As of 13 August 2026, AI can plan an MVP for your UK startup.

What to actually do

1. Use a tool built for this

   The route this page recommends

2. Do it yourself

   Second choice

   A chat interface, chat-fluent skill, and roughly 30 minutes until you can act on the result.

   How to actually do it

   1. Open a document and write the startup idea, target customer, problem, current alternatives, evidence, team skills, budget, timing and known UK constraints.
   2. Paste that information into the prompt, replacing every bracketed slot and leaving unknown items marked as unknown rather than guessing.
   3. Paste the completed prompt into ChatGPT, Claude or Gemini and ask for the MVP plan.
   4. Copy the proposed problem, user outcome, features, exclusions and assumptions into your working document, keeping supplied facts separate from AI assumptions.
   5. Compare the proposed scope with your available people, technical skills, budget and delivery constraints, then remove any feature you cannot support or validate.
   6. Use the validation plan to speak to potential customers or test a prototype, and record the evidence rather than treating the AI plan as market proof.
   7. Regenerate the plan with the evidence from those tests and make your own build, pause or change decision from the results.

   Prompt

   Act as a practical UK startup product manager. Help me plan the smallest credible MVP for this business idea.
   
   Business idea: [describe the idea]
   Target customer: [describe the customer and, if known, the buyer]
   Problem: [describe the problem and how customers handle it now]
   Evidence so far: [interviews, usage, sales, waiting lists, research or none]
   Proposed value: [what the product should help the customer do]
   Available team and skills: [people, roles and technical skills]
   Budget and time constraints: [constraints]
   Business model: [how the startup may make money]
   Known competitors or alternatives: [list them or say unknown]
   UK-specific constraints: [regulatory, data, sector or location constraints, or none known]
   
   Produce a concise MVP plan containing:
   1. The customer problem and the riskiest assumptions.
   2. A single recommended MVP user outcome.
   3. Must-have features, explicitly excluded features and the reason for each choice.
   4. The simplest plausible user journey.
   5. Technical, operational, data protection and commercial risks to investigate, without giving legal advice.
   6. A validation plan using specific customer conversations, tests or prototypes before substantial build work.
   7. Measures that would show whether the MVP is worth continuing, with no invented targets or market figures.
   8. A build sequence, dependencies and decision points.
   9. Questions that only I can answer.
   
   Separate facts I supplied from assumptions you introduced. Do not invent customer evidence, market size, competitor facts, costs, regulations or success rates. When information is missing, mark it as an assumption or question. Challenge the idea and recommend a smaller scope where justified. End with a one-page decision summary stating what to build, for whom, what not to build yet and what evidence is still needed.

   Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.

3. Hand it to a person

   The distant third

   A person who owns the outcome does this end to end, worth it when the failure is dear.