

A low price has little value if installation is confusing, plan information is unclear or support is difficult to reach. The long-term objective is to combine competitive pricing with simplicity, transparency and a customer experience that makes people comfortable choosing AloFor again for their next trip.

Price can attract a customer for the first purchase. Trust is what brings them back.

What have you learned from your early customers?

The biggest lesson is that simplicity matters even more than founders often expect.

For people working in technology, terms such as eSIM activation, data roaming, APN settings or network selection may feel familiar. For many travellers, they are completely new.

Early customer feedback has helped us understand where instructions need to be clearer, where support needs to be faster and which parts of the journey should become more automatic. Every question a customer asks teaches us something about how the product can become easier to use.

Those early customers are not only using AloFor; they are helping us shape the company and the product.

What are the biggest challenges of building a global eSIM business from Romania?

One of the biggest challenges is delivering a simple experience on top of a very complex global telecommunications ecosystem.

Different countries, mobile networks, devices and technical conditions can all affect the customer experience, so consistency is extremely important. Another major challenge is trust. A traveller may discover AloFor only days before an international trip and then depend on the service in another country. We have to earn that trust quickly.

Building from Romania also means thinking globally from the beginning: languages, payment preferences, device compatibility, support expectations and travel behaviour vary across markets.

At the same time, Romania is not a limitation. Digital infrastructure allows a technology company built here to serve customers globally. For us, the location is the base; the market is international.

International connectivity is increasingly shifting towards digital purchases before departure. How do you expect this trend to change the travel market?

I believe mobile connectivity will increasingly become a standard part of the pre-trip purchasing journey.

Travellers already book flights, hotels, transport and insurance before departure. eSIM allows mobile connectivity to move into the same digital workflow because the traveller no longer needs to physically obtain a SIM card at the destination.

Over time, I expect connectivity to become more integrated into travel platforms, booking experiences and other digital services. This also changes competition in telecom because the customer relationship is no longer controlled only by the operator in the destination.

Digital platforms can now compete on how easily people discover, buy, activate and manage international connectivity. That shift creates space for companies like AloFor to build a global brand around the customer experience.

AloFor relies on mobile operators’ networks while managing the customer experience digitally. Where do you see the greatest opportunities in this model?

The biggest opportunity is the digital layer between the telecom network and the traveller.

Mobile operators provide the essential infrastructure, but customers increasingly interact with connectivity through digital platforms. That creates room to improve pricing transparency, discovery, onboarding, activation, support and the overall user journey.

It also allows AloFor to bring many destinations into one consistent experience. A traveller moving between countries should not have to learn a completely different process each time.

I believe the strongest opportunity is to make global connectivity feel unified from the customer’s perspective, even when many different networks and markets exist underneath the service.

What are your next priorities for AloFor, and how do you plan to grow the platform internationally?

Our next priority is to turn AloFor from an emerging platform into an increasingly recognised global travel-connectivity brand.

That means continuing to improve the product, making the purchasing and activation journey even simpler, strengthening customer support, growing international distribution and building strategic partnerships that can extend AloFor’s reach.

We will continue developing the experience across web, iOS and Android while expanding both direct-to-consumer growth and international business opportunities.

AloFor already operates with a global product footprint across more than 200 destinations. The next phase is not simply about adding more countries; it is about building stronger brand trust, increasing repeat usage and proving that a company built in Romania can compete successfully in a global market alongside much larger international players.

What three pieces of advice would you give to founders who want to build an international startup from day one?

First, solve a problem that exists across borders. If the customer problem is genuinely international, the company can be designed for global scale from the beginning.

Second, stay extremely close to your customers. Early users will show you problems that no presentation, market report or business plan can reveal. Their questions often become the clearest product roadmap.

Third, do not wait for everything to feel perfect before entering the market. International companies are built through continuous learning. Launch, listen, improve and keep moving.

One additional lesson from my own journey is that geography does not have to define ambition. I am a Lebanese founder building AloFor from Romania for an international market. In technology, a founder can combine experiences from different countries and turn that perspective into a competitive advantage.

