Swiss Re · EC One · Redmond · 16 Jun 2026

Built to scale.

You named the AI paradox before most of the industry would admit it: adoption everywhere, impact still out of reach. And you already proved the way out. The MIM agent takes real time out of every major incident, and 97% of your people work with Copilot every day. The next move isn't more pilots. It's taking what works and scaling it across underwriting, claims, and contracts, on a platform you can see, govern, and stand behind, end to end.

Partner
Swiss Re
Moment
EC One · Redmond
Audience
Senior IT leadership
Frame
Governed scale, near-term
Scroll
00the four points · the thinking under the pitch
Problem
The human problem behind the business problem.

World-class AI adoption, a hero agent that works, and still the value won't scale. The people accountable for risk can't put their name on what they can't fully see.

97% Copilot adoption, a proven MIM agent, and the bottom-line return still just out of reach. The AI paradox, lived day to day.

Insight
An unspoken human truth that sheds new light on the problem.

A company that prices risk for a living will not run at scale what it cannot govern. For Swiss Re, trust isn't a feature. It's the product.

The brake on agentic AI here isn't ambition or talent. It's confidence: the ability to see, control, and answer for every agent.

Advantage
What makes our thing unique and motivating.

The platform you already trust can be the governed engine for agents: one identity, one security model, one audit trail, across Copilot, Copilot Studio, and even Palantir and ChatGPT.

Already deployed, already adopted. Copilot Studio plus Agent 365 give you the one thing a point solution can't: end-to-end visibility over the whole estate.

Strategy
A new way of seeing the thing, based on all of it.

Scale what works, governed. Make governance the accelerator, not the brake.

Take the agent that already proved itself, template it across core processes, and run it all on controls a risk company can stand behind. Built to Lead becomes built to scale.

“Every company thinks they can do everything, and then they realize the world doesn’t need them to. The world wants them to do the one thing.”
Satya Nadella · CEO, Microsoft

“We see AI as an Iron Man suit for our experts, not a Terminator that replaces them.”

Swiss Re · Transforming the insurance value chain with AI
Your one thing

Carrying the world's hardest risks so everyone else can move forward. The why behind every idea here: turn AI from a tool your people use into judgment your experts can trust, governed to a standard a risk company can stand behind.

01the opening · the AI paradox, named by you

Adoption is everywhere.
The impact still isn't.

Ninety-seven percent of your people work with Copilot. You have a hero agent, MIM, that takes real time out of every major incident and earned the AI Tour stage in Zurich. And still, like every large enterprise, the bottom-line return on AI stays just out of reach. You named this the AI paradox before most of the industry would say it out loud.

The way out isn't another pilot. It's taking the one thing that already works and making it ten, without losing sight of a single agent along the way.

02the moment · why now

There may never be
a better time to scale.

Built to Lead puts AI at the center of Swiss Re's next decade. You're moving from standalone pilots into the third wave, agentic AI inside underwriting, claims, and contracts. The proof exists, the platform is deployed, the people are ready. The one thing missing is the backbone to scale it safely.

Proven
MIM

A hero agent that already works.

Your Major Incident Management agent measurably cut time-to-resolution and earned the AI Tour stage in Zurich. The proof of value is done.

Adopted
97%

An organization already fluent in AI.

10,500 M365 Copilot licenses, 97% adoption, Azure as the default cloud. The hardest change-management work is already behind you.

The wave
Wave 3

Agentic AI, in core processes.

Built to Lead moves Swiss Re from productivity gains to agentic systems that reason and act inside underwriting, claims, and contract management.

The gap
0

No center of excellence, yet.

Agents live in silos, with no single owner, no reusable patterns, and not enough visibility. Close that gap and everything else compounds.

03the art of the possible · six governed agents, ready to scale

Six agents. One governed fabric.

Each is a different way to turn proven AI into bottom-line impact across Swiss Re's core. Every one runs on Copilot Studio, grounded in your enterprise data and your controls, and every one is visible, governed, and yours.

Idea 01Internal · the proof, scaled

Scale the MIM Agent

Your Major Incident Management agent already cut time-to-resolution and earned the Zurich stage. Don't leave it as a hero story. Make it the template: the first repeatable pattern in a library every team can build from.

