Interactive Prototype
Stage: Prototype
Performance Diagnostic
Live AI checks a training request against the team’s data, finds what’s behind the performance problem and who should own the fix, and drafts a four-week plan. You review it and make the final call.
Sample data · Fictional organization
Try one of the three sample tickets, or type your own request. All company data is fictional, and a person makes the final call.
Open full screen ↗
USER STORY
Primary user: L&D, HR, and talent leaders who field training and performance requests
TOOLS USED
Claude (diagnosis) · Lovable (app build and hosting) · Custom 115-factor library, built on Gilbert’s Behavior Engineering Model · Fictional company data set
POTENTIAL INTEGRATIONS
Jira (intake and backlog) · CRM and QA data · LMS · HRIS · Engagement platform · Slack (summarized themes only)
Ideas for future deployment. Not connected.
ASSUMPTIONS & WHAT I’D TEST
Assumes the data a team already has can point to likely causes before anyone builds a course. I’d test whether the diagnosis matches what experienced consultants conclude, and whether it shortens the time to the first stakeholder conversation.
WHAT GETS BETTER
Fewer training builds for non-training problems, clearer ownership when the fix sits outside L&D, and one agreed metric to watch for every request.
Potential benefits and how I’d measure them. No results claimed.
IN AN ORGANIZATION
How it would work for real
In a real organization, the tool would sit at the front of the request process. Every request gets a first-pass diagnosis, and a person decides what happens next.
1. A request comes in
Via Jira, ServiceNow, or an intake form
2. AI reviews the data
Quality, CRM, LMS, and HR records
3. AI drafts a diagnosis
Likely cause, owner, fix, and metric
Human-led
4. A person validates it
Talks with stakeholders and decides the fix
5. Work enters the backlog
Owner, pilot group, and manager plan
6. The team measures results
Scale, adjust, or stop at the checkpoint
AI’S ROLE
Reviews data from several systems at once
Spots patterns and compares groups over time
Drafts a hypothesis with its evidence
Drafts the plan, ticket, and stakeholder messages
PEOPLE’S ROLE
Talks with stakeholders and checks the findings
Adds context the data can’t show
Chooses and approves the intervention
Owns the decision and the results
WHY IT ADDS VALUE
Every request gets a diagnosis first, so effort goes where it will change performance.
The plan names the right owner up front: Operations, a manager, or HR.
Each request has one metric, a target, and a check-in date.
A four-week manager plan with huddles and coaching prompts keeps it going.