Interim product and commercial leadership for an AI venture, delivered as a contracted engagement.
I co-founded Lumination AI, a UK company building AI products for education, training and knowledge management. My ongoing role there is contracted through Azvai, covering product direction and commercial development rather than a full-time executive seat.
The portfolio now serves more than 90,000 users a month, and the B2B side has signed contracts worth over £180,000 a year.
What the role covers
- Product direction. Deciding what gets built, in what order, and what gets retired. Coordinating engineering, design and content across a distributed team.
- B2B business development. Building the institutional side of the business from first conversations through to signed agreements.
- Commercial and compliance groundwork. Pricing, data residency decisions, security posture and the documentation that procurement teams ask for.
The products
- AI Tutor. Course creation and tutoring, around 25,000 monthly users, with a white-label version launched in 2026 for academies and training providers.
- SMMRY. A summariser handling any format or length, around 30,000 monthly users.
- Socratic Owl. Scan-and-solve step-by-step problem solving on iOS and Android, around 20,000 monthly users.
- NoteGPT and Classx. Note summarising and classroom tooling, rounding out the portfolio.
Underneath them sits a shared inference platform with curriculum-grounded retrieval, subject routing, moderation and guardrails, usage analytics, and multi-language support. Organisations adopt it three ways: a full white-label product, the API on its own, or a plugin inside their existing learning management system.
Selling AI to institutions
Schools, universities and training providers buy differently from consumers. The deals turn on evidence rather than enthusiasm, and most of the effort goes into questions that have nothing to do with the model.
- Where the data lives. User accounts, content and hosting sit on UK infrastructure, which decides procurement conversations with UK institutions before features do.
- Security and data protection posture. Cyber Essentials certification and ICO registration, ready for vendor questionnaires.
- Predictable cost. Per-student pricing and usage tracking, so a finance team can budget a year ahead.
- Integration reality. Whether it works inside the systems teachers and trainers already use, which is why the LMS plugin path exists.
Building the product is the smaller half. Getting an AI tool through an institution’s procurement, data protection review and integration constraints is where most promising pilots quietly die, and learning to clear that path is the transferable skill.
The same role for other ventures
This is the shape of engagement Azvai offers elsewhere. Organisations bring us in when they need senior product or commercial leadership without hiring a full-time executive, whether that is an interim director covering a gap, a fractional seat alongside a founding team, or a mandate to take a product to institutional buyers.
It suits small ventures and consultancies that have capable delivery people and no one whose job is to decide what gets built, price it properly, and carry it through a client’s procurement process.
Why it fits this practice
Advising on AI adoption without having shipped an AI product leaves gaps that show up quickly in a client conversation. Running this venture means the advice comes from having priced inference, argued about data residency in a procurement review, watched users abandon a feature that tested well, and carried the cost of getting it wrong.
That is the foundation under the operating model in what an AI-augmented consultant actually does. To discuss an interim or fractional leadership engagement, see our services or get in touch.
Related work in this thread: the TutorAI learning platform and AI product prototyping.