As artificial intelligence moves from experimentation into enterprise deployment, the most interesting companies are increasingly solving the less glamorous problems that determine whether AI actually works. Reliability, access to proprietary systems, governance, voice interfaces, healthcare discovery and visual commerce are becoming as important as the underlying models themselves.
That makes HumanX Amsterdam a useful snapshot of where the AI market is heading. Taking place September 22–24 at RAI Amsterdam, the event is bringing together more than 2,500 attendees, with 60% at VP level or above, alongside startups, enterprise buyers, investors and technology leaders. Against that backdrop, these companies stand out for the practical problems they are tackling.
Unfold: Closing the Gap Between AI and the Systems Enterprises Actually Run On
Still operating in stealth, Unfold is working on the layer beneath most enterprise AI programs: the systems that hold both the proprietary data and the business logic, and were never built to give either one up. The company says it can make closed, undocumented or vendor-locked systems readable without requiring an API, documentation or vendor in the loop. Its approach is designed to operate inside the customer’s environment, with read-only access by default and full auditing. The broader proposition is simple: if the business logic embedded in a system remains inaccessible, modernization cannot replace it and AI cannot reason over it.
causaLens: Making AI Work More Reliably
causaLens is focused on one of the biggest challenges facing enterprise AI: turning promising agentic demonstrations into dependable systems that can operate in production. Its Digital Knowledge Workers use multi-agent architectures to automate repetitive knowledge-work processes.
The company’s platform combines pre-built Blueprints, a Factory for assembling customized workers and a governance layer called the System of Work. Its approach emphasizes causal reasoning, human-in-the-loop controls, automated evaluation, monitoring and self-healing, with the goal of producing AI systems that can deliver measurable enterprise outcomes rather than simply impressive demos.
Kensho Technologies: AI Built Around Trusted Data
Kensho Technologies operates as S&P Global’s innovation engine, combining AI and engineering capabilities with the company’s financial and business data.
Its technology connects large language models, agents and AI applications to trusted S&P Global information while structuring proprietary data for machine learning and generative AI workflows. For enterprises operating in finance and other data-intensive industries, that combination of domain expertise, data and AI infrastructure represents a significant part of the emerging enterprise AI stack.
Nuritas: Using AI to Discover What Nature Already Knows
Nuritas applies AI to peptide discovery, using its proprietary Nuritas Magnifier platform to identify and validate potentially useful peptides from nature.
The company says it has identified more than 8 million peptides and built a large library of peptides with known functionalities. Its technology is designed to move discoveries from computational prediction through clinical validation, creating ingredients for health and nutrition applications while shortening a process that can traditionally take years.
Trendium: Helping Enterprises Adopt AI With More Control
Trendium is addressing the security and governance challenges that accompany rapid AI adoption. Its ContextGuard platform provides visibility and control over AI infrastructure, while its AI Enablement Program focuses on helping engineering teams use AI coding agents responsibly.
The company’s framework centers on four areas: visibility, control, enablement and compliance. That positioning reflects a growing enterprise requirement: AI adoption cannot simply move faster; organizations also need to understand what their AI systems are accessing and doing.
AssemblyAI: Infrastructure for Voice AI
AssemblyAI provides speech-to-text and voice AI infrastructure for developers building applications around spoken language.
Its APIs support transcription, contextual understanding and real-time agentic workflows, allowing developers to incorporate voice capabilities into products without building speech infrastructure from scratch. As voice agents become a larger part of customer service, productivity and software applications, infrastructure providers like AssemblyAI are becoming an important layer underneath those experiences.
Speechmatics: Bringing Voice AI Across Languages
Speechmatics focuses on low-latency speech recognition designed for multilingual and multi-speaker environments. Its technology supports more than 55 languages and can be deployed in the cloud, on-premises or on devices.
The company serves use cases ranging from healthcare and legal transcription to contact centers, meeting platforms, live captioning and AI voice agents. Its emphasis on accuracy, latency and deployment flexibility addresses one of the practical challenges of making voice AI useful across global markets.
Otter AI: Turning Meetings Into Searchable Knowledge
Otter AI is expanding beyond automated meeting transcription toward what it describes as a conversational knowledge engine. Its platform records and transcribes meetings, identifies decisions and action items, and makes conversations searchable.
Users can also query their accumulated meeting knowledge, generate follow-ups and connect information with workflows such as CRM systems. With desktop, mobile and browser-based capture options, Otter is positioning meeting intelligence as an ongoing organizational knowledge layer rather than a simple transcription tool.
PhotoRoom: AI Visual Infrastructure for Commerce
PhotoRoom is building an AI-powered visual production platform for e-commerce companies, from individual sellers to large enterprises managing millions of product images.
Its tools support batch editing, automated quality assurance, brand controls and integrations with PIM, DAM and other commerce systems. The company is also emphasizing product fidelity, allowing businesses to generate and adapt images while maintaining consistency with the products actually being sold.
Where Enterprise AI Goes Next
The companies appearing around HumanX Amsterdam illustrate how the AI market is expanding beyond model development itself. The next phase is increasingly about making AI reliable enough to operate, connected enough to access the right information and controlled enough for enterprises to deploy at scale.
That shift is also reflected in HumanX’s broader programming, which is designed to connect AI startups with enterprise buyers and investors through dedicated programs such as VentureConnect and SolutionBridge. For attendees, the companies worth watching may ultimately be those solving the infrastructure and workflow problems that determine what AI can actually accomplish once it leaves the demo environment.
