EUCTech 2026 · aboard the Oslo–Kiel cruise, Norway · June 3, 2026 · ~45-minute keynote, Day 1. Also published as a special solo edition of the Citrix AI Hotsheet podcast (Episode 2).
A keynote for end-user computing veterans on where the EUC industry stands in the summer of 2026 and where it’s heading. The “AI isn’t worth it” narrative has become the “AI is too expensive” narrative, and both are diffusion stories, not capability stories. The talk reframes knowledge work around the invisible 80%, walks the updated seven-step roadmap for how AI enters work, then audits every EUC primitive, translating each into its AI-era successor. The close: “The Last Chapter of EUC” is the last chapter of book one. Book two is the next 30 years, and EUC pros are the ones who get to write it.
Capabilities vs. diffusion
The narrative flipped from “is AI even good?” to “AI is too expensive,” with companies imposing usage and spend caps. That flip is itself evidence AI works. Two clocks run: capabilities, still climbing, and diffusion, hitting a wall. Most “AI ROI” complaints are diffusion problems. AI-caused congestion is real: a worker who produces more doesn’t help if the business can’t absorb the output — the bottleneck just moves, the same way developers can now write code faster than it can be reviewed, tested, and secured.
The invisible 80%
Emails, documents, and transcripts are the visible outputs of knowledge work — maybe 20%. The real bulk is invisible: thinking, reasoning, judgment, the experience that comes from years in a role. IT lives in the visible 20%; the invisible 80% is where the business-transformation consultancies live. AI has to transform all of work, so EUC has to step outside its 20% bubble. That’s why AI hasn’t produced instant ROI.
How AI enters work: the seven steps, 2026 edition
An update to the seven-stage roadmap Brian first published in June 2025. Every worker is somewhere on this path: faster search, a better Google, one-and-done prompts, is where most workers (and most AI skeptics) still are. Thinking partner is back-and-forth, uploading documents, dictating, enabled by smarter models, longer context, and document upload. Cognitive extension, the second brain, flips it: instead of bringing papers to the AI, give the AI access to everything, in a context vault it reads and writes — Brian started in January 2026, and roughly 1% of workers are here. Multi-tool agent connects the AI to the world: MCP, connectors, browser control, and computer-using agents for desktop apps — on OSWorld, median human performance is around 72–74, and AIs are now in the 80s; the point isn’t automation, it’s extending the knowledge worker’s reach across all their tools and data. Fleet of AIs means multiple AIs working and talking to each other and to other systems’ AIs. The pod is the new unit of work: a worker plus their AIs, no longer bounded to business hours, with AIs acting, flagging, and escalating on their own — three worker types emerge here: cognitive owners (context and judgment, the source of expertise), cognitive operators (who run the fleet), and cognitive curators (who maintain context and skills). The published self, an optional fork, means publishing your context vault so others can subscribe to it — brianmadden.ai and mcp.brianmadden.ai do exactly this, underpinning a future of blended expert contexts inside companies.
The current EUC model and the audit
The current model assumes one person, one screen, one set of apps, one set of hours. AI breaks those assumptions. But most of EUC transitions over rather than disappearing. VDI stays, used by humans and AI workers, grounded in the idea that AI is not eating all software: shallow UX-wrapper apps struggle, middle horizontal SaaS gets squeezed, and deep regulated systems of record (Epic, SAP, Oracle, mainframes) don’t move — AI uses these apps, it doesn’t replace the deep ones. Image management becomes skill management; app virtualization and layering become skill virtualization and layering; profile management becomes context management — the same inheritance and layering mechanics, a new atomic unit, a knowledge nugget or a skill. Group policy becomes agent policy. Session recording becomes agent observability, and beyond that, cognitive observability — a new security vector, since layered context vaults can go rogue or be poisoned, so you need to trace what context shaped an AI’s judgment, not just govern the agent. Performance management becomes token management: token routing and efficiency mean the same task can run at 200K tokens (a computer-using agent driving Excel), 100K (browser automation), 10K (reading the .xlsx XML directly), 5K (a Python script), 2K (reasoning in context), or zero (handing it to a human) — plus where the model runs (cloud, on-prem, device) and device-trust and PII checks. Endpoint management stays, but endpoints become wearables and ambient devices, with the AI generating the UI on demand wherever you are. The Citrix receiver becomes the cognitive workspace: icons for tasks and knowledge sources, not apps. And the control plane stays and gets bigger — something has to manage the humans, agents, cognition, models, routing, and security. The throughline: the words change (agent, cognitive, skills, context) but how and why the work gets done doesn’t. The core IT skills transfer. The job gets bigger.
Book one and book two
“The Last Chapter of EUC” is the last chapter of book one — the past 30 years, all spent on the visible 20%. Book two is the next 30 years, on the invisible 80%: thinking, reasoning, judgment, taste. Page one of book two is step three — build your own second brain. You have to manage one before you can manage thousands. Then connect it to apps, to other AIs, and let it work without you. Book two doesn’t have a title yet, and EUC pros are the ones who get to write it.
Key formulations
“It’s still a story about ROI — but the AI is working, it’s actually working too well, and now it’s too expensive.”
“It’s not that AI can’t do it. It’s that companies can’t absorb it.”
“Emails and documents and transcripts aren’t the knowledge work itself. They’re the outputs of knowledge work.”
“You have to manage one before you can manage thousands.”
“This is the last chapter — but it’s the last chapter of book one. We are the ones who are going to write book two.”