Seven in ten senior leaders say they are making good progress toward a fully integrated AI-human workforce. But walk the floor of many of those companies, and the “AI-human workforce” turns out to be little more than an email containing a login to a promising new tool.

That gap is the real story of enterprise AI right now.

Adoption is Racing Ahead of Value

KPMG’s latest Global AI Pulse surveyed more than 2,000 senior leaders across 20 countries. The share of organisations actively driving AI adoption nearly doubled in a single quarter, rising from 13% to 22%. 75% now say AI is delivering meaningful business value.

And yet established ROI, the ability to prove that AI investment produces real returns, remains rare. Adoption, confidence and spending are climbing. Proof is lagging behind.

The tools are the easy part. The hard part is redesigning the workflow around them. As KPMG puts it,

“The next challenge is not deploying AI. It is redesigning work for a world where humans and AI operate together. That’s not a technology exercise. It’s a rethink of roles, workflows and how value gets created.”

What “Redesigning the Workflow” Really Means

That phrase gets quoted far more often than it gets understood. It is more complicated than it sounds, so here is a simple example.

Think about requesting information the old way, on a typewriter. You draft a letter, print it, mail it, and wait for a reply.

Then the laptop arrived. You draft an email, send it, and wait for a reply. Faster, yet the same shape.

Now picture it today. You tell an AI agent what you are looking for. The agent gathers context before it searches, realises the information already exists from last week when a colleague in another department asked for the same thing, strips out the irrelevant parts, answers your request, and then offers two recommendations on what to do next.

Notice the difference? Moving from letter to email only made the old steps faster. The agent version removed steps altogether. That is what redesigning the work actually looks like.

Access is not adoption, and adoption is not capability.

What you are really after is value, and value shows up when you rebuild the steps around what people and AI each do best: which parts move to AI, which stay human, and where the handoff happens.

Why We Keep Paving the Cow Paths

This is not a new problem. Michael Hammer, the MIT computer scientist known as the father of business process reengineering, spotted it decades ago and once described his life’s work as “undoing the Industrial Revolution.”

His most famous idea started with cow paths. A cow path is the crooked trail cattle wander to cross a field. For generations, towns paved right over those winding trails instead of building a straight road, so the inefficiency never left. It just got a smoother surface.

Hammer argued that companies do the very same thing with technology. They take a process built for a world that no longer exists, drop a shiny new tool on top, and then wonder why nothing really changed. His line from 1990 still holds today:

“Don’t automate, obliterate.”

That is the discipline behind workflow redesign. Before reaching for another tool, walk one of your own cow paths: pick a process you have always run the same way, and ask whether you would build it like that if you were starting from scratch.

The Same Shift, Department by Department

This is easier to see with concrete examples.

HR and Onboarding

Speeding up the same welcome email is automation. Redesign looks different. An AI agent assembles a tailored (yes, customized and personalized to the person) onboarding plan from the role, the team and past joiners, while your HR people spend their time on the human welcome that software cannot replicate.

Finance and Month-End Reporting

Rather than typing variance commentary a little faster, let AI produce a first-pass commentary straight from the ledger, and move your analysts to exception review and the decisions that follow.

Sales and Outreach

More cold emails per hour is easy to celebrate, but the outcome that matters is replies and conversions. Redesign so AI researches each account and drafts context-aware outreach, freeing your reps for the live conversations that actually convert.

Customer Support

Instead of copy-pasting macros faster, let AI resolve the repetitive tickets end to end, so your people handle the escalations and edge cases that need judgement.

As you can see from the examples across different departments, the biggest gain is a step that disappears entirely. We’re no longer building a tar road on top of the cow path; we’re redesigning the highway that connects critical parts of the town.

Why This is a Leadership Job

This is a management and enablement decision, and it belongs to the people who own the work. IT can enable it, but IT cannot own it.

The data backs this up. Organisations with clearly defined accountability for AI outcomes report established ROI at more than three times the rate of those without it, 14% against 4%.

Those with full visibility into AI operating costs are five times more likely to reach established ROI, 15% against 3%. At the same time, 49% of organisations have already delayed or scaled back AI agent deployments once costs began to outweigh benefits, and 78% of leaders expect roles to change for employees who do not build AI fluency.

The organisations pulling ahead are the ones redesigning how work gets done and naming who owns the result.

Here is the honest test. If your AI rollout disappeared tomorrow, would your team’s workflow look any different from last year? If the answer is no, you have added a tool, and the redesign is still waiting.

Ending Thought:

Adoption is the easy headline. Value comes from the quieter, harder work of redesigning how people and AI operate together, step by step and workflow by workflow. The survey data is consistent: the organisations proving ROI are the ones managing AI better and rebuilding the work around it.

If your teams have the tools but the results are still waiting, this is exactly the gap Radiant Institute helps close. Our AI enablement programs focus on redesigning real workflows and building the fluency and accountability that turn adoption into measurable value. If that is the challenge in front of you, let us talk.