EU AI ACTto the 2 August 2026 high-risk enforcement deadline.Check your tier →
Home · Webinars · Webinar 3
Live webinar · Wednesday, 9 September 2026 · 14:00 CET

Agentic AI in regulated industries — production patterns that ship

Agentic AI moved from research to production in 2025-26, and the customer conversations have shifted with it. Two years ago every CRO question about agentic AI ended with "can we audit what the agent did?" — in 2026 the answer is unambiguously yes, but only for agent runtimes designed from day one with audit as a first-class concern. This webinar covers the production patterns we've seen survive supervisor review across 17 regulated-industry engagements: bounded ReAct, planner-executor, multi-agent orchestration with hand-off contracts, and the four engineering controls that separate prototypes from production.

Duration
60 minutes
Audience
Architects, CIOs, CROs
Cost
Free
Register now
Save your seat
Calendar invite within 24h. Recording sent to all registrants. No follow-up sales sequence. Unsubscribe any time.

You walk away with

  • Three production patterns: bounded ReAct, planner-executor, multi-agent
  • Three failure patterns that fail supervisor review: silent loops, hidden tool calls, unbounded reasoning
  • The four engineering controls that separate prototypes from production
  • Sector breakdowns: BFSI, healthcare, public sector, pharma
  • Three-tier evaluation harness: deterministic, trajectory, safety
  • Live Q&A on agent runtime engineering

Agenda

0:00 – 4:00What "agentic" actually means in regulated production
The narrow definition we use. What it excludes. The framing supervisors increasingly expect.
4:00 – 18:00The three production patterns
Bounded ReAct with reasoning-trace persistence. Planner-executor with explicit plan persistence. Multi-agent orchestration with hand-off contracts.
18:00 – 28:00The three failure patterns
Silent loops, hidden tool calls, unbounded reasoning. How they fail supervisor review and how to prevent them.
28:00 – 38:00The four engineering controls
Hashed tool registry, content-addressed reasoning-trace storage, budget enforcement, separate verification pass.
38:00 – 48:00Sector breakdowns
Patterns we ship into BFSI, healthcare, public sector. Where the supervisory expectations differ.
48:00 – 60:00Live Q&A
Open mic for architecture and engineering questions.

Speakers

Saurabh Goenka
Founder & CEO · MindMap Digital
Has shipped agent runtimes into Saudi insurance, UK NHS, Canadian federal, Indian pharma R&D — each with different supervisory expectations.
MindMap Engineering
Engineering Practice Leads · MindMap Digital
The team behind the 17 regulated-industry agentic deployments referenced in the State of Agentic AI 2026 research piece.

Pre-reading

If you want to come in warmed up, the pillar page covers the substance from a different angle.

Read /agentic-ai
Talk to the product team