The move: from one agent that works to a pattern that scales.

Proof: Wells Fargo, Copilot Studio agents now live across 100% of branches · Rabobank, one chatbot rebuilt as three governed agents
Copilot StudioAgent BuilderM365Dataverse
View the concept → 01 Incident Operations dashboard · live preview
Idea 02Core · underwriting

The Underwriter's Iron Man Suit

An agent that pulls the submission, the contract, the prior treaties and the risk data into one view, drafts the first pass, and surfaces the why, while the underwriter keeps every decision. Augment judgment; never replace it.

The move: make your best underwriters faster, not your headcount smaller.

Proof: EY, core finance work augmented, submissions from 15 minutes to 60 seconds · Dow, an agent surfacing hidden losses autonomously, in minutes not weeks
Copilot StudioFoundryDataversePower Automate
View the concept → 02 Underwriting Workbench · live preview
Idea 03Core · contract management

Contract Intelligence

Reinsurance runs on contracts. An agent that ingests treaties and submissions, extracts obligations, compares versions, and flags the anomaly a tired reader would miss, with a citation behind every claim it makes.

The move: turn your most document-heavy manual process into minutes, with an audit trail.

Proof: Teck, customs-document extraction with 99% fewer errors · Dunaway, regulatory document review cut by 90%
Copilot StudioAI BuilderFoundrySharePoint
View the concept → 03 Contract checker · split-screen · live preview
Idea 04Core · claims

Claims at the Speed of Trust

An agent that triages incoming claims, summarizes the file, drafts the next step, and routes the edge cases to a human. Fast where it's routine, careful where it counts, logged at every step.

The move: speed where it's safe, human judgment where it matters.

Proof: EY PowerMatch, 80% of payments auto-cleared, up from 30% · Rabobank, 62% of chats resolved without a human
Copilot StudioPower AutomateDataverseFoundry
View the concept → 04 Solution architecture · the EY PowerMatch pattern · live preview
Idea 05Scale · the engine

The Pattern Library

Pravina asked for reusable design patterns. This is them: a center of excellence with templates, guardrails, and environment routing, so an agent proven in Zurich becomes ten across the business, without rebuilding from scratch or relaxing a single control.

The move: build once, govern once, deploy everywhere. Weeks, not quarters.

Proof: Heineken, 8,000+ environments run by a team of five · HSBC, a lending build live in 11 markets in 90 days · ZF Group, 25,000+ apps governed
Power PlatformManaged EnvironmentsALMDLP
View the concept → 05 Pattern library · Center of Excellence · live preview
Idea 06Governance · the trust spine

See Every Agent

Who built it, when, who it's shared with, how often it's used, for every agent, across Copilot Studio, M365, and even Palantir and ChatGPT. The four answers you asked us for, plus the single pane that finally lets you take the brakes off.

The move: govern the whole estate end to end, the one thing no point-vendor can give you.

Proof: NASA, 800+ apps across 300+ managed environments at zero margin for error · HSBC, secure low-code across 50+ regulated markets
Agent 365Managed EnvironmentsPurviewEntra
View the concept → 06 Agent 365 Control Plane · live preview
04proof · the company you'd keep

Regulated leaders already scale governed agents on Microsoft.

Pulled from the FY26 customer-story depot. Headline numbers verified against the public Microsoft customer stories where one exists; Wells Fargo and NASA are from the internal depot.

Wells Fargo · banking at scale
100%

Agents across every branch

One of America's largest banks runs Copilot Studio agents across 100% of its branches, automating operations with governance, compliance, and visibility built in. $4.2M in annual savings.

Microsoft customer story · FY26 depot
Rabobank · regulated bank
3 agents

Conversational banking, governed

A highly regulated bank moved its chatbots to Copilot Studio as three governed agents, one voice and two text, handling tens of thousands of calls and chats while keeping stringent controls.

Read the story →
EY · core finance process
60s

Submissions, from 15 minutes to one

EY's global finance org rebuilt SAP-integrated processes on Power Platform and Copilot Studio: 95% lower general-ledger lead times and 120K hours saved a year.

