EPPC 2026 · Closing keynote · Bella Center Copenhagen · 29 Jun

The Power of
the Platform.

Build what matters. Add AI where it pays. Scale it with control. This keynote is the business case for doubling down on Power Platform: preserved investments, agentic capability, lower lifecycle cost, and a European maker community already proving it in production.

Event
EPPC 2026
Speaker
Leon Welicki
Audience
European makers
Stage
Closing keynote
Scroll
00the four points · the thinking under the talk

One problem. One strategy.
Everything between earns it.

Problem
The human problem behind the business problem.

The people who bet on Power Platform are quietly afraid AI just made their bet — and their craft — obsolete.

When anyone can vibe-code an app, “why the platform?” becomes the doubt nobody says out loud.

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

We confuse building a thing with owning it. What feels free to make is what bills you for years.

Creation is T0. Fast to build is not cheap to own.

Advantage
What makes your thing unique and motivating in people's minds.

Build fast with Copilot Studio. Make it last with Power Platform. The only agentic path that also governs the cost, the risk, and the maintenance — on one managed platform.

Speed and staying power on the same managed foundation. Microsoft owns the maintenance curve; what you already built carries forward.

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

You made the right bet. Now double down.

Reframe the platform from “a way to build apps” to “the economical, governed way to own agentic software at scale” — so the faithful are ahead, not behind.

Framework — Mark Pollard's “Four Points,” Sweathead.

01Open · one local proof point, then the room

Start with Heathrow.
Then widen to Europe.

Open on a story people can feel before asking them to buy a framework. Heathrow gives us a concrete European proof point: high stakes, operational scale, a maker journey, and Power Platform running in the real world.

Opening video · Heathrow Airport

Problem / Stakes / Solution / Outcome

A deeper customer film can establish the keynote's stakes: real work, real risk, real scale, and a maker-led solution that keeps delivering. The goal is not a glossy case study; it is a grounded bridge into the economics of building.

ProblemOperational work that cannot wait for a perfect IT queue.
StakesAirport scale, safety, continuity, and trust.
SolutionA Power Platform maker/champion story with governed tools behind it.
OutcomeLong-running impact with low maintenance and room to evolve.
Why here, why now

The last keynote should feel like a celebration.

EPPC is not a generic enterprise audience. It is a builder audience, in Europe, at the end of a major community event. The structure should make them feel seen before it asks them to think about TCO, governance, and agentic strategy.

56–62M
Monthly active makers, globally
Every dept
Build moves past the IT backlog
In prod
Real solutions, running today
Posture
Celebrate European makers first. The local wins, the local names, the local case studies. We're here because of what they build.
The frame
“Everyone is a builder” isn't a slogan here — it's the room. The strategy that follows is about giving that room control.
Closing beat
Save the 2–3 minute builder montage for the end, where it can feel like a thank-you to the room.
Map of Europe · 8–10 makers · human impact, not technical depth.
Continuity from EPPC25

Last year said Power Apps is a vehicle for democratizing business innovation, and that apps still matter in the agent era. This year should not repeat that talk. It should build on it: if apps still matter, here is why the platform behind them matters economically, operationally, and ethically.

02The provocation · two ways the rush goes wrong

The fastest build
can become the most expensive one.

Before the answer, sit with the problem. Vibe-coded apps, custom agents, and raw-code shortcuts can look cheap at T0. The bill arrives later: maintenance, control, shadow IT, token sprawl, and security risk.

Failure mode 01 · the runaway bill

AI everywhere, with no hand on the dial.

Blanket AI everywhere reads as ambition. It bills like a leak. Costs that start cheap grow with scale, and nobody set the throttle.

$500K
One company's unmonitored token spend
Failure mode 02 · the new shadow IT

It's worse than the spreadsheet era.

The old shadow IT was a rogue Excel file. The new one is local apps, built with code, calling endpoints nobody vetted — a clean path for data to walk out the door.

Exfil
Unknown endpoints · unvetted code · no audit

The pitch isn't “use AI.” It's “own the curve.”

Everyone in the room can already turn AI on. The hard part — the part worth a keynote — is building in a way the organization can afford, maintain, secure, and evolve for years.

03The economics · creation is T0

Creation is T0.
Ownership is the real curve.

The Agent Academy deck proves the platform can make every app, workflow, and page agentic. EPPC needs the sharper business argument: don't optimize only for the first build. Optimize for the full life of the solution.

Power Platform · flattens to a plateau
Raw code · compounds with scale

Two curves, opposite incentives.

Custom build starts fast and cheap, then compounds as apps, dependencies, integrations, and agents multiply. Power Platform may ask for more discipline up front, but it flattens the lifecycle curve.

The crossover is the slide. Before it, a shortcut looks cheaper. After it — which is where every real organization lives — the managed platform is what keeps growth economically viable.

