Our vision for AI and the future of work

Citrix AI Think Tank · Paris, France · December 4, 2025

Presentation with slides. No video recording available. This talk followed the pre-November 2025 stump speech structure; the March 2026 DUCUG talk replaced it entirely after the model inflection made this version out of date.

Slides

Download slides (PDF)

Key frameworks

The talk

The AI disappointment paradox

Opened with the disconnect between AI hype and enterprise results, drawing on NYT, Fortune, and MIT reporting. "AI means lots of things" as a framing device: the conversation is fractured because everyone means something different. Meanwhile, workers are already using AI tools on their own (the shadow strategy reality), regardless of what the company has sanctioned.

Introduced the 7-stage evolution model as a way to understand where we are: moving from Phase 1 (AI as assistant) to Phase 2 (AI as colleague).

How AI accesses applications

Walked through the four pathways AI uses to interact with work today.

Connectors for modern apps: APIs, token-based authentication, purpose-built integrations. Browser control (the Operator model): AI driving a web browser the same way a human would. Computer Using Agents for legacy desktop applications: the OSWorld benchmark showed progress from 12% to 72% human-level performance. And direct file manipulation: Claude working with documents on the file system.

Tied these together through the four stages of post-app realization: the gradual recognition that applications as we know them are dissolving.

The workspace model

The evolution from helpers to colleagues to autonomous agents, and what infrastructure each phase requires. Workspace components: Apps and Data, Identity, Security, Context. The key insight that worker diversity makes top-down management impossible: every knowledge worker uses AI differently, and the workspace needs to accommodate that variation rather than enforce uniformity.

The Human-AI-Workspace interaction model: the workspace sits between the worker and their AI, managing security, governance, and context. UI flexibility becomes critical as workers access the same applications through different interfaces (chat, browser, voice, API).

The historical parallel

The horse and buggy transition image as metaphor. Then the specific 1990s parallel: everyone predicted the web would kill Windows apps. The web's promise was real (access anywhere, any device, better security). The reality was that migration took decades, not years.

Citrix's historical role: rather than waiting for the migration to complete, Citrix made existing applications accessible through the new paradigm. The same pattern applies now. "Keep doing what already works, with AI layered on top."

Alternate universe IT

Reframing shadow IT as "alternate universe IT." When workers build their own AI tools and workflows, that's not a threat to manage; it's innovation to learn from. Microsoft's "vibe working" concept referenced.

The management philosophy shifts from "manage the apps" to "manage the environment." Control the workspace, not the individual tools. This is where Citrix's value proposition evolves: from application delivery to workspace governance.