Happy New Year. This is my first “SNAP” of 2026, and it comes straight out of the final draft of my latest State of Workplace Wellbeing 2026 report. A key theme that keeps coming up is AI creating as much pressure as it is possibility? Not because it isn’t useful, but because it’s arriving in a world where the pace of work is already increasing. For many people, AI has been layered onto packed diaries, constant communication, and rising expectations. The result is we have a sense that we should be moving faster… even when the human effort doesn’t reduce. This is what I am calling the AI high performance paradox.
The paradox: faster outputs, same (or greater) human effort
This has raised a paradox in output expectations across many roles. Tasks can be drafted, analysed and iterated faster than before, but human effort has not (necessarily) reduced in line with that speed.
The AI high performance paradox is that while AI can save time by processing at greater speeds, this does not automatically translate into less work. In practice, it can just shift the work from producing the first version to evaluating, refining, aligning, and sense-checking multiple versions.
AI moves quickly. Humans still need time to think.
This is reflected in how many employees are feeling, a recent study published in HR Drive showed 1 in 4 feeling “often” or “always” overwhelmed with AI developments. A third also said that learning, using and checking AI takes as much time as their previous approach to work.
AI expands thinking before it reduces time
We’ve absorbed a story that AI equals speed: faster drafts, faster analysis, faster decisions. But in real working life, AI often does something different first. It expands.
It introduces:
- New perspectives you hadn’t considered
- Alternative ideas you now feel obliged to evaluate
- New wording that prompts a rewrite
- Multiple options that create yet more choices and decision fatigue
So instead of:
Task → outcome
It becomes:
Task → options → evaluate → refine → decide → sense-check → finalise
That doesn’t always speed things up, especially in roles where judgement, nuance, and accountability matter.
AI and cognitive overload
One of the most common impacts I’m seeing is decision overload. Often, “faster” creates more thinking, for example, when AI gives you five good routes, you still have to choose. You still have to judge tone, accuracy, risk, context, and audience.
So while AI can reduce the effort of starting, it can increase the effort of finishing. And at a time when the pace of work keeps rising (EY study shows 62% of UK respondents reported an increase in their workloads over the last year), that extra cognitive load can become the thing that tips people into overwhelm.
AI as a partner, not a shortcut
Here’s the shift I’m encouraging leaders to make:
See AI as a partner, not a replacement.
And (at least for now) it’s not always a speed tool.
Used well, AI becomes something closer to a:
- Virtual Assistant doing the legwork quickly
- Confidant a place to put messy thinking
- Guide helping you clarify what you want or challenge your ideas
- Mentor sharing wisdom from alternative sources
It’s someone to “think with”. That can enhance performance, but it still requires careful evaluation, judgement, and time.
There is a common misconception that agents act without control. Humans will remain the orchestrators and final decision-makers. Agents will function as powerful assistants to augment human-centric workflows.
– Albert Lai, Director, Telco, Media & Technology, Global Strategic Industries, Google Cloud.
What leaders need to stop assuming
If you lead a team, this matters: Don’t automatically assume AI will make everything faster.
It might, for repeatable admin tasks. But for knowledge work, it often increases the cognitive steps, at least until ways of working evolve around it.
The danger comes when organisations treat AI as a reason to raise expectations without adjusting capacity, effectively accelerating the pace again.
AI through a wellbeing lens
The future of work, I believe, sits at the intersection of wellbeing and high-performance because the goal isn’t wellbeing instead of results. It’s wellbeing as a core part of a sustainable high-performing culture.
So, how do we use AI to support focus, energy and sustainable high-performance, rather than adding pressure to an already fast-moving system?
A few practical starting points:
- Be explicit about the purpose. Are we using AI to reduce admin, improve quality, or generate ideas? Different purpose = different expectations.
- Reduce choice when needed. Ask for two options, not ten, or ask AI to recommend one approach and explain why.
- Build “thinking time” back in. If AI expands the decision set, people need time to evaluate, not just deliver.
- Normalise iteration. Using AI well is collaborative. It’s not “push button, receive perfect answer”.
AI can absolutely make work better. It can also make work feel heavier if we use it without intention, especially when the unspoken expectation becomes “now you can do more, faster”.
So as we head into 2026, here’s the central point:
AI can enhance outputs, but it doesn’t eliminate the need to think. And it doesn’t automatically reduce workload, even if it speeds up parts of the process.
That’s the AI high performance paradox. And it’s one leaders need to manage deliberately as the pace of work continues to increase.
Download the 2025 report (and be first to hear about 2026)
If you haven’t downloaded the 2025 State of Workplace Wellbeing report yet, you can do that now. If you’d like to be first to hear when the 2026 report lands, sign up for the newsletter (link in the footer), and you’ll be the first to know as soon as it’s released.