Picture credit: Private / AloFor S.R.L.

Thank you Ahmad Khodr for the Interview

Statements of the author and the interviewee do not necessarily represent the editors and the publisher opinion again.

HonwinOne: How Keith Honwin builds global infrastructure from eswatini

By

StartupValley

-

08/19/2026

0

HonwinOne: technology and infrastructure for growth Keith Honwin

HonwinOne is building technology and infrastructure designed to connect people, businesses and institutions across markets while reducing digital fragmentation.

Keith, could you start by introducing yourself and telling us what inspired you to found HonwinOne?

I’m Keith Honwin, founder and CEO of HonwinOne. I come from Eswatini, but a significant part of my education and professional development took place in Azerbaijan, where I studied International Relations at ADA University and later worked around technology, education and innovation through organisations including Taskool and INNOLAND.

Living and working between very different environments changed the way I thought about technology.

I began to notice that many of the problems people experience are not caused by a complete absence of technology. Often, technology already exists, but it is fragmented, difficult to access, disconnected or simply not designed around the person who is supposed to use it.

A business might need several different systems just to serve one customer. A young person looking for a scholarship might search through dozens of websites, social-media pages and application portals. Someone trying to discover a nearby service, attend an event or make a payment may move between several disconnected platforms.

That stayed with me.

HonwinOne grew from a belief that technology should make life simpler, expand access and give people more possibilities—not create another layer of complexity.

I also came to understand something personally: if you genuinely want to change something, you eventually have to be willing to dream beyond borders.

That means geographical borders, industry boundaries and sometimes even the limits people place on what they believe is possible.

Eventually, I stopped waiting for somebody else to build the systems I wanted to see and decided to start building them myself.

What exactly is HonwinOne, and which core problem or opportunity is at the heart of the company?

HonwinOne is a technology company building intelligent digital infrastructure for people, businesses and institutions.

The core problem underneath almost everything we work on is fragmentation.

A customer may discover a business in one place, communicate with it somewhere else, pay through another system and receive very little visibility afterwards.

A young person looking for an international opportunity may have to search hundreds of scattered sources.

A merchant may need separate systems for orders, payments, promotions, staff management, customer communication and reporting.

These appear to be very different problems, but underneath them is the same issue: systems do not communicate well enough, and people are often forced to do work that technology should be doing for them.

Our approach is therefore not simply to ask:

“What application can we build?”

We ask:

“Where is friction being repeated across an entire system, and what infrastructure could remove it?”

That thinking has led us into local commerce and discovery through Pickly, global youth opportunity infrastructure through GYAP, payment orchestration through H1Pay, event technology, APIs and our broader work around artificial intelligence.

The long-term opportunity for HonwinOne is not simply having several products.

It is creating reusable infrastructure around discovery, identity, intelligence, transactions and action that can support many different services.

HonwinOne works across AI, commerce, local discovery, payments, event technology and youth opportunity infrastructure. How do these different areas fit together within one company and one overarching strategy?

This is one of the most important questions about HonwinOne.

From the outside, it can initially look as though we are building several unrelated products.

Internally, we see something different.

We see one infrastructure company expressed through different use cases.

Pickly deals with local discovery and demand. Someone may want to find a restaurant, discover an event, book a service, order something or interact with a nearby business.

GYAP deals with another form of discovery: opportunity. Instead of asking, “What can I find around me?”, somebody may be asking, “Which scholarship, fellowship, university, internship, grant or programme is right for me?”

Our event technology addresses ticketing, verification, access, event operations and the infrastructure surrounding physical experiences.

Our work around H1Pay addresses the transaction layer and the possibility of intelligently orchestrating payments across different systems and payment rails without requiring every HonwinOne product to become a financial institution itself.

Artificial intelligence sits across all of these as an intelligence layer.

So the connection is not the industry.

The connection is the infrastructure underneath the industries.

