“Fall in love with the problem, not the solution.” – Ash Maurya
The Opportunity-Solution Tree (OST) provides a visual and logical structure for connecting user pain points to potential solutions and experiments. With AI, PMs can now scale this framework beyond whiteboards and gut-feel ideation.
Why OST Works: It fosters disciplined problem discovery, prevents premature convergence on solutions, and provides traceability from problem space to product delivery.
How AI Elevates It:
- Data Mining: NLP models analyze Zendesk tickets, G2 reviews, App Store feedback, and internal call transcripts to surface underexplored opportunity areas.
- Clustering: Unsupervised models group complaints by latent topics (e.g., onboarding, speed, integrations), creating layered branches in the opportunity tree.
- Framing: Generative models assist PMs in rephrasing opportunities in user-centric language (e.g., “Users struggle with X because Y happens when Z”).
Expanded Use Cases:
- Fireflies.ai + Miro integration that converts customer interview transcripts into opportunity trees.
- OpenAI plug-ins embedded in Retool dashboards for ops teams to elevate issues into PM triage.
- NPS segmentation: AI automatically links low NPS drivers to specific features or workflows.
- OST inside productboard: Opportunity paths auto-generated from user pain clustering.
- Voice of Customer AI tools: Like Wonderflow and Chisel, that feed structured OST diagrams.
Tools:
- Miro, Productboard, Useberry with LLM plugins
- Thematic, MonkeyLearn for text clustering
- AI-based customer insight tools like Wonderflow, Chisel, Idiomatic
Case Example: Duolingo uses internal GPT-4 fine-tuned agents to sift through in-app complaints and generate structured insights for their OST. Zendesk launched internal AI-based problem-mapping layers that suggest solutions tied to existing platform capabilities.
Risks & Watchouts: Blind reliance on AI may cause over-prioritization of high-frequency but low-severity issues. Weighting mechanisms and product judgment must be layered in. Additionally, some user problems only reveal themselves in direct conversations—not in data.
Tags: #AIUXResearch, #AIUXTools, #CustomerInsights, #Duolingo, #OpportunitySolutionTree, #ProblemDiscovery, #Productboard, #VoiceOfCustomer