Customer discovery and jobs-to-be-done framing for AI features
It is the systematic process of identifying user struggles, desired outcomes, and contextual constraints to define the functional, emotional, and social 'jobs' an AI feature must accomplish for a specific user in a specific situation.
This skill is the primary defense against building technically impressive but commercially irrelevant AI features, directly increasing product-market fit and user adoption. It translates ambiguous user needs into precise technical requirements, ensuring engineering resources are focused on high-impact work that drives retention and monetization.
1Careers
1Categories
9.2 Avg Demand
20%
Avg AI Risk
How to Learn Customer discovery and jobs-to-be-done framing for AI features
Focus on: 1) Mastering the core JTBD interview format ('Tell me about the last time you...') to uncover the real struggle. 2) Learning to separate solutions (a 'faster button') from the underlying job ('I need to feel confident I haven't missed a critical alert'). 3) Practicing the 'Job Statement' formula: [Action verb] + [Object of action] + [Context/Constraint] + [Optional: Emotional/Social payoff].
Move to practice by conducting live customer interviews for a specific feature, not a general product. Learn to map the 'Job Steps' (Define, Locate, Prepare, Confirm, Execute, Monitor) and identify where current solutions create anxiety or anxiety-driven workarounds. A common mistake is leading the witness by mentioning AI or a specific technology, which biases the discovery toward a solution rather than the problem.
Mastery involves synthesizing job statements across user segments to identify 'over-served' and 'under-served' job steps, informing a product roadmap. You must be able to frame AI's unique value proposition: Does it automate a tedious step (Execute), provide predictive confidence (Monitor), or surface previously invisible data (Locate)? This requires aligning the discovered jobs with business metrics and managing stakeholder expectations about AI's probabilistic nature.
Practice Projects
Beginner
Case Study/Exercise
Deconstruct a Mundane AI Feature
Scenario
A product manager has specified a 'smart search' feature for a B2B CRM that auto-suggests contacts based on partial names and past interactions.
How to Execute
1. Write down the assumed job: 'Quickly find the right contact.' 2. Conduct 5 hypothetical interviews (using a colleague role-playing a sales rep) using the prompt: 'Tell me about the last time you had trouble finding a contact in the system. What were you trying to do? What was at stake?' 3. Rewrite the job statement based on findings (e.g., 'Re-establish context for a high-stakes outreach after time away, under time pressure'). 4. List 3 AI-specific questions this job raises (e.g., 'How does the AI know what 'context' means for me?').
Intermediate
Case Study/Exercise
Job Mapping & Anxiety Point Analysis
Scenario
You are tasked with improving an existing AI-powered content moderation dashboard for social media moderators. Moderators complain it 'misses things' and 'flags too much noise'.
How to Execute
1. Map the job steps for a moderator: Prepare (gather queue), Execute (review item), Confirm (make decision), Monitor (track queue health). 2. Conduct interviews focused on the 'Confirm' step. Probe for: 'What specific information do you need to feel 100% confident in your decision? When do you hesitate?' 3. Identify the core job: 'Make a consistent, defensible moderation decision without second-guessing, at scale.' 4. Propose an AI feature framed to reduce anxiety at the 'Confirm' step (e.g., an AI-generated 'confidence score' with a cited precedent case).
Advanced
Case Study/Exercise
Strategic Job Portfolio Analysis for an AI Suite
Scenario
You are a product lead for a fintech company planning a suite of AI features for small business owners, including expense tracking, invoice generation, and cash flow forecasting.
How to Execute
1. Conduct discovery to uncover the overarching emotional job: 'Feel in control and empowered about my business finances, not embarrassed or anxious.' 2. Break this into functional jobs (e.g., 'Get paid faster,' 'Reduce time on bookkeeping,' 'Anticipate a cash crunch'). 3. For each functional job, analyze current 'job executors' (e.g., the owner, a bookkeeper, QuickBooks). 4. Define the AI's unique role for each: Does it fully automate (Execute), augment a human (Prepare/Confirm), or provide strategic insight (Monitor)? 5. Create a prioritization matrix based on job importance vs. current solution satisfaction.
Tools & Frameworks
Mental Models & Methodologies
Jobs-to-be-Done Framework (Ulwick/Moesta)Outcome-Driven Innovation (ODI)Customer Forces Progress Model (Struggling Moment -> Push/Pull -> Anxiety/Inertia -> Resolution)
Use JTBD and ODI to structure interviews and quantify outcome importance/satisfaction. The Forces model is critical for understanding why a user would switch to a new AI solution, mapping the emotional and rational drivers of adoption.
Transcription tools allow for verbatim analysis. Affinity diagramming clusters interview insights into themes. The Job Statement and Job Step Map templates are non-negotiable for structuring and communicating findings from raw data to actionable product specs.
Interview Questions
Answer Strategy
The candidate must reject the solution and reframe it as a job discovery problem. They should outline a specific plan to interview project managers about the last time they gathered and reported status. The strategy is to uncover the underlying job (e.g., 'Provide stakeholders with credible, timely evidence of progress to maintain trust without derailing my team's workflow'). A strong answer will mention specific interview questions and how the job statement might differ from the initial 'automate status' pitch.
Answer Strategy
This tests for intellectual humility and practical corrective action. The answer should follow the STAR method but be anchored in the discovery process. Look for: 1) Recognition of a disconnect (e.g., low adoption, negative feedback). 2) A return to user research, specifically job-based interviews. 3) A concrete pivot or refinement of the feature's framing based on the new understanding. The professional response should show how the candidate used the failure to build a more robust discovery practice.
Careers That Require Customer discovery and jobs-to-be-done framing for AI features