We are building around recurring functions:

discover, understand, identify, transact and act.

The individual products allow us to solve specific problems and reach different users.

The shared infrastructure gives HonwinOne leverage.

Over time, the objective is that improving one part of our infrastructure can strengthen several products rather than requiring us to rebuild everything independently.

Which products and technologies have you already launched, and which of them are currently the most important for HonwinOne’s growth?

Our two immediate priorities are Pickly and GYAP.

Pickly is our local discovery and commerce platform. It is designed to bring restaurants, stores, services, events, local experiences and other everyday needs into one intelligent discovery environment while giving merchants tools to establish and manage their digital presence.

Rather than forcing people to already know exactly where to search, Pickly is designed around discovery: helping somebody understand what is available, what is happening around them and what they can do next.

GYAP—the Global Youth Aspiration Platform—is focused on opportunity discovery.

It brings together scholarships, fellowships, internships, grants, training programmes, education pathways and other opportunities that are often scattered across the internet.

GYAP is particularly meaningful to me because education, youth development and international exposure have been an important part of my own journey.

Both Pickly and GYAP are publicly accessible and are now entering the most important stage for us: market activation, distribution and real-world adoption.

We have also developed significant technology around H1Pay and our event systems.

However, I think founders have a responsibility to distinguish between technology that has been built and infrastructure that has completed every commercial, regulatory and partner integration required for full-scale operation.

H1Pay, for example, remains in the integration and partnership stage rather than something we would describe today as a fully connected payment network.

One of our biggest lessons has been that having the ability to build many things does not mean we should try to scale everything simultaneously.

Our priority now is turning technology into sustained usage.

You are building HonwinOne from Eswatini with the ambition to serve global markets. What advantages and challenges come with building a technology company from a relatively small and underrepresented startup ecosystem?

Building from Eswatini gives you a very interesting kind of discipline.

In a smaller market, problems are close to you.

You can see how businesses actually operate. You see where people improvise because a system does not work properly. And you see where digital services are too complicated, too expensive or simply unavailable.

That proximity can be a major advantage because you are not solving theoretical problems.

You are surrounded by the consequences of systems that do not work well enough.

It also forces resourcefulness.

You cannot assume that unlimited capital, specialised talent or infrastructure will always be available. You have to find ways to build efficiently, reuse systems, automate repetitive work and make each decision count.

The challenges are equally real.

Founders from smaller and less represented ecosystems often have to prove two things simultaneously.

First:

Does the technology work?

And second:

Can a serious global technology company really come from this place?

Access to venture capital is more limited. Specialised technical talent can be harder to find. Certain APIs, payment systems and integrations that startups in larger ecosystems may take for granted can require considerably more work.

But I do not see Eswatini as something HonwinOne has to overcome.

I see it as part of our perspective and potentially part of our advantage.

If you can build technology that is simple, resilient and useful under real constraints, many of those lessons can travel.

And sometimes being underestimated gives you the freedom to think differently.

Many startups in emerging ecosystems initially focus on solving problems within their domestic markets. Why did you decide to think internationally from the beginning?

Partly because my own life taught me to think internationally.

I grew up in Eswatini, studied in Azerbaijan and have worked with people and organisations across different countries and cultures.

Once you experience different parts of the world closely, borders begin to look different.

You realise that people living thousands of kilometres apart can experience remarkably similar problems.

A small business in Eswatini struggling with digital discovery may have something in common with a small business in Mozambique, Azerbaijan or another emerging market.

A student in Southern Africa searching for an international scholarship may experience many of the same frustrations as a student in Central Asia or somewhere else in the world.

The interface may change.

The language may change.

The payment method may change.

The regulation may change.

But many of the underlying problems are universal.

That is why I never wanted Eswatini to become the ceiling of HonwinOne’s ambition.

It can be an excellent place for us to learn, build, test and prove things.

But the architecture has to be capable of travelling.

At the same time, international ambition cannot mean blindly copying exactly the same product into every market.

Countries have different languages, cultures, regulations, payment behaviours and commercial structures.

