Sep 21, 2026

The San Francisco Tribune Compares the AI Startups Competing at HumanX Amsterdam

Not every company attending HumanX Amsterdam is solving the same problem, but grouped correctly, the nine names The San Francisco Tribune recently highlighted fall into a handful of clear categories. The event runs September 22 through 24 at RAI Amsterdam, part of the HumanX Europe series, and comparing these companies side by side says a lot about where enterprise AI is actually headed.

The Access Layer: Unfold and Kensho Technologies

Both companies solve a version of the same problem: getting AI systems connected to data that was never designed to be shared easily. Unfold, still in stealth, works on enterprise systems that are closed, undocumented, or vendor-locked, making them readable without an API, documentation, or vendor cooperation, all while running inside the customer’s own environment with read-only access by default and full auditing. Kensho Technologies takes a more vertical approach, serving as S&P Global’s innovation engine and connecting large language models and AI agents to trusted financial and business data, while also structuring that proprietary data for machine learning and generative AI use. Where Unfold is building a general-purpose access layer, Kensho has built a deep, industry-specific one.

The Reliability Layer: causaLens

causaLens stands somewhat on its own here, focused specifically on the gap between a working demo and a system reliable enough for production. Its Digital Knowledge Workers use multi-agent architectures to automate repetitive knowledge work, supported by pre-built Blueprints, a customizable Factory, and a governance layer called the System of Work. What sets the approach apart is its emphasis on causal reasoning alongside human-in-the-loop controls, automated evaluation, monitoring, and self-healing, all aimed at outcomes that hold up once a system leaves the pilot stage.

The Governance Layer: Trendium

Trendium occupies a related but distinct space, focused less on making AI work and more on making sure enterprises know what their AI systems are doing. Its ContextGuard platform provides visibility and control over AI infrastructure, and its AI Enablement Program helps engineering teams use AI coding agents responsibly, built around four pillars: visibility, control, enablement, and compliance.

The Voice Trio: AssemblyAI, Speechmatics, and Otter AI

These three companies compete in the same broad category but occupy different niches within it. AssemblyAI is the developer-infrastructure play, offering speech-to-text and voice APIs for transcription, contextual understanding, and real-time agentic workflows. Speechmatics is the deployment-flexibility play, supporting more than 55 languages with recognition that works in the cloud, on-premises, or on a device, serving healthcare, legal, contact center, and captioning use cases. Otter AI is the outlier of the group, having moved beyond transcription into a conversational knowledge engine that identifies decisions and action items and connects meeting data to workflows like CRM systems.

The Applied Science Layer: Nuritas and PhotoRoom

The final two companies apply AI to problems well outside the typical enterprise-software conversation, and comparing them highlights just how broad this category has become. Nuritas uses its proprietary Nuritas Magnifier platform for peptide discovery, 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, with applications in health and nutrition. PhotoRoom applies a similar discipline to e-commerce imagery, offering batch editing, automated quality assurance, and brand controls for sellers of any size, with a consistent focus on keeping generated images faithful to the actual products.

Laid out this way, the pattern across all nine companies becomes clear: each is solving a distinct, unglamorous problem that determines whether AI actually works once it’s deployed, which is exactly the kind of conversation HumanX built its VentureConnect and SolutionBridge programs to support.