← All posts Article · Nov 16, 2026

What our AI adoption survey of 11 CEOs shows about payback

Lennard Kooy··7 min read
What our AI adoption survey of 11 CEOs shows about payback

In this AI adoption survey, we asked 11 CEOs, most of them running companies that make, move or sell physical products, how AI is paying off. All 11 use AI tools, 4 measure what it delivers and 1 is scaling it. The 3 companies furthest ahead share 3 habits: a goal in hours or euros, training for the team and AI running live past the pilot.

At Lleverage we think this explains most of the gap between companies that use AI and companies that profit from it. The tools are everywhere. An owner, a number to aim for and one process taken live are rare. We build AI agents that handle orders, quotes and customer questions inside the ERP a company already runs, and a demo shows how one works on your own documents.

How did we run the AI adoption survey?

Each CEO completed a questionnaire between 9 and 21 October 2026. We scored every company from 0 to 100 on 5 dimensions: strategy, ownership, systems, people and results. No company or CEO is named here, and every figure in this post is a count or a total across the 11.

The sample is small. It is a close look at 11 companies and does not describe a whole market. The value estimates are indicative and rest on the CEOs' own figures: the FTE that each CEO said go into manual work, multiplied by the share an AI agent could take over (between 40 and 90% depending on the process) and by €65,000 per FTE per year.

Where do the 11 companies stand on AI?

Only 1 of the 11 is scaling AI. 6 are still experimenting and 4 are implementing. None is at the exploring stage, which means every company has moved past the first conversations.

StageScoreCompanies
ExploringUnder 250
Experimenting25 to 506
Implementing50 to 754
Scaling75 and above1

Who uses AI, and who measures what it delivers?

All 11 use AI tools. 8 say half the office uses AI every week. Only 4 measure the result in hours or euros. The other 7 measure nothing or only tell anecdotes.

What we askedCompanies out of 11
Use AI tools11
Half the office uses AI every week8
Measure the result in hours or euros4

Using AI is common well beyond our 11. Statistics Netherlands reported that 66% of Dutch companies with 250 or more employees used AI in 2025 (a provisional figure), up from 42% in 2023. Tool use no longer sets a company apart. Measuring what the tools deliver does.

What do the leaders do differently?

We compared the 3 companies with the highest AI score to the 4 with the lowest. All 3 leaders have a goal for AI in hours or euros, train their team and run AI live past the pilot. None of the bottom 4 does any of the three.

Habit3 leadersBottom 4
A goal for AI in hours or eurosAll 3None
Training for the teamAll 3None
AI running live, past the pilotAll 3None

Using AI tools made no difference. At all 4 of the bottom companies, half the office already uses AI every week. With 11 companies we cannot say which habit causes which result. We can say that the 3 sit together, and that each is a management decision and not a technology purchase.

Where does the value of AI sit?

72% of the estimated value sits in 4 processes: orders, quotes, customer service and purchasing. The figures below are yearly staff time spent on manual work that AI could do, added up for all 11 companies.

ProcessEstimated value a year
Order entry€1.1M to €1.2M
Customer service€769k to €1.1M
Quotes and tenders€699k to €933k
Purchasing and supplier confirmations€282k to €340k

The remaining 28% is spread over processes such as production planning, master data and reporting. For a company starting with AI, the order flow is the clearest place to begin. An agent for customer questions and one for purchasing and supplier confirmations cover two of the four. A guide to AI order processing for manufacturers shows what the first one looks like in practice.

How big is the upside?

The CEOs counted 58 to 77 FTE of office time spent on manual work that an AI agent could do, in just 3 processes per company. At an average of €65,000 per FTE, that is €3.8M to €5.0M a year in staff time.

Freed up, that time lets the same team handle more orders without hiring. The figure leaves out the cost of errors. The 6 companies that estimate it lose at least €1.0M a year between them on mistakes in administrative processes, and 5 do not know or did not fill it in.

One of the 11 has already cut the time per quote from 6 minutes to 45 seconds.

Do budgets, goals and the ERP hold AI back?

Budgets and goals do. The ERP does not. Hardly any company gives AI a budget or a goal, while 10 of the 11 already have working links between their ERP and other systems.

What we foundCompanies out of 11
No separate AI budget9
Projects start bottom-up7
No written AI policy6
AI goal tracked in EBITDA1
Working links between the ERP and other systems10
ERP runs in the cloud7
Plan an ERP migration3

The 3 companies that plan an ERP migration have a moment to design AI in from the start. The rest can add an AI agent next to the ERP they have. At SIG Benelux, a building materials distributor, the automation sits next to Dynamics 365, and 73% of order emails now arrive in the ERP ready to review.

"I knew the technology made this possible. My biggest question was the quality of our data. It was a positive surprise to see the order history was enough of a base to get results this fast." Ditte Smit, Marketing Director, SIG Benelux

What does a first AI process in 90 days look like?

The 3 habits of the leaders turn into a plan of 3 steps for a first process. Each step has one owner and one number.

  1. Day 0 to 30: make 1 manager the owner of AI and set 1 goal in hours or euros.
  2. Day 30 to 60: automate 1 process, such as orders from email, and train the team on it.
  3. Day 60 to 90: measure the manual work before and after, and report the result in euros.

Our AI adoption playbook for manufacturing goes through each step, and a longer guide to AI adoption in manufacturing covers where to start, what to supervise and when a step can run on its own.

What should a company do with these findings?

Pick the process where your office spends the most manual hours, name its owner and set a goal in hours or euros. In the scan, that process was most often order entry, customer service or quotes. Then measure before you start, so the result in euros exists when the budget conversation comes.

A demo is a quick way to test it. Bring 5 of your own orders, invoices or quote requests, including one you find difficult, and see what an AI agent reads, what it would enter in your ERP and what it would hand to a person.

Frequently asked questions

Do most companies measure the return on AI?

Few do. In our survey, 4 of 11 companies measure what AI delivers in hours or euros, and only 1 tracks an AI goal in EBITDA. All 11 use AI tools, so the gap sits in measurement and goals. Measuring manual work before and after the first process closes it.

Which process should a company automate with AI first?

Start where the manual hours are highest. In our survey, order entry carried the largest estimated value, €1.1M to €1.2M a year across the companies that named it, followed by customer service and quotes. A good first process has high volume, an owner who feels the pain and mostly similar cases.

How much staff time could AI free up in a back office?

The 11 CEOs counted 58 to 77 FTE of office time on manual work that AI could do, in 3 processes per company. At €65,000 per FTE that is €3.8M to €5.0M a year. It is an indicative estimate based on the CEOs' own figures.

What separates the leaders from the rest?

The 3 leaders all have a goal for AI in hours or euros, train their team and run AI live past the pilot. None of the bottom 4 does any of these. Using AI tools made no difference, because half the office at all 4 of the bottom companies uses AI every week.

Does the ERP hold AI adoption back?

In our survey it does not. 10 of the 11 companies have working links between their ERP and other systems, and 7 run their ERP in the cloud. The gaps were budgets, goals and training. A company that plans an ERP migration, as 3 of the 11 do, can design AI in at that moment.

Give your back office an AI workforce.