The challenge is to build technology that can localise without losing its core.

I have come to believe that meaningful change requires us to dream beyond borders and to work with people who do not necessarily come from the same place we do.

The biggest problems do not respect borders, so our thinking should not be limited by them either.

Artificial intelligence is a central part of your work. How are you currently using AI within HonwinOne, and where do you see the greatest potential for AI across your products and infrastructure?

I believe artificial intelligence becomes most valuable when it stops feeling like a separate feature and starts quietly improving how an entire system works.

Within HonwinOne, we are applying AI in areas including discovery, matching, information organisation, business onboarding, personalisation and decision support.

A merchant should not need to understand complicated software just to establish a useful digital presence.

A young person should not have to manually search hundreds of opportunities to find the few that are genuinely relevant to their goals.

A customer should not eventually have to navigate exactly the same generic interface every time if the system can, with appropriate consent, understand their context and what they are trying to accomplish.

However, I am also careful about calling every automated process “AI”.

The longer-term thesis for us is much more ambitious.

Most software today still waits for the user to tell it almost everything.

You open an application, navigate menus, repeat information and constantly instruct the system.

We believe software will increasingly become adaptive.

It should understand where somebody left off, what they usually need, what they are trying to achieve and what the most useful next action may be.

That does not mean removing human choice.

It means removing unnecessary friction.

Our long-term thinking is toward increasingly personalised digital intelligence—systems that become more useful because they learn from context, history and user preferences, while keeping the user in control of that relationship.

For us, therefore, the future of AI is not simply putting another chatbot inside every product.

It is building adaptive digital infrastructure that progressively becomes more useful to the person using it.

You mention building multiple products with limited resources. How have these constraints influenced the way you develop products, set priorities and make decisions as a founder?

Constraints have shaped HonwinOne profoundly.

When resources are limited, you cannot solve every problem by hiring another department.

You have to think architecturally.

You ask whether something can be reused.

And you automate repetitive work.

You create shared components.

You become extremely conscious of where time and money are going.

It has also kept me very close to the products.

I spend a significant amount of time testing flows myself, looking for things that break, questioning assumptions and asking where we are creating unnecessary friction for users.

AI-assisted development has dramatically changed what small teams are capable of building.

A very small team can now experiment, design, test and develop at a speed that would have required considerably more resources only a few years ago.

But that creates another danger.

Software can sometimes be produced faster than it can be properly validated.

So testing becomes even more important.

One of the biggest things I have learned is that speed and quality should not be treated as opposites.

The objective is to create a development system where you can move quickly, test aggressively, learn from reality and continuously improve.

Constraints have also forced us to become more disciplined about priorities.

We have built a lot.

The next chapter is not about proving that we can build another product.

It is about proving that we can take the strongest products we already have and create meaningful, repeatable adoption around them.

What is the business model behind HonwinOne, and how do you plan to build a sustainable and scalable company across different products and markets?

The commercial model varies by product, but the broader HonwinOne strategy is to create recurring infrastructure and platform revenue rather than depending on a single source of income.

For Pickly, the business side of the network creates several potential revenue streams, including merchant services, subscriptions, transaction-related revenue, promotions and value-added operational tools.

For GYAP, the commercial opportunity is primarily institutional.

Universities, education providers, foundations, NGOs and other organisations can use the platform for opportunity distribution, recruitment, programmes, profiles and related infrastructure.

Our payment work is designed around orchestration and APIs rather than the assumption that HonwinOne itself must become a bank.

As the appropriate commercial and regulatory integrations mature, that can create infrastructure and usage-based revenue.

Our event systems create additional B2B opportunities around ticketing, access infrastructure, verification, cashless experiences and event operations.

At the HonwinOne level, there is another important layer.

We can license technology, APIs and specialised digital infrastructure to companies, institutions and governments rather than requiring every organisation to build the same underlying capabilities from the beginning.

So the broader opportunity is a combination of consumer distribution, business software, transactions, licensing and infrastructure.

