A new report out of Paris puts a number on something a lot of leaders already feel in their gut. The tools are in the building. The advantage is not.
Publicis Sapient released its 2026 Global Enterprise AI Report at VivaTech on June 17, 2026. The headline finding is a split between how widely enterprises have adopted AI and how ready their organizations actually are to capture its value. A vendor produced this study, so read the numbers as the consultancy's own read rather than neutral fact. The shape of the story still holds, and it is worth your attention.
The adoption number is high. The "core to the business" number is not.
The Publicis Sapient report found that 73% of respondents say they use AI regularly or across most business processes. That is close to universal at this size of company.
Then the floor drops out. Only 10% say AI is core to their business operations.
of large enterprises say AI is core to their business operations, against the 73% who report using it regularly.
Source: Publicis Sapient, 2026Sit with the distance between those two figures. Three out of four companies have AI running somewhere in the work. One in ten has built it into the center of how they operate. The first number is easy to reach. You buy licenses, you turn on a feature, you let teams experiment. The second number is hard, because it is not about the tool at all.
The report makes the same point another way. 47% believe AI currently meets their business needs. But 42% say something more pointed: the AI is capable, and their own organization cannot capture the value. Read that again. Nearly half are telling researchers the technology works fine, and the block is internal.
When you ask where the block sits, 22% name their own organizational operations as the primary barrier. Not the model. Not the budget. The way the work is organized around the model.
What the report calls "readiness," and why it is the real story
There is a difference between deploying AI and being ready for it. Deployment is a purchase decision. Readiness is an operating decision.
The U.S. figures show the split clearly. 71% of American respondents expect significant progress scaling AI over the next 12 to 24 months. Only 20% say they are fully equipped for that expected progress. The wave is widely expected. One in five feels ready to ride it.
Treat that as a baseline, a measurable starting point you can act on instead of a vague worry about being behind.
Nigel Vaz, CEO of Publicis Sapient, put it this way in the announcement:
"The enterprise was not designed for the speed, scale and autonomy that AI makes possible. Many organizations have successfully deployed AI, but deployment alone does not create advantage. The winners will be the companies that redesign how work gets done, modernize their operations and embed AI into the fabric of the business."
Redesign how work gets done. That is the line to take into your next leadership meeting.
The trap most teams fall into
The instinct, when you feel behind, is to assume you need a better tool. The right model. The right prompt library. One more subscription. So companies buy, and adoption climbs, and the advantage still does not show up.
AI amplifies whatever process it inherits. Point it at a clean, well-designed workflow and it makes that workflow faster. Point it at a messy one and it makes the mess faster. The 42% who say the technology is capable but the value is stuck are describing exactly this. They bought capability and dropped it onto operations that were never built to use it.
This is why the order of work flips. You design the workflow first. Then you build the agent to run it. A team that does this with one real process learns more than a team that buys five tools and hopes.
Where this maps in the 7 Levels of AI Proficiency
LaunchReady measures AI capability on The 7 Levels of AI Proficiency. It is the same idea the report is circling, with names attached.
Early levels look like adoption. People use AI tools, get comfortable, write better prompts. That is where most of the 73% sit. Real, useful, and not yet an advantage.
The higher levels look different. A Level 5 Captain is a design thinker who reshapes a process so AI can run it. A Level 7 Mission Director orchestrates AI across the operation as a system, not a set of one-off helpers. That is what "core to the business" actually requires. The 10% are operating up the levels. Most of the rest are not, yet.
The point is not that low levels are bad. Everyone starts somewhere. Climbing is a choice, and the climb is where the value shows up.
The picture is the same across borders
The report surveyed 1,550 AI decision-makers across six countries: the U.S., U.K., France, Germany, Australia, and the UAE. Respondents came from organizations with 500 or more employees and $100M+ in annual revenue, and they were responsible for evaluating or selecting enterprise AI. Fieldwork ran from April 29 to May 14, 2026.
The regional cuts, as the report presents them, tell a consistent story:
- In the U.K., 51% report fundamental business transformation and 60% report high or full AI workflow embedding.
- In France, 24% report fundamental transformation, and 51% cite internal data as the primary constraint.
- In Germany, 35% use AI as a "colleague," while 10% have full enterprise integration.
- In Australia, 53% report high or full workflow embedding and 38% report fundamental transformation.
- In the UAE, 60% report coordinated cross-team AI, and 5% have full enterprise integration.
Different starting points, same lesson. Plenty of usage, far less true integration. Globally, 38% say AI is fundamentally changing their operations, which leaves the majority still treating it as a bolt-on.
What to do this week
You do not need a transformation program to act on this. You need one process and an honest look at it.
Choose a workflow that costs your team real time. Write down how it works now, including the parts that are messy. Find the one step where AI could do the heavy lifting, and design the handoff so a person still owns the decision. Run it once. Measure what changed.
That is the move from the 73% to the 10%. One workflow at a time.
If you want a baseline before you start, take The 7 Levels of AI Proficiency assessment. It takes about ten minutes and tells you where your own AI capability sits today, so the work you do next has a number to measure against.
Sources
Related reading: Level 5: The Captain (Design Thinker).
Frequently Asked Questions
Is this just a consultancy selling its services?
Publicis Sapient is a digital transformation firm, and the report supports its business. That is why the numbers above are attributed to the report rather than stated as settled fact. The internal-readiness pattern it describes shows up across enough independent research that the direction is credible, even if you discount any single figure.
Does a high adoption rate mean my competitors are ahead of me?
Not necessarily. Adoption is the easy 73%. The harder question is whether anyone has redesigned a core workflow around AI. Most have not. The opening is in the integration, where only about 1 in 10 are operating.
What is the smallest first step?
Pick one workflow you own. Map how it actually runs today, step by step. Then ask where AI could carry a piece of it. That single exercise tells you more about your readiness than any vendor scorecard.
Find your AI Proficiency level
The free 7 Levels assessment places you across seven stages of AI capability. Under ten minutes. Research-backed scoring.