AI UX for Strategy
Overview
Strategy is the UX phase most likely to be under-resourced and over-rushed. AI compresses the data-gathering and synthesis tasks that typically consume 60% of strategy time. Perplexity for competitive intelligence (cited, verifiable), Claude for brief synthesis and assumption stress-testing, Miro AI for post-session clustering. The human judgment layer stays in framing the right questions and prioritising what matters to this client, at this moment, in this market.
AI workflow
9 steps
- 01
Stakeholder Note Synthesis
Claude
Paste raw stakeholder interview notes or meeting transcripts. Prompt: "Extract the key tensions, priority signals, unspoken assumptions, and success criteria from these notes. Structure as a problem brief." Review critically - AI surfaces patterns but may miss political subtext.
- 02
Competitive Landscape Research
Perplexity
Use Perplexity (not ChatGPT) for competitive research. It cites sources in real time. Prompt: "Analyse the competitive landscape for [product category] in [market]. Include pricing models, key differentiators, user sentiment from reviews, and recent product changes." Verify all claims.
- 03
Assumption Mapping
Claude
List your core project assumptions. Prompt Claude: "For each assumption, what is the strongest counter-argument? What data would disprove it? Rate each assumption by risk level." This de-risks strategy before design begins.
- 04
Opportunity Space Mapping
Miro AI
After a human-led workshop, paste sticky note content into Miro AI for automated clustering. Use the output as a starting point, not a final answer. AI clustering misses nuance - always do one human review pass.
- 05
Problem Brief Drafting
Claude + Notion AI
Draft the formal problem brief using Claude as a writing partner. Use Notion AI to restructure for client-readiness. Include: problem statement, user segment, success criteria, out-of-scope boundaries, and key assumptions.
- 06
HMW Question Generation
Claude
Prompt: "Generate 15 How Might We questions from this problem brief, ranging from narrow execution to broad systemic reframing." Use the widest ones as creative provocations in ideation sessions.
- 07
Strategy Stress-Test
Claude
Final check: "What are the three most likely ways this strategy fails? What early warning signals should we track? What assumptions are we most likely to be wrong about?" Build these into your project risk register.
- 08
High-Impact Decision Triage
Claude + Notion
Not every design decision deserves the same time. Sort the open decisions by two questions: how many users and screens does it touch, and how hard is it to undo? One-way doors - navigation model, information architecture, core data model, pricing and plan structure, onboarding flow, design system foundations - get research, options and a review. Two-way doors - copy, colour of a secondary button, icon choice - get decided in minutes and tested live. Prompt Claude: "Here are the open design decisions on this project. Classify each by reach and reversibility, and tell me which three deserve the most time this sprint."
- 09
Decision Log
Notion
For every one-way-door decision, write five lines: the decision, the options considered, the evidence, the principle it rests on, and what would make you revisit it. Months later, when a stakeholder asks why the navigation works the way it does, the answer is a link rather than a meeting - and a new designer inherits the reasoning, not only the result.
Tools
| Tool | Use | Free tier | Client-safe |
|---|---|---|---|
| Perplexity | Cited competitive research, market landscape | Free tier available | Yes |
| Claude | Brief synthesis, assumption stress-testing, HMW generation | Free tier available | Yes |
| Miro AI | Opportunity clustering post-workshop | 3 boards free | Yes |
| Notion AI | Client-ready strategy documentation | Free tier available | Yes |
| QoQo | Persona and problem statement generation | Free tier available | Yes |
| Kraftful | Feedback and review synthesis for existing products | Free tier available | Yes |
Constraints
5 constraints
- Honest Assessment: Perplexity Pro ($20/month) significantly outperforms the free tier for competitive research depth and source quality. The free tier is useful but limited.
- Fixed Bid: Cap strategy at 2 days maximum. AI compresses competitive research from 8 hours to 90 minutes. Use the saved time for stakeholder alignment, not more research.
- Consultancy: All strategy outputs should be client-presentable. Use Perplexity citations as evidence, not AI-generated text without sources.
- Warning: ChatGPT and Claude will fabricate competitor pricing, feature lists, and market share data. Always verify with primary sources.
- Time spent is not the same as impact. Teams routinely spend a week on a button style and an afternoon on the navigation model. Put the effort where reversibility is lowest and reach is widest.