However, at our current stage, I believe there is something more important than trying to monetise every possible feature.

We first need to prove repeatable usage and measurable value.

If people genuinely choose the products, businesses receive measurable outcomes and institutions can achieve better results through the infrastructure, sustainable monetisation becomes much easier to build.

What stage is HonwinOne currently at in terms of users, customers, partnerships and traction, and which milestones have been particularly important so far?

We are still an early-stage company, and I prefer to be transparent about that.

We do not yet have the enormous user, GMV or revenue figures that later-stage technology companies can point to, and I would rather build those numbers than manufacture a story around them.

Our most important milestone so far is simpler:

We moved beyond ideas and presentations into working products.

Pickly is publicly accessible and includes the foundations for consumer discovery, merchant onboarding and digital commerce.

GYAP is also publicly accessible and is evolving into a technology platform for global opportunity access.

Alongside these products, we have developed payment, ticketing, AI and API capabilities.

We are now actively pursuing merchants, institutions, businesses, ecosystem organisations and public-sector relationships across different markets.

That is the stage HonwinOne is entering now:

commercial validation and distribution.

The milestones that matter next are very clear:

active merchants, active users, repeat usage, transactions, institutional customers, strong partnerships and evidence that our products are becoming part of people’s normal behaviour.

We are also entering HonwinOne’s pre-seed chapter because the bottleneck has changed.

Earlier, the biggest question was:

Can we actually build?

We have answered a significant part of that question.

The next question is:

What happens when that ability to build is matched with the capital, team, partnerships and distribution required to scale it?

That is the chapter we are now working toward.

From your perspective, what needs to change for more technology companies from African and other emerging startup ecosystems to compete successfully on a global level?

The first thing is access to serious risk capital.

There are many competitions, programmes, incubators and small grants across emerging ecosystems, and they can be valuable.

But companies with global ambitions eventually need investors who are prepared to finance uncertainty.

Technology companies are built around things that have not been proven yet.

Someone has to be willing to take that risk alongside the founder.

The second issue is access to markets.

Governments and large corporations regularly speak about supporting innovation, but startups need procurement systems that allow young companies to become actual suppliers—not only participants in conferences, competitions or innovation programmes.

Third is global connectivity.

Talent, investors, technical partners, mentors and distribution are not evenly spread around the world.

A founder’s access to the right network can sometimes matter almost as much as the quality of the technology.

But I also think something needs to change in the way founders from emerging markets see ourselves.

We should not believe that every African technology company must remain an “Africa-only” company.

We should solve local problems deeply.

And we should understand our environments better than anybody else.

But we should build to global standards.

Innovation does not have a nationality.

A technology company should ultimately be evaluated on whether its technology works, whether people need it, whether its economics make sense and whether its team can execute.

The next important global technology company does not necessarily have to come from the place everybody is currently watching.

It can come from somewhere unexpected.

And creating more pathways for those unexpected companies to access capital, markets and global networks will benefit everyone.

Where do you want HonwinOne to be in the next three to five years, and what is your long-term vision for the company?

In three to five years, I would rather have two or three HonwinOne products creating genuine value at meaningful scale than twenty impressive demonstrations.

I want Pickly operating across multiple markets with strong local merchant and consumer networks.

I want GYAP to become genuine infrastructure connecting young people with universities, institutions, programmes and opportunities across borders.

And I want our payment and API infrastructure integrated wherever the appropriate commercial and regulatory partnerships exist.

I want businesses, governments and institutions to be able to use HonwinOne infrastructure rather than rebuilding the same underlying systems from the beginning.

And across all of those products, I expect artificial intelligence to become less visible but considerably more important.

It should quietly improve discovery, personalisation, decision-making and how people interact with technology.

But my long-term vision goes beyond products.

During this journey, I came to understand that if we genuinely want to help change the world, we have to be willing to dream beyond borders and work together across them.

No individual founder, company or country solves the world’s biggest problems alone.

And I am very aware of the odds.

