Adaptive Track

AI closes the loop
theory → practice.

Trainee takes the course. Fails the evaluation. AI diagnoses which competencies in the framework they were weak in. Generates, automatically, a personalised roleplay template to drill exactly those gaps. Manager reviews and publishes. Trainee runs it.

Request Demo Back to overview

How it works, step by step.

No magic. Four deterministic stages, each observable by the manager.

Course in Academy

Trainee walks through the chapters and takes the final evaluation, multiple choice + open-ended.

Gap diagnosis

AI maps every wrong question to the competencies tested. No keyword heuristics, only what the template declared.

Match in the framework

Gaps are positioned in your competency matrix, not in a platform-wide catalogue.

Roleplay generated

Template auto-created with persona + scenario + criteria. Manager reviews, publishes and assigns, trainee runs it.

Real example

"João missed 40% of the Anchoring questions in the quiz."

Instead of forcing him to retake the entire course, the Adaptive Track generates a negotiation roleplay focused on price anchoring with a specific scenario (a customer asking for an unexpected discount). After he runs it:

Anchoring jumped to 78% on the next diagnostic. No 8-hour course redo. No coordinator hand-holding.

Breaks the "trained and forgot" cycle, a classic in pharma and banking, as reported in a pilot with a large regional bank.

Anchoring (before the roleplay) 40%

Adaptive track completed

Anchoring (after the roleplay) 78%

What makes this possible.

The Adaptive Track only works because the evaluation is anchored in a framework. Otherwise, it's just "another AI guessing".

Frozen criteria at dispatch

When a trainee is enrolled in the course, the evaluation criteria are frozen at that framework version. Even if you edit the framework later, their evaluation stays on the original criteria. No noise. No drift.

Manager-review gate

AI never publishes a roleplay on its own. The generated template lands in the manager's queue, they review persona, scenarios, criteria and edit whatever they want. It only ships if the manager approves. Compliance sleeps easy.

Level 1 manual + Level 2 automatic

Configure per course: Level 1 shows a preview of the suggested roleplay, the manager decides case by case. Level 2 generates + publishes automatically after review. You pick how much automation to delegate.

Auditable end to end

The generation logs: which questions were wrong, which competencies were mapped, which prompt the AI used, which framework version. For regulatory audit, the entire cause→effect chain is documented.

Availability