Stop taking orders. Start advising.
Every training request, diagnosed before anyone builds. Stop paying for training without business impact.
For L&D professionals and HR business partners.
- Performance Consultant AI
- 12-month learning program
A training request names the solution before anyone knows the problem.
A manager sees a problem and blames the people.
L&D gets a training request and delivers the training.
But the cause usually sits in the system around people.
So the training doesn’t help, and the problem stays.
Workplace learning program
Learning on your own requests with a senior performance consultant: four sessions in six months.
Leading experts
Guidance by the leading experts in performance consulting.
Four sessions
Ninety minutes each, in the first six months.
Your own requests
Personalized for your cases and training requests.
More coaching
Online, whenever you ask for it.
The program in twelve months
Working, learning and implementing at once.
Working and learning: Month 1 to 6. Learning and implementing: Month 7 to 12.
This is LOOP: team-based workplace learning with AI, on your own requests. And it does not stop after month twelve.
See the whole program
LOOP
Learning
A session with the whole L&D team and a senior consultant.
Operate
The team runs its next real requests through the Performance Consultant AI.
Optimize
The team looks back together: what did we miss, what do we ask next time?
Put it to work
What the team learns goes back into the work, and into the review at month twelve.
Performance Consultant AI
Diagnoses every request in minutes, in the AI your organization already uses.
Yearly license
Per team, included in the program.
25 years of experience
Right at your fingertips, in your own context.
Strong reports
Faster than ever before.
More cases
Use it in your work and handle more requests than ever.
How it works.
- The Sales Training Request ($5.4M Budget)
- The Request: A two-day sales training for 10,000 reps. The win rate suddenly dropped from 24% to 17%.
- The AI Diagnosis (Duration: 3 minutes): The AI discovers that the drop coincides exactly with the introduction of a new pricing model. Legal now takes 8.7 days to approve quotes instead of 2.5 days. Deals are stranded in the queue.
- The Advice: This is not a skills problem. Speed up the approval process. Cancel the ineffective training.
- The Onboarding Request (Nursing)
- The Request: A longer onboarding program for new nurses. A third leaves within the first year.
- The AI Diagnosis: The AI sees that the departures concentrate on three wards with structural understaffing. New colleagues there run night shifts on their own within two weeks. The onboarding program is the same on every ward.
- The Advice: This is not a knowledge problem. Don't schedule new colleagues alone at night for the first three months and give them a fixed point of contact. Don't extend the onboarding program.
- The Safety Training Request (Production)
- The Request: A mandatory safety training for 1,200 production workers. The number of incidents rose by 40% in six months.
- The AI Diagnosis: The AI discovers that 80% of the incidents happen on one line, since the switch to a new shift schedule. During a breakdown the break is dropped and the checklist is skipped. Nothing changed on the other lines.
- The Advice: No training for 1,200 people. Restore the break during breakdowns and make the checklist part of the handover. Focus on that one line.
- The Customer Service Training Request
- The Request: A conversation training for 300 customer service employees. Customer satisfaction dropped from 8.1 to 6.9.
- The AI Diagnosis: The AI finds that the drop starts on the day a new knowledge system went live. Search time per question doubled and waiting time rose to eleven minutes. Customers rate the waiting time, not the conversation.
- The Advice: This is not a conversation skills problem. Fix the search function and the most asked questions in the knowledge system. Put the training on hold.
- The Leadership Program Request
- The Request: A one-year leadership program for 120 team leads. In the employee survey, 'my manager gives direction' scores lowest.
- The AI Diagnosis: The AI sees that the lowest scores sit with team leads who have more than 25 direct reports and three ongoing reorganizations. They spend 70% of their time on reporting. Giving direction is not a skills gap here but an agenda problem.
- The Advice: Don't start with a one-year program. Shrink the teams, cut duplicate reporting and give team leads a fixed moment with their team. Then measure again.
- The E-learning Request (Code of Conduct)
- The Request: A new e-learning on the code of conduct for 8,000 employees. The audit found too many deviations from the procedure.
- The AI Diagnosis: The AI discovers that 90% of the deviations sit in one process, where the procedure no longer fits the system. Employees take the detour because the official route requires three approvals.
- The Advice: No e-learning for 8,000 people. Adapt the procedure to the work and remove two approvals. The audit resolves itself in the process, not in the inbox.
What is this training request really worth?
Value Flip
Three numbers are enough. You see what saying yes costs, and what the business avoids when the cause lies elsewhere. In money your CFO recognizes. It also shows why the program and the AI assistant are an investment, not a cost: you earn them back within a few requests.
