A framework mapping the scientific method onto lean product development, guiding AI product decisions through vision setting, market analysis, and strategic planning. It addresses SWOT evaluation, moat assessment, and risk mitigation for AI products—particularly the legal, privacy, and safety risks introduced by generative AI and the EU regulatory landscape.
Semantic Classification
Content
- automatically published
Academic science mindset, is business product mindset
- Scientific inquiry maps to product management. Central Role of Product Design Managers: Deciding what to do next.
| Scientific Method | Lean Product Development | General Product Development |
|---|---|---|
| Observe | Build | Where do we want to go (Vision) |
| Hypothesise | Measure | Where are we now (Data/Analysis) |
| Test | Learn | Where should we go next (Strategy) |
-
When I started my formal postgraduate machine learning training risks were couched in biases, now this is Safety and alignment, with the UK positioning itself as a global leader. This likely does have impact on your business goals.
-
Question 1: Where Do We Want to Go?
-
Vision and Goal Setting: Defining a clear destination or North Star.
- My northstar (as an example) is
Equity of opportunity of access to AI, to support a fairer world. -
**Where Are We Coming From?
-
-
Product market analysis** : In your business this is more likely to be a function of your product methodolgy
- I am coming from a position of understanding collaboration in groups, across technology, where some members of the group are likely to be AI.
- Where Should We Go Next?
-
Strategic Planning: Deciding the next steps based on vision and current status.
- For me, this mean helping build B2B capabilities,
- Education, and interweaving of people and AI through storytelling,
- Distributed, global, AI enabled infrastructure,
- Clearly communicating why,
- Building communities to help.
-
This won’t be you and your company, but this is a great time to work out these checklists.
- SWOT Analysis: Evaluating strengths, weaknesses, opportunities, and threats.
- I am trying to build this bit of the business journey through these presentations
Everyone’s pivoting to generative AI.
https://twitter.com/petergyang/status/1667539634577092609
- Link to Tweet
- But my alarm bells go off when I see:
- ? A crowded landscape
- ? FOMO driven decision making
- ? Sky high valuations for an early space
- If you took the word “AI” out, is the product still solving a customer problem?
- AI is a solution, not a problem. Ask yourself:
- What is the pain point?
- How many users share this pain?
- Is the pain big enough to take action?
- Is the pain underserved by non-AI tools?
- How accurate does the solution need to be?
- Plot the problem on a fluency vs. accuracy grid.
- Gen AI today is great for high fluency + low accuracy problems (e.g., productivity).
- It’s not great for solutions that need high accuracy (e.g., financial decisions).
- How fast will incumbents move?
- Incumbents like Microsoft, Google, and Adobe have moved incredibly fast on AI.
- Startups that overlap with core incumbent use cases might struggle.
- AI presentation startups need to be MUCH better than AI in Powerpoint to thrive.
- Is there a moat? Examples moats include:
- Access to proprietary data and models
- Exclusive contracts with large customers
- Great product even without AI
- Exceptional talent in the selected field
- Business models that incumbents avoid
- And of course…speed of execution.
- Is it overvalued?
- If an AI product already has $100M+ valuation, you should think:
- Can it continue to grow and (more importantly) retain users?
- In a crowded space like AI copywriting and productivity
- that could get hard.
- To recap, here are 5 questions to ask to evaluate AI products and companies:
-
- Without “AI”, is it still solving a problem?
- How accurate does the solution need to be?
- How fast will incumbents move?
- Is there a moat?
- Is it overvalued? 7/ I hope these questions also help builders who are thinking of creating new AI products.
-