Startups fail.

Technology changes quickly.

Markets can reject you.

Capital can disappear.

Things can go wrong.

There are thousands of reasons why an ambitious idea may never become what its founder imagined.

But I think there is something worth pursuing even in the smallest probability.

Even if there were only a 0.0001% chance that something we build could eventually improve the lives of millions of people, create opportunities for people who previously did not have them, help a small business grow, or allow somebody to access something they once believed was beyond their reach, I believe that possibility is worth working toward.

That is the kind of company I want HonwinOne to become.

A company built from Eswatini, working across borders, collaborating with people from different parts of the world and using technology to expand what is possible for others.

I also want a young person from a small country or an overlooked community to see HonwinOne one day and realise that where you begin does not have to determine where your ideas can go.

We are under no illusion that what we are trying to accomplish is easy.

The odds of failure are real.

In many ways, we are trying to prove something that can sometimes feel impossible.

But perhaps that is precisely why it is worth attempting.

We are here to build useful technology, help people, work with others and pursue that tiny possibility that something we create can leave the world better than we found it.

We are here to try to prove the impossible.

And if HonwinOne can help demonstrate that meaningful technology can be built from anywhere, that people from different countries can build together, and that even the smallest starting point does not have to define the size of an ambition, then this journey will have meant something far beyond the company itself.

Picture credit: Eddie Honwin / Private

Thank you Keith Honwin for the Interview

Statements of the author and the interviewee do not necessarily represent the editors and the publisher opinion again.

From Construction Sites to AI Founder: How Alin Hoisan Built Prompt Director

By

StartupValley

-

08/18/2026

0

Prompt Director: AI for creators with Alin Hoisan -Founder

Prompt Director combines AI with a practical creative workflow, helping creators turn their ideas into images, videos, campaigns and other creative projects.

Alin, could you start by introducing yourself and telling us how your journey from working on construction sites in the UK eventually led you to become the founder of Prompt Director?

I am Alin Hoisan, a Romanian creator living in the UK. My path into this world did not begin in a startup office. It began on construction sites. I started as a labourer, then became a traffic marshal, supervisor and eventually an assistant manager. Those years taught me discipline, responsibility and what it means to keep showing up even when the work is hard.

At the same time, I always had ideas. I wanted to write, create music, make videos and build things, but I did not have a team, a big budget or a technical background. When I discovered generative AI, I realised that the distance between an idea and a finished project had become much smaller. I began learning every day after work and family time. Prompt Director grew out of that journey: it is the platform I wished had existed when I was trying to turn imagination into something real.

You describe opening ChatGPT in 2023 as a moment that changed everything for you. What happened after that first encounter with generative AI, and when did curiosity turn into the ambition to build something yourself?

At first, I opened ChatGPT simply out of curiosity. I wanted to understand what it could do. Very quickly, it became clear to me that this was not just another app. It could help a person learn faster, organise ideas, write, research, solve problems and create things that previously required several specialists.

Curiosity became ambition when I started finishing real projects. I wrote and published books on Amazon, even though they did not make money. Then I built BounceVoid, my mobile game, working on it before the gym and after work for almost a year. Each project proved the same thing: I did not need permission to begin. I needed patience, consistency and the willingness to learn. Once I saw that I could take an idea all the way to a public release, I stopped seeing AI only as a tool. I started seeing it as a bridge between having an idea and having the ability to build it.

What exactly is Prompt Director, and what problem are you trying to solve for creators with the platform?

Prompt Director is an AI creation platform designed to help people move from a rough idea to a clear creative result. It began as a professional prompt library, but it has grown into a wider studio for creating images, video concepts, storyboards, AI advertising ideas, music-related concepts and creative campaigns.

The problem I am trying to solve is simple: many people have strong ideas but feel blocked by the technical side of AI. They do not know how to write a useful prompt, which tool to choose, how to keep a visual style consistent or how to turn one thought into a complete piece of content. They end up jumping between websites and copying prompts without knowing why something works.

