MVP scope, business case, and technical co-founder strategy — written honestly for a solo design-led founder building something that genuinely matters.
Where Bloom is today
What Bloom has already, what the full vision requires, and how we sequence from here to product — with investment.
The hardest part. A fully-formed product vision, a design system, UX flows, curriculum framework, AI architecture, and the pedagogical reasoning behind every decision. All of it mapped before any build spend.
A team of 8–10 engineers, 18 months minimum, and significant funding. The AI architecture as designed is sophisticated. The knowledge graph alone is a multi-month engineering project. This is the version we're building toward — not the version we're starting with.
We don't need the full system to prove Bloom works. We need a version that creates the right feeling for a child and their family — with enough real personalisation to demonstrate the concept. That version is achievable with investment. Then we use it to raise and hire.
Section 1 — MVP
Not a simplified version of the full product. A carefully chosen slice that proves the core thesis: that an AI learning service can make a neurodivergent child feel genuinely seen, safe, and capable. Everything else follows from proving that.
One child. One subject. One session. Does it feel different? Does the child feel safe? Does the parent feel informed? Does the AI respond as if it actually understands this specific child? If the answer to all four is yes — Bloom has something real to build on.
The child experience must be complete, warm, and dignified from day one. There is no "we'll improve it later" for the child-facing product. If it feels like a beta, parents will not trust it with their vulnerable child.
Parents are Bloom's true customer in the MVP. They pay. They decide. They will recommend Bloom to every parent in their network if it works — and the SEN parent network is exceptionally vocal and connected. Make the parent experience a product in itself.
The MVP AI is far simpler than the full architecture. That is the right decision. We build sophistication on top of real data, not on top of assumptions. Start with something that works well. Learn. Then build what the data tells us to build.
Safeguarding is not a Phase 2 feature. It is the price of admission to working with children. No child touches this product without it in place. This is both ethical and legal.
Being clear about what we're not building is as important as what we are. Scope creep is how solo founders fail. These are real features that will be built — just not first.
Section 2 — Build Phases
Four phases. Each one has a clear output, a clear milestone, and a clear decision point. You never move to the next phase until the current one is done — and done means real children using it, not "it's mostly ready."
Everything that doesn't require code. Build the product soul before the product itself. This phase ends when we have a technical co-founder committed — or strong enough prototype evidence to raise a pre-seed to hire one.
Build the real product with real children. Closed, invited, consenting families only. No public launch. Every session observed or reviewed. Weekly iterations. The goal is not perfection — it is learning fast enough to know what to build next.
Expand carefully. Add the educator view. Add a second subject. Begin building real adaptive capability using the session data from Phase 1. Raise a seed round on the back of Phase 1 evidence.
The full product vision starts to become real. Self-hosted models. LSTM emotion detection trained on real session data. IRT knowledge graph. First school partnerships. Series A raise to build the team properly.
Built on UK clinical evidence, a trusted brand, and a validated playbook. The US requires curriculum localisation and IEP reporting — bounded, planned work. The AI architecture, pedagogy, and experience design transfer entirely. IDEA's strong enforcement and the scale of the US SEN market make this a compelling next step.
Section 3 — Business Case
Not a charity. Not a side project. A defensible, scalable commercial product with genuine competitive moats — built on top of something that actually helps children. The two are not in conflict.
Section 4 — Market
Three distinct customer segments, each with different willingness to pay, different buying journeys, and different value propositions. The MVP focuses entirely on one.
Market Context
Bloom launches in the UK. Not because the UK is the only market — but because it is the right market to prove the model, build the evidence base, and establish the trust that opens every other market.
Section 5 — Revenue Model
Three revenue streams, unit economics that work, and a path to profitability that doesn't require becoming a venture-scale machine to survive.
Section 6 — Technical Co-Founder
Not just a developer. A true co-founder — someone who owns the technical vision the way you own the product and design vision. The wrong hire here is worse than no hire. Take the time to find the right person.
Someone who has shipped a production web application before. Ideally with experience in consumer products (not just internal tools), some exposure to AI/LLM integration, and a personal connection to the mission — a neurodivergent family member, a background in education, or a deep belief in equitable access to good learning.
The wrong technical co-founder will slow you down, create technical debt that takes years to unwind, or misalign your incentives at the worst moment. Better to take longer to find the right person than to move fast with the wrong one.
Section 7 — Action Plan
Not a vague roadmap. A concrete list of things to do, in order, that builds momentum and ends with a technical co-founder committed or a pre-seed round in progress. Most of this is work Bloom can progress now, before writing a line of code.
The unglamorous but necessary work that lets everything else happen.
Your assumptions about what SEN parents need will be partially wrong. Find out which parts before committing them to a design system.
Work that can be done now, with Claude, that will be handed directly to your technical co-founder as a complete specification.
The pitch deck is also your co-founder recruitment document. Build it to do both jobs.
This is the most important thing you do in Phase 0. Do not rush it. Do not settle. But also do not let perfect be the enemy of good — a committed, capable, mission-aligned engineer is what you need.