Read the story →
HSBC · regulated build
50+

Secure low-code in a global bank

HSBC innovates across 50+ markets on Power Platform with customer-managed keys and lockbox controls, moving fast while meeting strict regulatory requirements.

Read the story →
NASA · zero margin for error
300+

Governed at the highest stakes

An organization with some of the strictest risk constraints in the world runs 800+ custom apps across 300+ managed environments, $3.5M saved, secure by design.

Microsoft customer story · FY26 depot
HEINEKEN · governance at scale
5 people

8,000 environments, one small team

Managed Environments and a center-of-excellence model let a five-person team govern 8,000+ environments, 10,000 apps, and 3.1M hours of productivity.

Read the story →
05expected outcomes · what a close partnership compounds to

If you scale this the way
top-tier enterprises already have.

Not a forecast: a direction. Each agent is a step; together they turn high AI adoption into impact you can measure, governed to a standard you can defend. This is what your next chapter of Built to Lead can look like.

Outcome 01

Adoption finally becomes impact.

AI shows up on the bottom line, the way EY turned a 15-minute submission into 60 seconds and Dow let an agent surface millions in hidden value in minutes. Inside core processes, not just inboxes. The paradox, resolved.

Outcome 02

An estate you can answer for.

End-to-end visibility and governance across every platform: the model that lets a five-person team at Heineken run 8,000 environments and lets NASA operate at zero margin for error. Audit-ready, board-ready, regulator-ready, including the agents you don't build yourself.

Outcome 03

Weeks, not quarters.

Reusable patterns and a center of excellence turn one proven agent into a portfolio, the way HSBC stood up a lending build across 11 markets in 90 days. Time-to-production measured in weeks, with the controls baked in.

Outcome 04

Experts, amplified.

Underwriters, claims teams, and operations with an Iron Man suit. Humans in the loop, judgment multiplied, headcount respected.

You were built to lead.
Now scale what works, governed.
Built to Lead → built to scale.
06appendix · the flow, mapped

Paradox → Stakes → Agents → Outcomes.

The spine of the page, and where each piece came from in the inputs.

#
Beat
The idea to land
Section
00
The Four Points
Scale what works, governed. Make governance the accelerator, not the brake.
Strategy
01
The AI paradox
Adoption everywhere (97%), a hero agent (MIM), and impact still out of reach.
Opening
02
Why now
Built to Lead, wave 3 agentic AI, proof + platform ready, missing the CoE backbone.
Why now
03
Scale the MIM Agent
Turn the hero story into the first reusable template · their own proof.
Idea 01
04
Underwriter's Iron Man Suit
Augment underwriting, keep human decision · their "Iron Man, not Terminator" frame.
Idea 02
05
Contract Intelligence
Agentic contract management · named in Built to Lead.
Idea 03
06
Claims at the Speed of Trust
Triage and route claims, audited · core process.
Idea 04
07
The Pattern Library
CoE + reusable design patterns · Pravina's explicit ask.
Idea 05
08
See Every Agent
Agent 365 + the 4 metadata fields · multi-platform governance wedge.
Idea 06
09
Proof to take home
Wells Fargo · Rabobank · EY · HSBC · NASA · HEINEKEN.
Proof
10
Expected outcomes
Impact, a governable estate, weeks not quarters, experts amplified.
Outcomes

Sources: the 16 Jun 2026 Swiss Re EC One briefing (Ryan Cunningham); the Swiss Re Executive Docket (Built to Lead, the AI paradox, Pravina Ladva's three waves and "reusable design patterns," the "Iron Man suit, not Terminator" framing, the MIM hero agent, the missing center of excellence, the Agent Management Graph API / four-metadata-field ask, and Agent 365 interest). Bridge quotes: Satya Nadella on doing "the one thing" (core competencies) plus Swiss Re's own "Iron Man suit" line. Customer proof wired in from the FY26 depot: Wells Fargo, Rabobank, EY, HSBC, NASA, HEINEKEN in the Proof section, plus Dow, Teck, Dunaway, and ZF Group in the agent cards (Wells Fargo and NASA are internal-depot, no public link). Open item: add real keyart / concept UI to each agent.