Prompt Director gives them a clearer path. It combines ready-to-use professional prompts, visual prompt-building tools and direct creation features in one place, including Romanian-language support. The goal is not to make everybody a prompt engineer. It is to help more people express what they already have in their heads.

How does Prompt Director work in practice? If a creator comes to the platform with just an initial idea, what can they create and what does the process look like?

A creator can begin with something very simple: a product, a song, a character, a mood or even one sentence. From there, Prompt Director helps turn that starting point into a structured creative direction. They can use the prompt library, a visual prompt builder or dedicated tools for storyboards, AI ads and directing scenes.

For example, someone with a product idea can create a cinematic image concept, develop it into a short video sequence, build a storyboard and write the creative prompt needed for generation. A musician can shape cover art ideas, visual scenes and campaign content around a release. A small business can build a clearer AI advertising concept without needing an agency for every first draft.

The platform is made to reduce the friction between the first idea and the first usable result. It does not replace taste or imagination. It gives creators a practical structure so they can spend more time developing their idea and less time fighting with the tools.

There are already numerous AI tools for generating images, videos, music and marketing content. What makes Prompt Director different from using several individual AI tools separately?

The individual tools are powerful, but using them separately can be confusing, especially for someone who is new to AI creation. You might have an image generator open in one tab, a video generator in another, a prompt document somewhere else and no consistent system connecting them.

Prompt Director is built around the creative workflow, not just around one model. It gives users direction before generation: structured prompts, visual builders, storyboards, AI advertising formats and creative guidance that help an idea stay coherent from one stage to the next. The platform also brings those parts together in a simpler environment, with a focus on practical use rather than technical language.

I built it from the perspective of someone who actually creates music videos, visual campaigns and content every day. I know that the hard part is often not pressing Generate. It is knowing what to create next, how to describe it and how to make all the pieces feel like they belong to the same project.

Who is Prompt Director primarily designed for, and which types of creators or businesses do you believe can benefit most from the platform?

Prompt Director is for independent creators first: musicians, content creators, filmmakers, designers, writers, photographers, social media creators and people who have an idea but do not yet have a production team. It is also useful for small businesses that want to create more original visual content, product concepts, short-form ads or social campaigns without having to start from zero every time.

I especially care about people who may feel outside the traditional creative or technology world. You do not need to come from a design agency, a software company or an art school to have a powerful idea. Many people have creativity, life experience and stories, but they need a more accessible way to bring them to life.

For experienced creators, the platform can save time and help organise a workflow. For beginners, it can remove the fear of the blank page. In both cases, the purpose is the same: make AI creation feel more understandable, more practical and more human.

You built Prompt Director without coming from a traditional technology or startup background. How did you acquire the skills necessary to turn your idea into a working platform, and what were the biggest challenges along the way?

For me, learning came through building. I used AI as a learning partner, but I did not expect it to do the work for me. I spent countless hours testing, researching, watching tutorials, asking questions, breaking things and trying again. Every project taught me something different: publishing books taught me how to finish; BounceVoid taught me product thinking and persistence; music and video work taught me creative direction; Prompt Director brought all of those lessons together.

The biggest challenge was not one technical problem. It was carrying a big idea while still working, looking after my family and doing almost everything alone. There were moments when I did not know what the right next step was, when something did not work and when it would have been easier to stop.

But not coming from a traditional background also gave me a useful perspective. I build for people like me: people who want to create but do not speak in technical terms. My goal has always been to make complicated things feel more possible.

What is the business model behind Prompt Director, and how do you plan to turn the platform into a sustainable company?

Prompt Director currently uses a simple one-time lifetime-access model. I wanted the entry point to feel fair and accessible, especially for independent creators who are already paying for several AI services. The platform is designed to give value through a growing prompt library, creative tools, visual builders and a more connected creation workflow.

Long term, sustainability will come from continuing to improve the product around real creator needs: useful features, reliable generation workflows, premium tools where they genuinely save time, and partnerships or services that fit the platform naturally. I do not want to build a business that depends on keeping users confused or locked in. I want people to stay because Prompt Director helps them make better work.

