Industry executives and experts share their predictions for 2026. Read them in this 18th annual VMblog.com series exclusive.
By Brian Madden, technology officer and futurist at Citrix
If you listen to vendor marketing or conference keynotes, you’d think AI will be fully embedded in every job, workflow, and decision by the end of the decade. But that’s not how technology adoption actually works inside real organizations. As we look towards 2026, the more realistic story isn’t “AI everywhere,” but uneven AI adoption. Small groups of workers will fundamentally change how they work, while most others continue experimenting in the margins.
In this article, I’ll share the six predictions for AI in the enterprise I’m betting my year on. Taken together, you’ll see that I’m not describing a sudden step-change or a single breakthrough moment. There are no big bombshells here. Instead, they extend trends we’re already starting to see today: worker-led AI adoption, uneven usage across teams, and companies slowly realizing that their existing policies and controls aren’t enough. The real shift over the next couple of years is that these patterns will become so visible (and uncomfortable!) that organizations will no longer be able to ignore them.
1. AI will be a real coworker for a minority of workers but barely present for everyone else.
By the end of 2026, we still won’t see “AI everywhere” when it comes to the enterprise. What we will see is a widening gap. A relatively small group of workers will be using AI all day, every day, quietly getting 20-50% more done. But most people will still be dabbling in the early stages of their AI journeys-using AI a few times a week for emails, summaries, or brainstorming-but not fundamentally restructuring how they work.
As the year goes on, organizations will start to recognize that this gap exists, but most will end the year still being unsure how to address it.
2. Worker-led AI will still outpace corporate AI programs.
Real transformation will continue thanks to workers bringing their own tools (ChatGPT, Claude, Gemini, etc.) into their workflows, not from top-down “AI initiatives.” Sure, companies will spend heavily on pilots, platforms, and frameworks, but most day-to-day productivity gains will come from individual employees quietly wiring AI into everything they do.
The pattern mirrors what we saw with the consumerization of IT in the 2010s (smartphones and SaaS in that era) where workers move first, and IT tries to catch up.
3. The “AI skills” conversation inside companies will become less abstract and more personal.
AI will stop being just a board-level or strategy-deck topic and start appearing in everyday performance conversations. Managers will ask practical questions: Why does one person get twice as much done using AI while others don’t? What do we now expect a “competent” knowledge worker to be able to do with AI?
This doesn’t mean mass AI-driven layoffs, but it does mean that AI fluency becomes part of what “good at your job” looks like in many roles.
4. Companies will finally admit that “don’t paste sensitive data into ChatGPT” is not an AI strategy.
Ever since ChatGPT was released in late 2022, most companies have treated it primarily as a data loss prevention problem. They create generic policies about confidentiality, acceptable use, and risk, but by 2027, it will be clear that this approach isn’t sufficient.
This coming year, we’ll see more explicit AI usage policies, targeted training, and “safe spaces” where employees can use AI with corporate data in approved ways. This won’t be clean or universal, but it will mark a shift from “ignore or block” to “enable with guardrails,” as companies finally acknowledge that AI use is already happening regardless of policy.
5. Workers will juggle multiple AIs every day, and the dream of “standardizing on one AI” will mostly fail.
A typical knowledge worker today might use a corporate-approved copilot, a personal AI account they trust, AI embedded in their CRM or office suite, and additional AI features baked into individual SaaS tools. Companies will talk about “our AI,” but in reality, workers will stitch together three to five different AI platforms depending on what they’re doing.
Smarter organizations will stop trying to force everyone onto a single model and instead focus on where and how AI is allowed to operate-across specific workspaces, applications, and data-regardless of whose AI it is.
6. “Citizen developer” stories will be real but still concentrated in a small group of power users.
Throughout 2026, we’ll certainly see more genuine examples of non-developers leverage AI to build automations and mini-apps. But as I’ve written about previously, these apps won’t replace core enterprise systems, and won’t mean every worker becomes a developer. Instead, every department will have a handful of AI-powered superheroes who quietly become critical to how teams operate.
The Year AI Stops Being New and Starts Being Real
When 2026 closes and turns over to 2027, AI won’t feel new to enterprises anymore. But it also won’t feel neatly solved. Sure, a small subset of workers will be dramatically more effective, powered by the multiple AI tools they’ve stitched into their daily routines. But most companies will unfortunately miss this, instead trying to standardize, control, and/or formally “roll out” AI from the top down.
The biggest shift over 2026 won’t be technological; it will be organizational. Expectations around productivity, skills, and trust will change faster than policies and platforms can keep up. The companies that navigate this well won’t be the ones chasing a single “enterprise AI strategy,” but the ones who accept the messy reality I’ve described here-that workers moved first, that the gap is widening, and that “don’t paste sensitive data” was never a strategy. Companies that focus on enabling people to work effectively and safely, regardless of which AI they’re using, will be in great shape moving into 2027.
So, my bombshell prediction for 2026 is that there will be no bombshells. Just reality catching up.
##
ABOUT THE AUTHOR
Brian Madden is a technology officer and futurist at Citrix, where he explores how AI is reshaping knowledge work. He writes regularly at citrix.com/blogs, discusses workplace AI on LinkedIn, and gives talks globally on the future of work. Brian has spent the past 31 years focusing on end-user computing, digital workplace trends, and the evolving nature of knowledge work. Along the way he’s started and invested in companies, written six books, published more than 2,000 articles, and delivered over 1,000 talks around the world.
