Closing the gap between your AI eyes and stomach at HumanX
Fresh off my jaunt to AGNTCon + MCPCon Europe last week, I headed back to Amsterdam for the first European installment of the artificial intelligence-focused HumanX conference so I could meet with another round of AI startups. As always, my goal was to uncover the untold stories of innovation across the software landscape – and […] The post Closing the gap between your AI eyes and stomach at…
The HumanX conference in Amsterdam 2026 brought together artificial intelligence startups showcasing innovative solutions to address the gap between the desire for AI and the ability to successfully deploy them in production environments. Through a variety of approaches, these ten vendors aim to help their customers close this gap.
8wave Oy provides a platform that tracks the lineage of the data feeding AI applications, connecting AI activities to KPIs important to the business. This enterprise resource planning (ERP) for AI operations enables organizations to better understand and manage their AI initiatives.
Bosun B.V. leverages AI to automate software maintenance across legacy codebases, ensuring security fixes, dependency updates and legacy migration activities without risking the behavior of the original applications. The company's key innovation is the parity file, which tracks the behavior of each legacy application and allows organizations to compare the original and updated applications.
Clarifeye SAS combines AI with tribal knowledge gathered from experts within an organization to create process maps, business rules and playbooks that accurately represent business operations. Clarifeye's unique innovation is its ability to incorporate hard-to-reach tribal knowledge into AI-driven decision-making.
Distil labs GmbH focuses on simplifying the creation of small, specialized language models (SLMs) that can be trained with corporate and synthetic data to deliver tailored solutions for specific business requirements. Their key innovation is an automated, iterative approach to creating SLMs, enabling businesses to lower token costs compared to large language models (LLMs).
Reasoning Engines B.V., operating as LangWatch, offers simulation-based AI agent testing and evaluation. The service uses agents to simulate human behavior in testing applications, iteratively writing and improving tests based on their output. LangWatch's key innovation is its early adoption of the Jev AI model, which promises to accelerate agentic evaluation and LLM-as-a-Judge capabilities.
Orq.ai Holding B.V. provides an agentic AI development and deployment platform that combines centralized governance with decentralized, domain-specific delivery capabilities. The centralized platform serves as an agentic center of excellence, offering organization-wide governance, policies, and audit trails. Meanwhile, individual teams can leverage the platform to build agentic applications while adhering to the guidelines set by the CoE.
Orq.ai's innovation lies in using platform engineering best practices to balance centralized governance with diverse business needs.
Safe Intelligence Ltd. offers a platform for testing agentic applications to eliminate failure conditions, improving their resilience. The platform runs through numerous test cases, identifies failures and their causes, and provides recommendations for fixes to enhance application resilience in subsequent iterations. Safe Intelligence's key innovation is its ability to identify various causes of failures, such as underlying models, governance guardrails, and business logic issues.
Softr Platforms GmbH offers a no-code platform for creating portals, internal tools, and other applications, providing enterprise-class back-end capabilities including scalable deployment, security, access control and policy enforcement. While not exclusively focused on AI, Softr's visual development environment complements the growing demand for no-code solutions in the generative AI landscape.
Written by urgent.news from SiliconANGLE's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.