AI was supposed to make work easier.

That was certainly the promise. Faster decisions, less admin, more time for thinking, judgement and creativity.

But the reality across many businesses is more mixed. PwC’s 29th Global CEO Survey (2026) found that 56% of CEOs say their company has seen neither higher revenues nor lower costs from AI over the last 12 months, while only 12% report both of these positive impacts.*¹

Boston Consulting Group’s “Where’s the Value in AI?” research adds to that picture: 74% of companies have yet to show tangible value from their use of AI.*²

Those figures matter, because they mirror what many leaders are experiencing day-to-day.

In many of the organisations I hear from, teams feel busier. Work feels more fragmented. Roles feel less clear.

Despite new tools, dashboards and automation, progress still stalls, and frustration quietly builds.

This tension really struck me during Stanton House’s recent Winning with AI leadership event.*³ Listening to leaders from across finance, data, transformation and technology, one message came through consistently: the organisations making real progress weren’t the ones talking most about tools, they were the ones starting with the problem they were trying to solve.

That resonated deeply.

Because in practice, what I see far too often is technology being introduced into ways of working that were never designed for it. And when that happens, AI doesn’t fix the problem, it exposes it. Often creating confusion and disengagement with the tech itself.

AI Isn’t the Problem, Work Design Is

AI hasn’t arrived into a neutral system.

It’s landed inside organisations that were already stretched, layered and often poorly designed for speed. What it’s doing now is acting as a stress test - revealing where tasks, processes and decision-making no longer line up with how work actually happens.

One of the most common things people miss when new technology is introduced is the need to redesign the work around it.

Not just roles. Not just job descriptions. But expectations, tasks, processes and workflows end-to-end.

If you introduce AI into a process without redesigning how work flows, all you really do is accelerate existing friction. The inefficiencies don’t disappear, they just move faster.

Tasks Are Changing Faster Than Jobs

One of the biggest misconceptions about AI is where its impact is really landing.

So far, AI hasn’t eliminated work at scale. What it has done is fundamentally change tasks — how work is done, in what order, and by whom.

Activities that once took hours or days now take minutes. Drafting, analysing, modelling and summarising have all accelerated. But that acceleration on its own doesn’t guarantee better outcomes.

In fact, EY’s 2025 Work Reimagined Survey found that while AI can unlock up to 40% more productivity gains when used effectively and on stable talent foundations, most employee usage remains basic, and only 5% report using AI in advanced ways that genuinely transform how they work.*⁴

That gap helps explain what many organisations are experiencing next.

In most cases, the roles those accelerated tasks sit within haven’t evolved at the same pace. Job titles, org charts and responsibilities are still built for a world where work changed slowly. Today, tasks evolve continuously, sometimes weekly, while roles lag behind.

Research increasingly shows that AI is changing how work is done at the task level far faster than it’s eliminating jobs, which helps explain why roles are struggling to keep up.*⁵

When tasks change but roles don’t, clarity erodes. And when clarity erodes, people fill the gaps themselves — with workarounds, duplication and extra effort.

The Overlooked Challenge: Tool Overlap and Unclear Expectations

Another issue I see regularly is tool sprawl, often well-intentioned, but poorly integrated.

New technology gets added alongside existing platforms, with little thought given to overlap or expectations. Teams are left asking:

• When should I use this tool versus that one?
• Which output is the “source of truth”?
• Is this meant to replace part of the process, or sit alongside it?
• What’s good enough, and what still needs human judgement?

Without clear answers, people default to caution. They double up. They sense-check everything manually. They spend more time managing tools than benefiting from them.

The result isn’t efficiency, it’s cognitive load. And when technology speeds everything up without fixing that, the mental load just increases. *⁶

And once again, the issue isn’t the technology. It’s the absence of intentional work design around it.

The Stress Test: What AI Exposes When Work Design Is Weak

AI doesn’t create organisational problems. It reveals the ones that were already there.

If any of the following feel familiar, the issue probably isn’t adoption, it’s design:

• AI tools exist, but decisions still bottleneck
• Automation increases volume, not clarity
• Teams feel busier rather than more effective
• Accountability feels fuzzier, not sharper
• New hires struggle to understand where they add value
• Leaders can’t clearly explain who owns what - or why
• People are experiencing AI & tech change fatigue

These patterns aren’t isolated. They’re showing up across organisations, sectors and leadership teams.

