The companies getting the most attention in enterprise AI right now aren’t the ones with the flashiest demo. They’re the ones solving problems nobody wants to think about: whether a system can actually read the data it needs, whether an output can be trusted, whether a voice agent understands what’s being said. That’s the thesis worth carrying into HumanX Amsterdam, covered recently by The San Francisco Tribune as part of the broader HumanX Europe series, running September 22 through 24 at RAI Amsterdam.
The Thesis
Enterprise AI matured the moment it stopped being about what a model could produce and started being about what could actually be deployed. Nine companies at this year’s event make that case better than any keynote could.
The Evidence
Start with Unfold, still operating in stealth, building a way to make closed, undocumented, or vendor-locked enterprise systems readable, without an API, without documentation, and without needing the vendor involved, all while running inside the customer’s own environment with read-only access by default and full auditing. The argument underneath it is simple and correct: business logic that stays locked away can’t be modernized, and it can’t be reasoned over by AI. Kensho Technologies proves the same point from inside a single industry, serving as S&P Global’s innovation engine and connecting large language models and AI agents to trusted financial and business data, while structuring that proprietary data for machine learning and generative AI use.
causaLens makes the case for reliability specifically, closing the distance between an agentic demo and a system that survives production through its Digital Knowledge Workers, built on pre-built Blueprints, a customizable Factory, and a governance layer called the System of Work, with causal reasoning and human-in-the-loop controls running throughout. Trendium makes the case for governance, giving enterprises visibility and control over their AI infrastructure through its ContextGuard platform and helping engineering teams adopt AI coding agents responsibly through its AI Enablement Program, organized around visibility, control, enablement, and compliance.
Voice AI supplies three more data points. AssemblyAI is the infrastructure layer, offering speech-to-text and voice APIs for transcription, contextual understanding, and real-time agentic workflows. Speechmatics is the deployment layer, supporting more than 55 languages with recognition that works in the cloud, on-premises, or on a device, serving use cases from healthcare transcription to live captioning. Otter AI is the furthest along, having moved past transcription into a conversational knowledge engine that identifies decisions and action items and turns meetings into searchable, queryable knowledge connected to workflows like CRM systems.
The final two companies close the case in fields outside the usual enterprise-software conversation entirely. Nuritas is applying AI to peptide discovery through its proprietary Nuritas Magnifier platform, having identified more than 8 million peptides and built a library of known functionalities, moving discoveries from computational prediction through clinical validation far faster than the traditional process allows. PhotoRoom is doing the equivalent for e-commerce, building AI visual infrastructure with batch editing, automated quality assurance, and brand controls, anchored by a firm commitment to product fidelity.
The Takeaway
None of these nine companies will get the loudest applause at HumanX Amsterdam. But each one is solving a problem that decides whether AI actually works once it leaves the demo environment, and HumanX’s own VentureConnect and SolutionBridge programs exist specifically to put that kind of work in front of the people who need it most.
