The Human-in-the-Loop

Designing Hybrid AI Systems for Critical Banking and Insurance Workflows

Most financial institutions already know AI can help. The harder question is how to design systems where models and people work together – where automation accelerates decisions without eliminating judgment, and where governance exists by design, not as an afterthought

This session brings together three practitioners who’ve built these systems at scale and are now applying those lessons to help banks and insurers ship AI that regulators will approve.

Your Benefits

Four questions we'll tackle

How do banks detect the dual threats of hard and soft attrition?

Not every decision needs a human checkpoint. We’ll look at real examples from underwriting, collections, and fraud to show where intervention adds value – and where it just slows things down.

How do you design for explainability without killing speed?

Regulators want transparency. Business units want fast answers. We’ll discuss architecture choices that satisfy both, including what documentation actually gets reviewed in audits versus what just sits in folders.

What does "governed AI" mean in practice?

Beyond the buzzwords: maker-checker workflows, model lineage, approval trails, and risk monitoring. What works in production, what’s theater, and where most institutions get stuck.

How do you move from pilot to production without a three-year roadmap?

The gap between “proof of concept” and “running live” kills most AI projects. We’ll share patterns from institutions that shipped, not just tested.

Senior Speakers

Three people who've spent their careers building AI systems that actually ship

Suman Kumar Singh

Founder & CEO, iTuring

Milind Joshi

VP of Payment Solutions, Striim

Valsan Ponnachath

Co-Founder, President & COO, iTuring

Moderated By

Bemnet Tesfaye

Strategic Partnerships at iTuring Former U.S. Diplomat

Who is this for?

This conversation assumes you're past "should we use AI?" and onto "how do we do this right?"

If you’re a Chief Risk Officer or Model Risk lead trying to balance innovation with audit requirements, you’ll hear governance frameworks that actually get used.

If you run Collections, Underwriting, or Claims and need to lift performance without losing oversight, you’ll see working examples of hybrid systems.

If you’re leading AI/ML teams and tired of pilots that never launch, you’ll get architectural patterns that clear compliance hurdles.

And if you’re in Legal, Audit, or Compliance and need to understand how explainability works when humans and models collaborate, you’ll get straight answers.

Frequently Asked Questions