We replaced long, static training courses with an AI tool that spots skill gaps from live work data and builds short learning paths for each role. Course completion rose 40%, and people built skills 25% faster.
Industry: HR TechService: AI & AutomationUpdated
Overview
The challenge
The client's older learning management system (LMS) offered static courses that lasted hours and had little to do with daily work. Completion ran 40% below industry benchmarks, because employees saw training as something that got in the way of their jobs.
Everyone in a broad group, such as all managers, got the same course, whatever their skill level. Success was measured in clicks, not in whether people actually got better at the work.
Fixed course plans meant field staff worked from outdated protocols for 3–6 months before being retrained.
Millions went on licensed content that only 12% of the global workforce actively used.
Safety certifications were tracked by hand, which left serious gaps and regulatory exposure at audit time.
The solution
We built a live data flow that reads work signals, such as missed KPIs and ticket resolutions, to spot skill gaps as they appear.
AI tools compare each gap with a structured skills map and build a custom path of 5-minute lessons.
Each lesson and the performance change that follows it are saved in an auditable record, to prove return on investment and compliance status.
Before and after
Before
Long courses
Generic 60-minute modules on a fixed quarterly schedule.
After
Short lessons
5-minute lessons delivered when live performance data shows they are needed.
Before
By role only
Broad groups, such as all managers, got the same training regardless of skill.
After
By individual skill
Each person gets their own path, based on a map of their current skill levels.
Before
Completion clicks
Success was measured by clicks rather than real improvement.
After
Measured results
Success is measured by the link between training and KPIs recovering.
Key features
Graph neural networks predict which skills teams will need next, based on new project data and industry shifts.
Business impact
Zero outdated course plans
Performance scores are anonymised, which meets EU GDPR requirements while still giving the learning team useful insight.
Business impact
100% GDPR aligned
Cloud delivery keeps path generation under 2 seconds, even when staff around the world log in at the start of their day.
Business impact
40% engagement lift
Results
40% Completion Lift
Short lessons removed the time barrier, and more employees chose to take part.
Before: LowAfter: 40% Higher
25% Time Reduction
Targeted lessons shortened the time between spotting a skill gap and confirming it was closed.
Before: SluggishAfter: Precise
1.8s p99 Response Time
Personal paths are ready almost instantly, however many people are using the tool.
Before: < 5sAfter: 1.8s
Faster delivery
Reusing these components saved about 8.5 weeks of setup, which went into the client's skills map and lesson content.
Provided the structured skills map that each gap is compared against, which we filled with the client's own roles.
Gave the team tested logic for turning a skill gap into an ordered path of 5-minute lessons.
Pulled KPI and ticket data from Jira and the ERP into one live feed for gap detection.
Tracked how much each path cost to generate and deliver, so spend could be compared with results.
Trust and control
All employee performance data is encrypted at field level and stored in region-specific private cloud environments.
GDPR-alignedEvery recommendation explains why that content was assigned to that person.
Explainable AIThe cloud-based platform stays reliable during busy shift changes.
Cloud nativeOn completion, the client owns 100% of the IP in the skill graph and the AI learning tool.
100% ownedThe old LMS offered generic 60-minute courses on a fixed quarterly schedule, so training felt like time away from real work. Completion ran 40% below industry benchmarks, and licensed content saw only 12% active engagement. Field staff could work from outdated protocols for 3–6 months before being retrained.
It reads live work data, such as missed KPIs and ticket resolutions, to spot skill gaps within a minute. AI then compares each gap with a skills graph and builds a personal path of 5-minute lessons in under 2 seconds. Every recommendation explains why that content was assigned.
Performance scores are anonymised for analytics, in line with GDPR. All employee performance data is encrypted at field level and stored in region-specific private cloud environments. Learning activity is kept in an auditable record, which also helps the company show its compliance status at audit time.
If your courses are long, generic and rarely finished, we can help you build short learning paths driven by the work your people actually do.