The AI adoption illusion
Teams are using AI every day. Whether the actual work has changed is a different question.
Here is how to tell the difference, and the five moves we see from the organisations closing the gap.
Why the gap matters
Most of the AI conversation in large organisations has landed in the same place: the divide is organisational rather than technical. What most pieces stop short of is the second question, which is what to do about it when the accountability sits on your desk.
Across marketing, insights and data, teams are using AI every day. The operating models around them have not moved with them. That is the adoption illusion, and it becomes more expensive to close the longer it sits.
A team can use AI every day and keep the same workflow, approvals, measurement and agency model around it. How the business runs has not shifted. Agentic AI raises the stakes further, because agents only pay off when the work around them has been rebuilt to give them something meaningful to do.
When these programmes stall, the technology is rarely what killed them. Our Resilience Report points at human and organisational barriers instead: ownership that was never quite assigned, and a quiet caution around being replaced that surfaces as delay.
€1.25bn
Operating profit at stake for a typical consumer goods and luxury company when AI transformation stalls on organisational barriers rather than technical ones.
Source: Pivot & Co Resilience Report 2025
The advent of AI requires a fundamental shift in mindset, from merely identifying problems to actively seeking out interesting solutions.
Axel Adida
CDO Beiersdorf. Pivot & Co Resilience Report.
How to tell
A simple check we use with clients: name a workflow AI has actually changed in your organisation, and say who owns it now. If that takes more than a moment to answer, you are somewhere in the 95%.
The leaders who could answer quickly have something in common: they act. Overthinking is one of the quieter ways an AI programme stalls. Committees form, frameworks multiply, decision papers circulate, and the ground moves before anyone commits. The organisations closing the gap run smaller decisions faster and take the learnings as they come.
Five patterns come up in those leaders, drawn from Resilience Report interviews with senior data and transformation people across consumer goods and luxury.
The five moves
01
Assign ownership before the pilots multiply
New AI capabilities usually start in one corner of the business and spread. Without clear ownership and decision rights early, teams build different tools and different answers to the same problem. The organisations moving fastest assign one owner per use case before the work sprawls across functions.
Gartner has at least half of gen AI projects abandoned after proof of concept, most often because ownership and business value were never clear at the start.
What ready looks like
Every AI initiative has one named owner and clear decision rights, agreed before work starts.
Question to ask now
For each of your AI initiatives, can you name the single person accountable for it?
02
Rebuild the workflow end to end
A faster brief, image or analysis makes little difference on its own. Value compounds when the whole path is rebuilt around AI, from insight through brief, creative, approval, activation and learning.
Across our Resilience Report interviews, value only showed up where the whole workflow around the tool had been rebuilt.
What ready looks like
At least one end-to-end workflow rebuilt around AI, with a named owner and a changed operating rhythm around it.
Question to ask now
Which workflow have we redesigned end to end, and what actually changed in how the team works?
03
Cluster use cases on one prioritised workflow
Running a single use case in isolation is one of the easier ways to look busy without changing much. Pick one workflow top-down, insights into activation for example, and gather several high-value use cases on it. Named owners per use case, a cap on live use cases, and a scale-or-kill gate keep the cluster honest.
MIT's Project NANDA found 95% of pilots deliver no measurable return, mostly because they ran as isolated experiments.
What ready looks like
Every active use case sits on a mapped workflow with an owner and a scaling path.
Question to ask now
Which workflow are we actually transforming, and how many of our AI use cases sit on it?
04
Build the foundations through the use cases
No AI ambition survives poor data foundations. Foundations built in the abstract, ahead of any use case, tend to sit unused. Build them in sync with the priority use cases, so fixes are targeted and appetite grows with each round of visible value.
As Ian Curd, Global Lead Data Foundations at Diageo, puts it, a resilient approach has to be "in-built" rather than an "add-on".
What ready looks like
Foundations funded alongside the priority use cases they support. Every fix ties to a live workflow, and AI spend is logged against outcomes.
Question to ask now
Which foundation is our next use case going to expose, and are we ready to fund the fix as it goes in?
05
Measure whether the work moved
This is where the illusion becomes visible. Usage and outputs go up while the underlying business stays where it was. The organisations that get past this define success up front on a shared set of lenses, agreed by business and technology, so any programme has to show the work itself changed.
Our Resilience Report found chaotic measurement is one of the main reasons programmes lose credibility. When anyone asks for actual impact numbers, the answer is already thin.
What ready looks like
Success on a shared set of lenses, agreed by business and tech: operational efficiency, commercial value, user behaviour.
Question to ask now
Do our AI metrics show the business has changed, or only that we have been busy?
Is your organisation using AI, or has AI changed how it runs?
The mandate is broad, the path unclear. That is where we work.
How Pivot & Co can help
-
AI Strategy & Partner Ecosystem
Set the AI direction, choose partners independently of any vendor, and turn the roadmap into how the business runs.
-
Marketing Workflow Transformation
Take the workflow as the unit of change, and redesign how marketing runs day to day so AI and data land in the work.
-
Operating Model Design & Adoption
Build the operating model that holds, and the team equipped to keep it running.