At this early stage, I am realistic. The focus right now is product quality, trust and listening carefully to how users create. If I build something people return to because it removes friction from their creative life, the business model can grow in a healthy way around that value.

What stage is Prompt Director currently at in terms of users, traction and product development, and what kind of feedback have you received so far?

Prompt Director is live and still developing quickly. It has grown far beyond the first prompt-library idea and now includes more than 260 professional prompts, visual prompt-building tools, storyboards, AI Ads, AI Director, AI Beats and direct image and video creation tools. I keep improving it from real use, because I am also one of the people using it every day in my own creative work.

I prefer to be honest about the stage: this is an independent, bootstrapped product, not a company claiming huge numbers before they exist. The most meaningful traction for me is that people understand the purpose immediately: they want a clearer, less intimidating way to work with AI. The feedback I value most is when a creator says, “I finally know how to start,” or when a feature saves them from losing hours between different tools.

The product is not finished, and that is exciting. I see it as a living platform that becomes stronger as more creators show me where the real obstacles are.

You also created the AI music artist Selyna Turcu and developed projects such as BounceVoid. What did these experiments teach you about AI-driven creativity, and how have they influenced Prompt Director?

Selyna Turcu and BounceVoid taught me that AI is most powerful when it helps a person finish a complete vision, not just generate isolated pieces. With Selyna, I created the artist identity, music, production, visuals, videos, lip-sync content and social storytelling. The project has now generated more than 30 million organic views across TikTok and Facebook, with no advertising spend. That showed me that AI-made work can still connect emotionally when there is real creative direction behind it.

BounceVoid taught me a different lesson: building something alone requires patience, testing and a willingness to improve the small details nobody sees. I spent around a year on it, often working in the small hours around my job and family life.

Both projects shaped Prompt Director. They showed me that people do not only need a generator. They need a process: how to begin, how to develop an idea, how to keep it consistent and how to turn it into something they are proud to release publicly.

Generative AI is rapidly changing the creative industries while also raising questions around copyright, originality and human creativity. How do you see this development, and how do you address these issues with Prompt Director?

I understand why people have concerns. AI is moving quickly, and creative work is personal. Questions about copyright, consent, training data and attribution deserve serious attention, not just marketing answers. We need clearer rules, responsible platforms and respect for the people whose work and identities are involved.

At the same time, I do not believe AI removes human creativity. A tool can generate options, but it cannot replace the life experience, taste, decisions and emotional intention of the person using it. The difference is visible when someone simply presses a button compared with when someone directs a project from beginning to end.

With Prompt Director, my approach is to make the creative process more intentional. The platform is built to help users develop their own ideas and describe their own vision, rather than copy somebody else’s identity or work. I also believe transparency matters. With Selyna Turcu, for example, I have been open that she is an AI artist. Honesty is important if this new creative space is going to earn trust.

Where would you like Prompt Director to be in the next three to five years, and what is your long-term vision for the company?

In the next three to five years, I want Prompt Director to become a trusted creative home for people who have ideas but do not yet have a full team, technical background or large budget. I would like it to be known not only as a place for prompts, but as a practical studio where someone can shape an idea into images, videos, campaigns, stories and new creative projects.

I also want to keep the platform accessible. My own story is proof that talent and ambition do not always arrive with the traditional opportunities around them. A person can come from construction, retail, parenting, music or any other walk of life and still build something meaningful when the right tools and guidance exist.

The long-term vision is bigger than one feature or one AI model. I want Prompt Director to help creators keep their voice while technology changes around them. If it gives someone the confidence to start, the structure to improve and the tools to finish, then it is doing what I built it for.

Picture credit: Ionut Alin Hoisan, AI-assisted portrait based on my own original photo

Thank you Alin Hoisan for the Interview

Statements of the author and the interviewee do not necessarily represent the editors and the publisher opinion again.