Generative AI making personalised learning accessible at scale. One of several threads in this practice. See the systems-change thesis.
AI Tutor is a generative AI learning platform built for corporate training and development. Azvai co-developed it end to end, from market research through to implementation, covering AI consulting, user experience design, marketing, sales and product management.
The platform serves over 30,000 users every month.
What TutorAI.me does
TutorAI.me uses generative AI to structure and create courses automatically, tailored to what each user needs. That speeds up the production of training content and personalises it, which matters most for organisations whose knowledge is scattered across documents, systems and individual people.
What we contributed
- Product management. Coordinating interdisciplinary teams against corporate objectives.
- Market research. Identifying real needs and trends in the training sector.
- AI consulting. Strategy and selection of the generative technologies, for scalability and cost efficiency.
- UX research. Making the platform accessible to training teams rather than to technologists.
- Marketing and sales. Positioning in a crowded learning market.
- Rapid prototyping. Building prototypes, plugins and integrations for e-learning systems in hours.
Grounded in usage data
Product decisions drew on published research rather than assumption. We analysed the Anthropic Economic Index, covering usage data from thousands of professionals, to understand how educators actually use generative AI across their tasks.
Pie chart. Data table with 6 rows and 2 columns follows.
| Course content | 34.53% |
|---|---|
| Instruct | 31.7% |
| Assess | 12.62% |
| Student support | 10.3% |
| Communications and coordination | 6.49% |
| Data and information management | 4.36% |
What building this kind of platform takes
Organisations considering an AI tutoring or training platform usually underestimate the same things, so it is worth naming them.
- Getting the knowledge in. Source material arrives as documents, slides, recordings and things people simply know. Pulling that into a form the model can teach from is most of the work.
- Keeping content current. Generated courses go stale like any others. The platform needs a path for updating material as the underlying knowledge changes.
- Data residency and privacy. Corporate training material is sensitive. TutorAI runs on Microsoft Azure infrastructure in Europe, under GDPR.
- Cost control at scale. Token usage compounds quickly with tens of thousands of monthly users. Tracking and rate limiting belong in the design from the start, not after the first invoice.
A worked example of the content side is in building custom courses from your own material.
Why it fits this practice
Knowledge management is a systems problem before it is a technology one. Information sits in the wrong places, training goes stale, and people rebuild the same understanding repeatedly. Building a platform that addresses that at scale, with data held on European infrastructure under GDPR, taught us where generative AI genuinely helps in learning and where it quietly makes things worse. That experience is what we bring to consulting engagements on AI capability and training.
Related work in this thread: AI product prototyping for learning and development. To discuss an AI learning platform or training programme for your own organisation, get in touch.