Training spend
0
How the Value Flip works, request after request
Price the request
Participants, fees, travel and hours away from work. You see what saying yes costs, before anyone builds.
Test the training
Will this training remove the cause? The Performance Consultant AI tests whether knowledge or skill is the cause, or something in the work: tools, procedures, incentives, roles.
Advise and report
When training won’t remove the cause, L&D advises the fix that will. The budget that does not go to training without effect is avoided cost: value L&D creates, in a memo for your manager or CFO.
How to present the outcome to your management or CFO.
A number is not an argument. Most L&D business cases die in the first minute of the meeting, and always for the same reason. What finance does respect is avoided spend: budget that stays in the building, not a theoretical productivity gain. Here is how this one survives it, in the words that work.
Why this calculator exists at all
Finance stopped believing learning business cases because they add everything up and call the total a saving. A CFO discounts a number he cannot verify, and from then on the meeting is about your credibility instead of the decision.
So this one counts a single thing: money the organization approved and did not have to commit, on requests where the decision actually changed. It shows what leaves the building and what goes back into the operation, refuses every number that cannot be defended, and hands the interpretation back to finance.
What it counts, and what it refuses to count
Two things the request already counted: the money you pay out, and the hours of work the training takes, at the hourly cost your own finance team uses. One total as the request arrived, one after the analysis, and the difference between them.
The hours count because they were in the request. If they count when the organization buys the training, they count when it does not.
It refuses productivity gains, revenue effects, ROI multiples and the cost of the problem that stayed unsolved. And every request where the analysis confirmed the training counts zero, which is why the rest is believed.
What to bring, and what to leave at your desk
Bring three things on one page: the register extract for the period, one request worked out from arrival to decision, and the summary the calculator writes for you, with the split visible.
Leave behind the method, the framework and the names of the models. A CFO is not buying the instrument. He is deciding whether the number in front of him is one he can use.
The conversation, in six steps
Open with the decision, not with L&D. We received a request, we analyzed it before spending, this is what changed.
One subtraction, out loud. What the request asked for, what it costs after the analysis, the difference, and the split between money paid out and hours back into the operation.
Say what you are not claiming. No ROI, no productivity effect. Avoided spend only.
Hand over the choice. Avoided cost becomes a saving the moment finance decides not to release the budget, and capacity if it does. Then stop talking.
Show the register, not the total. Every amount has a date, a request and the cause that was found. Offer the register itself.
Ask for one thing. The next request routed to your team before the date is set.
The four things finance will say, and what to answer
“That money was never going to be spent anyway.”
It was approved, dated and in the plan, and here is the request. If it truly was not going to be spent, that belongs in the register too.
“Those hours are not real money.”
They were real enough to go into the request, at the rate the business case used. Hours out of a queue before peak season come back as overtime, agency cover or a longer queue, so they land in somebody's budget either way.
“How do I know the analysis was right?”
You do not have to take it on trust. Every entry carries its confidence level and what was still unconfirmed. Start with the entries at eight and above.
“So L&D is now saying no to training.”
No. Four of these twelve requests were confirmed as training and sit in the register at zero. What changed is that we know which four, and why.
This conversation is part of the program. Session four is about exactly this, because an analysis that never reaches the people who decide changes nothing.
The thinking behind it.
- White paper · 2026The request is not the problemThe Performance Consultant AI: how a training request leads to a grounded decision about performance.
- White paper · 2026The performance improvement processTwo doors, three analyses, five questions, and the map of the whole series.
- White paper · 2026Human performance technologyThe discipline in full: its history, its four traditions and the ten standards.
- White paper · 2026Environment before capabilityWhere the measured evidence points, and why the person comes last in a cause analysis.
Start with your team in three steps.
Thirty minutes
One real request from your team, diagnosed together. Free.
Your proposal
Per team per year, AI assistant included, priced for your region.
Day one
The AI assistant runs in your own AI. The first session is in month one.
Frequently asked questions.
What if the Performance Consultant AI gets it wrong?
Every conclusion states how certain it is and what is still unconfirmed. You and the requester decide. The Performance Consultant AI advises. AI drafts, the professional decides.
What does it cost?
You pay per team per year, with the Performance Consultant AI included, at a price that varies by region. You receive a quote after the intake. Your first diagnosis is free.
What happens to our data?
The Performance Consultant AI runs as an MCP server inside the AI your organization already uses, so your case material stays in your own AI environment, under the agreements you already have. You ask the requester for permission before you record anything. Trust and AI
Can we not just use ChatGPT?
You can, and next week you will get a different answer. We plug the performance consulting method into the AI you already use, so every request gets the same analysis.