Cost isn't the build. It's the years after.

Total cost of ownership is cost multiplied by time across the operating life of the solution. Most decisions only price the first moment. Power Platform wins on the long arc.

T0 · build

Create

One-time cost to stand the solution up. The phase everyone estimates.

One-time spend
T1 · maintain

Keep it running

Years of operation. Microsoft updates the backend. Many solutions need zero ongoing maintenance.

Near-zero on platform
M1 · evolve

Adapt over time

No chasing React or Vite vulnerabilities. The platform abstracts the framework churn away from you.

Platform absorbs the churn
Whiteboard sketch comparing custom build cost rising over time with model or canvas build effort decreasing over time.
App building economicsVibe code may win the first week. Power Platform wins the years after launch.
Whiteboard sketch of a solution lifecycle with build, maintain, evolve, and agent moments for tokens, refactor, and audit.
Lifecycle thinkingBuild once, then maintain and evolve. Agent moments happen where they do one job well.
Whiteboard sketch of an IT backlog with projects of different sizes stacked in a queue.
Backlog pressureThe platform is not replacing IT; it expands safe capacity beyond the queue.
“We got fast software. But software, if you're going to build a real company, can't be throwaway.”
Tony Fadell · co-creator of the iPod & iPhone, founder of Nest
Build fast with Copilot Studio. Make it last with Power Platform. One managed platform — secure, compliant, and governed by default.
04The strategy · agentic, on your terms

Use AI like a scalpel,
not a fire hose.

The differentiator is control over where the tokens go. Choose AI where it earns its cost. Use plain automation everywhere else. The design decision is no longer “add AI or not” — it's “AI here, deterministic there.”

Tokens

Spend AI where language, ambiguity, or judgment earns the cost. Leave deterministic work deterministic.

Refactor

Bring in an agent for a bounded improvement job, not as a permanent source of toil.

Audit

Use agents as focused reviewers when the lifecycle calls for scrutiny, traceability, or control.

automationIntake form
AI · earns itRead the messy attachment
automationPolicy check
automationRoute & notify
AI · earns itDraft the reply
automationPost & reconcile
AI applied surgically — two steps, where judgment pays Deterministic automation — cheap, predictable, the rest of the flow

Two kinds of agents. Know which curve they touch.

B

Agents that build

They help the maker create the solution — generating apps, flows, logic, pages, schemas, and tests. The cost sits at design time, with a human watching.

R

Agents that run

They live inside the solution at consumption time, acting for the user in production. This is the spend that scales, so this is the one to supervise and govern.

Fastest path for existing customers

Add agents without a rewrite.

Your existing Power Platform solutions don't get thrown away to go agentic. Configuration switches turn on in-app agents in what you've already shipped. The investment is preserved; the AI is additive.

In-app agent Existing solution, untouched Token budget, set by you
05The answer · governed innovation

Make every build safe.
Then make every solution agentic.

This is where the Agent Academy flow comes back in. The platform gives makers a full toolset, brings agents into existing solutions, and keeps governance, security, operations, and availability in the managed layer.

Ungoverned · the new shadow IT

Innovation you can't see.

  • ✕Local apps calling endpoints nobody approved
  • ✕Data exfiltration with no audit trail
  • ✕Token spend invisible until the invoice lands
  • ✕Framework vulnerabilities you now own
  • ✕Security finds out last
Governed · the same energy, contained

Innovation you can administer.

  • ✓Full admin control over what users can do
  • ✓Enterprise guardrails on every environment
  • ✓DLP, audit, and visibility built in
  • ✓A safe place to build, by default
  • ✓Security is in the loop from day one

The bet isn't “lock it down” or “let it run.” It's both: a workforce free to build, on a platform that keeps it safe, visible, and economically viable.

06The payoff · the builder economy

Productivity doesn't
shrink the team.

The fear in the room is that “everyone's a builder” means “fewer people.” Name it and flip it. The gain isn't fewer heads; it is the same people able to solve more business problems than the backlog ever allowed.

The fear
More automation means a smaller workforce.
What actually happens
The same workforce becomes a workforce of builders — and the org's capacity expands.

What “everyone is a builder” unlocks.

56M+
Reach

People already building

Monthly active makers worldwide — the base this strategy stands on.

0
Backlog tax

Past the IT bottleneck

Work that used to wait in a queue gets built by the people who need it.

+1
Net effect

Capability, not cuts

Headcount holds. What each person can ship goes up — safely, on platform.

07The proof · power of the platform

Show the platform.
Then celebrate the builders.

Land the talk on proof. Every demo shows multiple products working together toward one outcome — never a feature tour. Then close as the last keynote should close: with the European makers who already live it.

Demo strategy · the rule

“Power of the Platform.”

