ai-differentiator-solution-ideation

Ideates 3 distinct AI-native solution directions for the chosen AI differentiator need (AI wedge) through three lenses — competitor, future, and connect-the-dots — so the team doesn't fall in love wit

SKILL.md

GitHub

AI Differentiator Solution Ideation

Purpose

Prevent the most common ideation mistake: latching onto the first solution idea that comes to mind and never comparing it against alternatives. This skill forces the ideation of three genuinely different AI-native solution directions for one chosen AI differentiator need (AI wedge), using three different lenses so the directions aren't just the same idea in different words.

Based on

  • The methodology of an external "AI-first SaaS Product" workshop, applied by the owner to one own case — see ../../references/ai-first-saas-workshop-source.md and the worked example ../../cases/ai-decision-coach-mvp-case.md, section 5. Note: applied only once so far (the owner's own case) — not broadly validated.
  • The email-marketing personalization example from the workshop illustrates the "genuinely different" requirement: 1:1 history-based personalization vs. 1:1 contextual social personalization vs. 1:group cluster-based personalization — three AI-native approaches to the same need (relevance), not three variations on the same idea.

Method (draft — to be filled in further)

  1. Start from one chosen AI wedge (from ../customer-vision-to-jtbd/SKILL.md) — don't ideate for more than one need at a time, it breaks the comparison.
  2. Ideate solution direction 1: the competitor lens. Ask: how do existing players (TTOs, incubators, generic AI tools, consultants) solve this need today, and what would AI make possible that ISN'T possible for them (scale, continuity, personalization, speed)? Goal: differentiation from existing competition.
  3. Ideate solution direction 2: the future lens. Ask: how was this need solved BEFORE (via workshops, templates, sticky notes, frameworks applied solo — a heavy, energy-draining process that always has to be refitted to context), and how will it be solved in the FUTURE, when context is continuously built and taught to AI, an AI mentor asks the right questions to reach shared understanding, and the user is in continuous interaction with AI that keeps producing small building blocks (e.g. an ICP definition, a value proposition, a revenue-model hypothesis)? Goal: imagine a solution that doesn't just digitize the old process but changes the shape of the process itself.
  4. Ideate solution direction 3: the connect-the-dots lens. Ask: what other separate tasks/activities (a+b+c+d) does this user do around this need that could be COMBINED into a single AI-native experience? Goal: a solution that emerges from combining several separate steps, not from speeding up one step.
  5. For each of the three directions, capture: a name, the AI-native concept (2-3 sentences — how it works), the primary output for the user, and why it's distinctive relative to competitors/the status quo (not replicable with a static checklist or a generic LLM prompt).
  6. Compare the three directions side by side in a table: core value / why it fits the chosen AI wedge. Resist the urge to pick right away — take all three to ../rice-scoring-and-mvp-synthesis/SKILL.md for objective scoring before choosing.

What this skill does NOT do

  • Does not choose the MVP for you — it produces three comparable alternatives; the choice is made in the next skill via RICE scoring.
  • Does not assess technical feasibility or build cost — only the distinctiveness of the concept. Feasibility/effort is assessed in ../rice-scoring-and-mvp-synthesis/SKILL.md.
  • Does not guarantee that all three directions are equally good — it deliberately also produces weaker/riskier options, so the comparison is genuine rather than artificial.

Refinement notes

This skill has so far been applied to one case (the owner's own case). As you apply it to more businesses, add:

  • your own additional lenses or variations on the three lenses that have proven useful
  • typical pitfalls (e.g. when the three directions end up too similar, and how to notice that)
  • concrete examples from other cases in the ../../cases/ folder

Once this section has been filled in with multiple cases, raise skills_index.json's maturity field to validated (see ../../../../meta/maturity_levels.md).

Continue from here

  • Preceding skill in this pack: ../customer-vision-to-jtbd/SKILL.md — produces the AI wedge that is ideated on here.
  • Next skill in this pack: ../rice-scoring-and-mvp-synthesis/SKILL.md — scores the three directions and selects the MVP.
  • Worked example: ../../cases/ai-decision-coach-mvp-case.md, section 5.
  • The pack's shared guardrails: ../../CLAUDE.md

References

  • ../../references/ai-first-saas-workshop-source.md — source information
  • ../../cases/ai-decision-coach-mvp-case.md — worked example
  • ../../CLAUDE.md — the pack's shared guardrails