Unibright AMA Recap: AI Compliance, Provenance, and Real-World Proof
April 2026 — This AMA marked a turning point: proof of execution is no longer theoretical, but is now being applied to real-world regulatory challenges — particularly around AI auditability, content provenance, and the practical need to make blockchain infrastructure invisible to end users.
The fourth installment of Unibright’s monthly AMA series introduced a new voice to the conversation and went deep into these developments.
Joining moderator Zoe were Marten (Co-Founder & CEO), Stefan (Co-Founder & CTO), Pierre (Co-CEO of SPO Consulting and founder of Audiogroup), and a new participant: Martin from Tunesat.
A New Voice: Martin and the Tunesat Story
The session opened with an introduction to Martin, whose 25-year journey through the music industry brings a concrete use case to the Unibright ecosystem.
Martin’s path started as a professional guitarist and music producer before evolving into TV production and rights management. In 2009, he joined Tunesat — a company specializing in audio fingerprinting technology to detect music usage on television. The work involved helping publishers, composers, and labels claim royalties based on technology-verified data rather than traditional reporting.
Martin: “Getting fingerprinting technology accepted by the rather traditional structures of broadcasters and PROs was not easy. But finally we achieved that for the first time in German history, a publisher got paid on fingerprinting monitoring data. Which was a huge success.”
As Tunesat expanded from linear TV monitoring to internet detection, a new challenge emerged: proving that something happened at a specific point in time on the internet, where content can change or disappear.
Martin: “With internet monitoring, you find something and then you have to make a proof that at a specific point in time, something happened — because the asset can be changed or taken down. Proving this in a verifiable and tamper-proof way is rather critical.”
This is precisely where blockchain adds the missing link. For Martin — self-described as “not a blockchain guy at all” — the appeal is purely practical: having an independent, tamper-proof foundation that makes digital rights evidence legally defensible.
The Missing Link in Enterprise Software
Stefan framed the Tunesat integration as exactly what the restructuring efforts of the past two years aimed to achieve: adding blockchain’s proof capabilities to established, working business processes rather than building from scratch.
Stefan: “The complete use case around Tunesat has a lot of parts — data, digital assets, users that need access and management. 80 or 90% of the work is in a classical software-as-a-service environment. And then a proof of execution layer adds to something that has existed since 2009. It’s not that this whole industry waited for blockchain to appear. It already worked. But now we can put some final missing links to the complete picture.”
Pierre emphasized how this changes the go-to-market dynamic. In the past, getting beyond pilots and test cases was a struggle. Now, with the combination of blockchain technology and full software services, the entire stack can move into commercial usage quickly.
Pierre: “In the past, it was a struggle to get past test cases and pilots. Now with this combination of blockchain technology and software services, we are very quickly in a position where this whole stack is being put into real commercial usage.”
EU AI Act: Verifiable Execution as Regulatory Requirement
The conversation turned to how the proof of execution platform serves companies needing to demonstrate compliance under the EU AI Act.
Marten explained what “verifiable execution” means in practice: being able to prove after the fact — or in real time — what an AI system did, how it arrived at its output, and under what conditions it operated.
Marten: “If you train a model and later do inference, you have to be able to prove what information was used to get to that output. And with AI agents acting as a group, you have to prove how these agents worked and what was the basis of their doing. That’s what the AI Act is about.”
This isn’t the only regulatory driver. Digital product passports, ESG reporting, and supply chain requirements all create similar demands: proving that specific events happened in specific ways.
Marten: “There are a lot of things forced by law where businesses have to prove that a certain event happened, whether it was an AI system or traditional system. That’s a prime use case for our platform.”
Beyond Hashes: Building a Trust Mesh
A recurring theme throughout the session was that effective proof of execution requires far more than simply writing hashes to a blockchain.
Marten: “As we learned the last eight years, it’s not just about putting a hash on a blockchain. There’s much more to come up with a proof that is valid in front of a court, that is valid with current regulation. It’s about having off-chain auditors in place. It’s about this idea of building a trust mesh.”
The platform incorporates multiple layers: different players with different access levels, versioned proofs with amendments, privacy features where some data is public while other data is visible only to auditors.
Stefan elaborated on the role-specific needs within a single use case like Tunesat:
Stefan: “There are actually different roles participating in that system at different points in time. And the users would not be happy if you just pointed them to a hash in a blockchain. You need to put this into perspective of the process.”
Martin illustrated this from the practical side: a license manager needs different information than a lawyer, and both need different views than the party who may have been using music without proper licensing. Each role requires appropriate disclosure of the relevant data subset.
User Experience: Making Blockchain Invisible
Perhaps the most emphatic point across all speakers was that end users must never need to understand blockchain technology.
Martin: “In real-world business, you always need some front end where people really understand what they are seeing. You cannot talk about hash sums with real clients. That does not work.”
Marten recalled a pivotal realization from years ago:
Marten: “We stopped selling blockchain a few years ago. We had this epiphany — it’s not about blockchain first. Of course, under the hood, everything is blockchain first. But that’s nothing a customer wants to see. A huge part of our system is a web front end where every role is served.”
Pierre turned this into a validation methodology: because he and Martin aren’t blockchain experts, they serve as ideal first-line testers. If they don’t grasp something, external clients certainly won’t either.
Pierre: “We are the best initial testers, because as soon as we don’t grasp it, we can be sure that other people will also not understand it. This has been a very fruitful partnership — it automatically puts everything into the client’s perspective.”
Provenance: From Internal Use to Industry Scale
The discussion of AI provenance moved from abstract concept to concrete implementation through Audio Group’s sound effects business.
