What AI changes in every job at your company,
seen task by task, so you can decide together.
For HR leaders, executive leadership and transformation teams. AI is already part of your teams' daily work, your executive committee expects a figure and your staff want answers. Starting from your job descriptions, we give you a tool to measure what AI changes, put figures on the options and discuss them with everyone. And we stay by your side for what comes next.
Your org chart will never tell you what AI will change.
It shows, for example, 40 payroll officers, 120 positions in customer relations and 15 in purchasing. But behind each job title lie very different activities, and AI affects each one in its own way. According to the Anthropic Economic Index (2025, US data), AI is used for at least a quarter of tasks in 36% of jobs, and for three quarters of tasks in only 4%.
Reasoning position by position means moving forward blind, with the risk of automating the very tasks through which your junior staff learn their job. To see clearly, you have to go down one level, to the task.
AI doesn't replace jobs. It changes tasks.
Four words are enough to see clearly. The position appears on the org chart; the job brings similar positions together; the task is a specific activity within the job; the skill is what you need to know how to do to carry it out. AI acts at the level of the task. Here, as an example, is the job of payroll officer:
Entering pay data received by email
A repetitive, well-defined task that AI can take on entirely.
Answering routine questions from staff
AI prepares the answer; a person reviews it, adjusts it and sends it.
Checking anomalies before payroll closes
AI spots the discrepancies; the payroll team interprets them and decides.
Supporting someone on long-term leave
A delicate situation, where listening and judgment do all the work.
Checking what AI has produced
A task that didn't exist yesterday. It goes into the job description and calls for a new skill.
You start from your job descriptions, you leave with a clear view.
- Étape 1
You share your job descriptions
Along with headcount per job. No personal data: we work on jobs, never on people.
- Étape 2
Every task goes through four questions
What AI can do, what your teams want to hand over to it, what needs preparing in your company, and how fast things will move. Two people check every answer, as our method explains.
- Étape 3
Decisions are built together
We present the results to your executive committee, then the tool stays in your hands. Leadership, managers and social partners then move forward on shared facts.
Answers to the questions you're already being asked.
Where AI is capable and your teams are willing
According to Stanford (2026, 844 office tasks studied), that is the case for 43% of tasks. That's where buy-in comes most naturally. The rest is prepared first with the works council.
How far AI goes
One automation level set for the whole company, then adjusted job by job. Every setting immediately shows its consequences in figures.
For executive leadership HR leadershipThe skills to develop
For each job, the new skills and the order in which to acquire them: the raw material for your strategic workforce planning (GEPP) and your skills development plan.
Prepare workforce planning Executive leadership · HR leadershipTime freed up and budget
Time freed up in hours, in full-time equivalents and in euros, and the cost of each option: everything you need to answer your finance team, figure by figure.
Put figures on the transformationNo scoring, no predictions about people
No individual scores, no sorting of applications, no decision made about a person.
« The AI shift is first and foremost a question of skills. And good decisions start with a good map. »
The questions we're asked most often.
Do we need to send you data about our staff?
No. Job descriptions, headcount and average salaries per job are enough. The questionnaires for your teams come later, after internal communication, and remain anonymous.
What do we ask of our teams?
For each task, how much they want to hand it over to AI, from 1 (not at all) to 5 (completely), and the time they spend on it. That's what separates what AI can do from what it makes sense to hand over to it.
What about training our teams?
Since February 2, 2025, the EU AI Act has required employers to ensure their staff have a sufficient level of AI literacy. Our two programs start from your company's jobs: Altitude for the executive committee, Cordée for teams. All the details are on the AI Act page.
Behind every diagnosis, our skills graph.
Every skill is linked to its prerequisites and its neighbors: that is what reveals the pathways between jobs.
And at your company, where do you start?
In 30 minutes, you walk us through your situation. We show you the tool in action on a fictional company, then we sketch together what a pilot would produce at your company. No commitment.
30 minutes on video with Skillberg, no commitment. No personal data required.
- Anthropic Economic Index, Handa et al., 2025, p. 1-2 and p. 13 (preprint; Claude.ai usage, US data from late 2024)
- Shao et al., Future of Work with AI Agents, Stanford, arXiv v3, February 2026, p. 3 and p. 24 (preprint; 1,500 US workers, 844 office tasks)
- Regulation (EU) 2024/1689 on artificial intelligence, Article 4