McKinsey’s 2025 Global Survey on the State of AI reflects this at scale. Despite widespread AI use, only 39% of respondents report a measurable EBIT impact at the enterprise level, highlighting how many organisations are still struggling to turn adoption into sustained business value.*⁷

None of these are technology failures.

They’re signals that work hasn’t been intentionally designed for the reality it’s now operating in.

And when that happens, the burden lands on people.

Where the Human Cost Shows Up First

When work isn’t designed, people absorb the complexity.

Cognitive load increases. Ambiguity becomes normalised. High performers quietly carry more than their share.

This doesn’t land evenly.

Neurodivergent employees are often forced to decode unclear expectations. Experienced hires are expected to “just work it out.”

Those with non-linear careers or different working styles are judged as misaligned - when the real issue is design, not capability.

Poor design doesn’t just slow organisations down. It quietly excludes talent, accelerates burnout and increases disengagement.

Designing Work for an AI-Enabled Reality

Good workforce design doesn’t start with tools or structures. It starts with intent.

In the organisations navigating this well, a few principles consistently show up:

Start with the problem, not the technology - Be clear about what needs to change - and why - before choosing tools.

Design work before assigning roles - Clarify the tasks and workflows first, then decide who owns them.

Be explicit about where human judgement matters most - Not everything should be automated - and that’s a strength.

Set clear expectations around tools - When to use them, when not to, and what “good” looks like.

Treat design as iterative, not a one-off exercise - Because work will keep evolving.

Employees are allocated time to adapt and experiment with new ways of working - Pilot groups are deployed to help understand and challenge.

This isn’t about getting it perfect. It’s about being intentional.

From Efficiency to Effectiveness

AI makes it tempting to optimise purely for efficiency.

Faster outputs. More volume. Less friction.

But effectiveness doesn’t come from speed alone. It comes from clarity, flow, judgement and trust.

I’ve seen organisations move very quickly with new tech - and still end up stuck. Not because the technology didn’t work, but because the work around it wasn’t designed to.

AI doesn’t reward organisations that move fastest. It rewards those that are designed to move well.

The Leadership Moment We’re In

AI will keep evolving. Tools will improve. Capabilities will expand.

But poorly designed work will always hold organisations back, no matter how advanced the technology becomes.

The real leadership challenge right now isn’t just adopting AI.

It’s redesigning work so people can do their best thinking, not just keep up.

In the age of AI, leadership isn’t about planning headcount.

It’s about designing work for a future that’s already here.

Sources

_{1. PwC, 29th Global CEO Survey (2026): Leading through uncertainty in the age of AI — “More than half (56%) say their company has seen neither higher revenues nor lower costs from AI, while only one in eight (12%) report both of these positive impacts.}

_{2. Boston Consulting Group (BCG), AI Adoption in 2024: 74% of Companies Struggle to Achieve and Scale Value / “Where’s the Value in AI?” — “Seventy-four percent of companies have yet to show tangible value from their use of AI.”}

_{3. Stanton House Leadership Event Insights - “Winning with AI”}

_{4. EY 2025 Work Reimagined Survey: “AI can unlock up to 40% more productivity gains… [yet] only… 5% are using it in advanced ways to transform the way they work.”}

_{5. U.S. Census Bureau & research survey analysis showing AI’s greater impact on tasks than employment levels - “Will the adoption of AI by businesses substitute for worker tasks or jobs?”}

_{6. Researchers in organisational psychology and information systems have found that unclear processes and decision rights increase cognitive load - meaning people spend more mental effort just staying coordinated, especially when technology accelerates work without clarifying how it should be done. See: “The impact of information overload and unclear decision structures on employee performance and stress”, Journal of Business Research.}

_{7. McKinsey & Company, The State of AI: Global Survey 2025: Reports that while AI adoption is now widespread across organisations, only 39% of respondents say AI has delivered a measurable impact on enterprise-level EBIT, highlighting the gap between adoption and realised business value.}

_{Outspoken is Stanton House’s thought leadership series where we raise the volume on the real-world challenges shaping work today - through the lens of leadership, inclusion, and human potential.}