On-demand keynote session

The Operating Architecture for Accountable AI

AI governance cannot stop at policies. How do you govern the work agents actually do?

As AI agents move beyond assistance to taking action, organizations need an operating architecture that brings governance into every step, from intent to outcome.

 

In this Governed AI Summit 2026 session, Shahir Daya explores human approval gates, auditability, proof of work and cost controls. He also shares what Zafin learned as Customer Zero, alongside a real banking implementation.

ShahirDaya

Featured speaker

Shahir Daya

Chief Technology Officer & Head of AIOS
Zafin

What you’ll learn

Close the AI
governance gap

Understand why deploying agents is not enough. Learn what changes when agents can access tools, make decisions and take action.

Govern work
from intent to outcome

Reimagine workflows with the right human approval gates, permissions and controls built into execution.

Make every action
accountable

See how audit trails, workflow versioning and proof of work help reconstruct what an agent did and who approved it.

Learn from
real-world execution

Hear how Zafin used AIOS internally and how a bank converted a mobile workflow in three hours with two human approval gates.