alexe-ev Ideate AI Features
Purpose
Help teams generate, evaluate, and frame AI feature ideas that create real user value and are technically feasible.
Skill type
Conceptual skill
Use this skill when
- A team wants to explore how AI can improve their product
- AI feature ideas exist but lack clear value framing
- Leadership is asking "where can we use AI?" and needs a structured answer
- A product area has a clear problem that AI might address
Do not use this skill when
- The goal is model selection or technical AI architecture (use assess-model-capabilities)
- The goal is workflow design for an already-decided AI feature (use design-human-in-loop-workflow)
Required inputs
- Product area or problem to address
- User segment and key pain points
Optional inputs
- Current product capabilities
- Known user pain points from research
- Competitive AI features in the market
- Technical constraints
Upstream context
Works best when:
- User problem is clearly defined
- Product strategy direction is known
- Some technical feasibility context exists
If upstream context is missing
Generate a broad ideation set and flag that prioritization requires feasibility validation.
Downstream handoff
Output can feed:
- assess-model-capabilities (validate feasibility of top ideas)
- design-human-in-loop-workflow (design the execution of chosen AI feature)
- formulate-experiment-hypothesis (test AI feature impact)
Instructions
- Map the user problem or inefficiency the AI feature could address.
- Generate a range of AI feature ideas (breadth first, evaluation second).
- For each idea: describe the user value, the AI mechanism, and rough feasibility.
- Evaluate ideas on: user value, feasibility, differentiation, risk.
- Shortlist top 3 ideas with rationale.
- Identify validation steps before committing.
Output
Provide:
- Problem framing for AI intervention
- Idea list with: user value, AI mechanism, feasibility signal
- Evaluation matrix (user value / feasibility / differentiation / risk)
- Top 3 recommended ideas with rationale
- Validation steps (user research, technical spike, prototype)
- Open questions
Risks / caveats
- "Use AI" is not a feature — the user value must be the starting point
- Feasibility must be checked with engineering before committing
- AI features create user expectations that are hard to roll back — frame carefully