ISO/IEC 42001 Certification: AI without losing control

Gartner sayis 40% of enterprise apps will have AI agents by end of 2026.

Deloitte says only 11% of organizations have them in production today.

That gap is where things go wrong.

Companies are rushing to ship agentic AI. 

Fewer are asking who governs it once it's live – routing queries, accessing proprietary data, making decisions on behalf of users. 

We asked that question before we shipped.

Last October, Hotovo achieved ISO/IEC 42001 certification, among fewer than 50 organizations worldwide at the time. Not as a checkbox. Because we were already building AI agents for clients like g-Xperts and Energy intelligence provider, and we needed the governance to match the capability.

ISO/IEC 42001 is the first international standard providing requirements for an Artificial Intelligence Management System (AIMS). It helps organizations develop and deploy AI systems responsibly, ethically, and safely by managing risks, ensuring transparency, supporting AI risk management, and guiding responsible AI practices.

What that looks like in practice:

→ An agentic AI assistant for energy intelligence provider that routes each query to the right module across research, time-series data, and API docs

→ An AI copilot for g-Xperts helping insurance underwriters evaluate property and vehicle risks

→ Both deployed under a certified AI management framework

The EU AI Act high-risk deadline hits August 2, 2026. Five months from now.

If you're outsourcing AI development, ask your partner one question: 

How do you govern what you build?

The answer matters more than the product demo.