Pierre explained the internal use case: Boom Library, one of the world’s most successful sound effects libraries, has content created by roughly 20 sound designers who receive revenue shares based on their contributions. When new content is created from existing material — whether through human mashups or generative audio — tracking what was used becomes essential for fair distribution.
Pierre: “Whenever we create new content out of existing content, we need a clear trail of what has been used to do what. If at a certain point we start doing generative audio based on our original content, all of a sudden we’re doing new stuff that needs to be distributable in terms of revenue. This is exactly where the provenance layer comes in.”
The scaling path is clear: prove it works internally first, then extend to a larger marketplace with hundreds of suppliers, then demonstrate to major rights holders dealing with the same challenges at massive scale.
Pierre: “We can show this system to people who have this on a much larger scale. We can tell them: look, it’s already working in these two reputable brands. We can easily expand to the bigger scope.”
AI Agents and the Hallucination Question
A community question raised the issue of AI hallucinations undermining trust in blockchain records. The team addressed this from multiple angles.
Marten confirmed that AI agents can interact with the platform directly — either through APIs or smart contract calls — to prove their actions automatically. But the hallucination concern points to something deeper.
Stefan reframed the problem: in a proof of execution setup where AI agents do work, human beings remain in the loop to verify automated processes before giving final approval.
Stefan: “We have AI agents that work together but need the human part for the actual proving that the process they prepared is really the right one. This could be one way to understand hallucinations — the AI is doing something wrong, and exactly then you need a proof of execution framework where the last proof part is done by humans.”
Marten connected this back to regulatory requirements: under the EU AI Act, hallucinations aren’t just undesirable — they’re non-compliant. Systems must be deterministic and verifiable.
Marten: “Under the AI Act, it has to be deterministic or verifiable. Hallucination is just not allowed to happen.”
Scaling for High-Frequency Transactions
With AI agents potentially driving massive transaction volumes, the question of scalability arose.
Marten explained that the platform is blockchain-first, built on Layer 2 solutions like Optimism or Arbitrum where gas costs are minimal and throughput is high. The infrastructure components built by Unibright — APIs, user interfaces — are enterprise-grade systems with years of zero-downtime operation.
Stefan added a crucial detail about the credit-based token model: UBT needs to be bridged to the L2 in advance, which takes time. The incentive structure encourages users to place credits before they’re needed rather than during high-demand periods.
Stefan: “In the moment where high transaction volume is needed, the credits should already be in place. There will not be a bottleneck at that point to feed the transactions connected to the process.”
Strategic Focus: Media First
When asked to prioritize among target sectors — ESG, AI, media — Pierre and Marten aligned on media as the natural starting point.
Pierre cautioned against over-indexing on AI hype, drawing parallels to the blockchain excitement of years past. Real use cases will emerge after the market settles.
Pierre: “AI will level out to a point where it will be much clearer what is a real use case and what is just hype. What is very clear to us is the media thing, because that is totally stable. It’s very clear what can happen.”
From a sales perspective, the Audio Group connection provides something invaluable: the ability to demonstrate working solutions rather than pitching theoretical capabilities.
Pierre: “If you want to get new technology out to enterprise customers, you should never try to sell something to them. You should give a solution to them, help them. Then it comes naturally that they want it because it gives them a direct advantage.”
Marten noted that the abundance of potential use cases — ESG, digital product passports, supply chains — means the team must choose where to focus. Media provides the clearest path while regulatory-driven use cases continue to develop.
Coming Soon: White Paper and New Website
The session concluded with a concrete announcement. Marten confirmed that a white paper and focused new website are in development, targeting release in the coming weeks.
Marten: “We will come up with a white paper and a new website — a highly focused website where we can see this funnel of all our expertise and products going into this platform. As we are already developing the platform, it’s time to come up with a white paper.”
Pierre added that the team will work to share actual usage examples and tangible front ends where possible, acknowledging the perennial B2B challenge: client work often can’t be disclosed publicly for extended periods.
Pierre: “We’ll try to show actual usage and front ends — something more tangible — as far as we can without disclosing things we’re not allowed to disclose.”
Looking Ahead
This fourth AMA demonstrated the ecosystem taking shape around proof of execution — not as an abstract concept but as infrastructure supporting real business processes with real users. The addition of Martin and Tunesat illustrates exactly the integration pattern the team has been describing: established industry expertise meeting blockchain capabilities to solve problems that neither could address alone.
Future sessions will continue exploring these themes as the white paper and platform development progress.
The next AMA is scheduled for May 2026. Questions for the next AMA can be submitted in advance via Email to ama@unibright.io or through our Telegram Communication channel (t.me/unibright_io), using Hashtag #AMA.
About Unibright: Unibright is a team of blockchain specialists, architects, developers and consultants with 20+ years of experience in business processes and integration. Unibright offers consulting with Unibright Solutions, Enterprise-grade Multi-Party-Synchronization with Unibright ONE, low-code-integration tools with the Unibright Framework, programmable DeFi with Unibright Freequity, and the Universal Business Token UBT. Unibright is actively engaged in blockchain initiatives and associations, offers Baseledger as a solution to orchestrate and trust-enhance B2B processes and offers ERP-Blockchain-integration together with a growing network of consulting and implementation partners.
About SPO: SPO, since its founding in 1989, is a strong player in the field of SAP and enterprise software consulting, development and integration, with a reputation for driving innovation and delivering comprehensive solutions to clients across a range of sectors. By integrating the latest technologies, SPO is committed to helping businesses thrive in a rapidly evolving digital landscape.
SPO has successfully implemented projects worldwide for companies of all sizes for more than 35 years.
