The three companies
Each audit looked at one company, or one part of it. These are the results.
| Company | Sector | Recoverable hours a year | Also found |
|---|---|---|---|
| Koskamp | Auto-parts distributor · 13 branches | 3,900 | ≈ €77,000 a year |
| C&L Lichtreclame | Sign manufacturer | 660 | ≈ €30,000 a year |
| Webron | B2B engineering group · sales team | 1,200 | ≈ 70% of sales prep time |
| Total | 5,760 |
How we count recoverable hours
We talk to the people who do the work. We map each process and count how often a task happens and how long it takes. Then we estimate, per task, which part AI can take over. That gives a number of hours per year.
Recoverable hours are not job cuts. They are capacity: time your people can spend on customers and on better work.
Where the time goes: four patterns
1. Typing the same data again
At C&L Lichtreclame, work orders still ran on paper. At Webron, salespeople typed data over between systems. Every time data moves by hand, it costs time and it can go wrong.
2. Searching for knowledge that already exists
At Koskamp, the answers to many support questions were already there, in thousands of solved tickets. A knowledge base built from those tickets is one of the two solutions behind the 3,900 hours.
3. Handling exceptions by hand
When an order at Koskamp is short, someone has to sort it out. Automatic shortage handling is the second solution. Exceptions feel small one by one, but they add up across 13 branches.
4. Preparation around the real work
In the Webron sales team, a large part of the day went to finding and qualifying leads before any customer contact. About 70% of that prep time can be won back.
What this means for your company
These patterns are not unique to these three companies. Any SME with repetitive, manual or knowledge-heavy work has them. Ask yourself four questions:
- Where does someone type data over from one system to another?
- Which questions does your team answer again and again?
- Which exceptions are handled by hand every week?
- How much of the day goes to preparing instead of doing?
For each answer, multiply the hours a week by the number of people and by about 46 working weeks. You will quickly see how big it is.
Where to start
Start with one process, not a big plan. Koskamp decided to build straight from the audit report. Our free AI Scan is the quick version: in one hour we name the 2–3 places where AI saves you the most.
Founder of StellarTech AI. Before AI, he ran a sales business and made over 30,000 cold calls. Now he builds AI systems for SMEs across Europe.