Roughly 35 minutes, mostly demos. No single-tool feature tours — each one starts from a business outcome and shows the products combining, because that's where the impact compounds. Build fast with Copilot Studio. Make it last with Power Platform.

Apps Automate Copilot Studio Dataverse Agents → one outcome

Leon's demo lineup.

Demo 01 · rich attended automation

A clunky legacy step becomes a beautiful app — in seconds.

A desktop flow automates an old system; the Canvas MCP builds the attended app from the flow itself, brand-aligned with Copilot and WorkIQ inside.

Power AutomateCanvas MCPCopilotWorkIQ
Demo 02 · enhancing everyday work

Your everyday app, now with Copilot, agents, and WorkIQ.

A modern model-driven app answers from your data, invokes an agent, drafts a Word or PowerPoint doc, and renders app fragments inside Copilot via the App MCP.

Model-drivenCopilotCopilot StudioApp MCPDataverse MCP
Demo 03 · reaching a large audience

From a markdown spec to a live, governed site.

“Keeping it Real”: the Pages MCP builds the site from a spec, a security agent reviews quality, and a back-office app lets humans approve what ships.

Power PagesPages MCPGitHub CopilotModel-driven
Demo 04 · agents + humans, together

Your makers and agents, building on one app.

GitHub Copilot joins a Canvas app as an agent — analyzing, refactoring, improving — with Claude Code side by side. Many humans, many agents, one app.

Canvas MCPGitHub CopilotClaude CodeSkills Marketplace
Demo 05 · mobile (tentative)

Built for mobile — natively.

Power Apps with native mobile components, on the device where the work happens. Customer scenario still to be chosen.

Power Apps mobileNative components

European proof on stage.

ISV · UK
Customer film · 1:20

Provance

A software business built on the platform. Leeds City Council replaced Remedy at enterprise scale — real workloads, not one-task apps.

app ISV · candidate
Customer story

Peter Charquero

A second ISV story to show a business built on the platform, if we can land it alongside Provance.

airport
Opening film

Heathrow / Sumit

The human proof point for long-running impact, maker growth, and low-maintenance platform value.

finale · ~2:30
Closing video

European maker film

Flies across Europe — eight to ten makers, each with a phrase, their scenario, and the impact. Telefónica (Spain) added to the lineup.

You made the right bet. And the world only moves forward because people like you keep building it.
So keep building.
Now double down on the platform
08appendix · whiteboard themes, mapped

Where each idea
came from.

Remapped after the Agent Academy deck, the EPPC25 transcript, and the updated economics whiteboards. The shape below keeps continuity without repeating last year's keynote.

#
Whiteboard theme
The idea to land
Module
01
Cover
The Power of the Platform · Build what matters. Add AI where it pays. Scale it with control.
Hero
02
Heathrow opening film
Problem / Stakes / Solution / Outcome · one local story earns the strategy
Open
03
From EPPC25 to now
Last year: apps still matter. This year: the platform economics make the bet durable.
Continuity
04
The IT backlog
The platform expands safe capacity beyond the central queue
Builders
05
The AI shortcut trap
Fast custom/vibe builds create later cost, control, and shadow IT risk
Trap
06
App building economics
Custom cost curves rise; platform lifecycle cost flattens
Economics
07
Creation is T0
TCO is cost x time across build, maintain, evolve
Lifecycle
08
Agent moments
Tokens, refactor, audit: bounded jobs-to-be-done, not agent sprawl
On your terms
09
Every app/workflow/page becomes agentic
Carry forward the Agent Academy flow: inject agents into existing solutions and add agentic authoring
Platform
10
Power of the Platform demos
~35 min, 5 demos: attended automation · everyday work · large audience · agents+humans · mobile. Build fast with Copilot Studio, make it last with Power Platform.
Demos
11
Governed innovation
The new shadow IT is local apps, unknown endpoints, token leakage; platform gives visibility and control
Governed
12
Existing investments preserved
No restart. Add agents to what customers already built.
Fast path
13
European proof
Heathrow · Provance/Leeds (replaced Remedy, 1:20 film) · Peter Charquero · Telefónica
Stories
14
Closing maker celebration
~2:30 film across Europe; each maker a phrase, scenario, and impact — communal and earned
Finale
15
Final line
You made the right bet. Now double down on the platform.
Close

Sources: “Presentation: Leon for Agent Academy Live,” “EPPC25 Keynote — Democratizing Business Innovation,” “EPPC — Early whiteboard thinking_Leon,” the Sally/Leon maker celebration notes, “The Power of the Platform — Leon's notes for demos,” and Tony Fadell on Lenny's Podcast (the one borrowed quote). Open questions to resolve next: confirm demo build owners (Leon OOF 6/9–6/20), final customer permissions, and the mobile demo scenario.