The next phase of enterprise artificial intelligence may depend less on increasingly capable models and more on everything surrounding them. Businesses need ways to access proprietary systems, govern AI deployments, process information and integrate new capabilities into established workflows.
The San Francisco Tribune reveals list of top startups working on those practical challenges as HumanX Amsterdam approaches. Taking place September 22–24 at RAI Amsterdam, the event will bring together more than 2,500 attendees, including startups, investors, enterprise buyers and technology leaders, with 60% of attendees at VP level or above.
The Systems Beneath AI
Enterprise AI often runs into an obstacle before the model even enters the picture: important business information remains locked inside existing systems.
Unfold is developing technology around that problem. Still operating in stealth, the company says it can make closed, undocumented or vendor-locked systems readable without requiring APIs, documentation or vendor participation. Its technology is designed to run within a customer’s environment, with read-only access by default and full auditing.
Kensho Technologies is tackling the data side of the equation. As S&P Global’s innovation engine, it combines AI and engineering capabilities with financial and business data. Its technology connects large language models, agents and AI applications with trusted S&P Global information while preparing proprietary data for machine learning and generative AI workflows.
These approaches highlight a broader requirement for enterprise AI: models need access to relevant information and business context if they are going to participate in real workflows.
Making AI Operational
Reliability and governance become increasingly important once AI moves from experimentation into production.
causaLens is building Digital Knowledge Workers around multi-agent architectures to automate repetitive knowledge-work processes. Its platform includes pre-built Blueprints, a Factory for creating customized workers and a governance layer called the System of Work. The company emphasizes causal reasoning, human oversight, automated evaluation, monitoring and self-healing.
Trendium is focused on the controls surrounding AI adoption. Its ContextGuard platform provides visibility and control over AI infrastructure, while its AI Enablement Program is designed to help engineering teams use AI coding agents responsibly. Its framework covers visibility, control, enablement and compliance.
The Expanding Voice Layer
Spoken language is becoming another important interface for AI applications.
AssemblyAI provides speech-to-text and voice AI infrastructure through APIs covering transcription, contextual understanding and real-time agentic workflows. The technology allows developers to incorporate voice capabilities without building speech infrastructure from scratch.
Speechmatics is focused on multilingual speech recognition, supporting more than 55 languages as well as cloud, on-premises and device deployment. Its technology is used across areas including healthcare, legal transcription, contact centers, meetings, live captioning and AI voice agents.
Otter AI is taking a broader approach to workplace conversations. Its platform captures and transcribes meetings, identifies decisions and action items, and makes conversations searchable. Users can query accumulated meeting knowledge, generate follow-ups and connect information to workflows such as CRM systems.
AI Applied Beyond Software
The companies represented around HumanX Amsterdam also demonstrate how AI is moving into specialized applications.
Nuritas applies AI to peptide discovery through its Nuritas Magnifier platform. The company says it has identified more than 8 million peptides and created a library of peptides with known functionalities, using its technology to move discoveries from computational prediction through clinical validation for health and nutrition applications.
PhotoRoom is applying AI to visual commerce. Its platform supports batch editing, automated quality assurance, brand controls and integrations with PIM, DAM and other commerce systems. It serves businesses ranging from individual sellers to large enterprises managing millions of product images.
The Infrastructure Opportunity
Taken together, these companies point toward an AI market increasingly concerned with implementation. The challenge is no longer limited to what models can generate; it also involves how AI accesses information, operates within business environments and becomes part of established processes.
HumanX Amsterdam provides a window into that transition, bringing startups and enterprise stakeholders together around the technologies that could shape how AI is deployed beyond experimentation.
