Continuous Discovery in the Age of AI: Webinar Recording + Playbook

Former Uber Eats research leader Jeanette Mellinger sat down with Eran Dror to map a pragmatic path to decision-grade discovery--how to mix qual + quant, where AI fits (and doesn't), and the weekly habits that turn conversations into roadmap-ready moves.

Make AI do the mechanics--you still decide.

Webinar video

Follow our Youtube channel for past webinar recordings →

TL;DR

Discovery isn't extra homework. It's fuel for prioritization under uncertainty.
Qual explains why. Quant sizes how much. Use both before you bet.
Keep three lenses in view: Desirability · Feasibility · Viability.
Run a simple weekly loop: one conversation -> 30-min synthesis -> one decision input.
AI should remove manual work (no transcript tagging). Don't outsource judgment.

Why continuous discovery matters

Shipping faster doesn't mean learning faster. Most teams over-optimize for what's easy to measure and miss the early signals that prevent expensive detours. Treat discovery as a decision problem:

What do we believe right now?
What evidence would change that belief?
What decision will we make differently if it does?

The 3-lens framing for every discovery question

Discovery is about balancing Desirability (do users want it?), Feasibility (can we build it?), and Viability (can we sustain it?).

Jeanette and Eran argue that teams should start with desirability—not to please users blindly, but to validate whether the problem matters to them. Then reality-check against engineering constraints and business models.

Mixing qual and quant: why you need both

Teams ask: "Should we do customer interviews or run A/B tests?" The answer: Both, in sequence.

Qualitative (interviews, observations, usability tests):

Reveals why users struggle, how they think, what motivates them
Spots opportunities you'd never think to A/B test
Uncovers the "jobs" they're hiring your product to do

Quantitative (analytics, surveys, experiments):

Sizes the impact: affects 5% or 50% of users?
Validates hypotheses at scale
Settles debates with hard numbers

The practical rule: Use qual to find promising directions. Use quant to commit resources.

What AI can—and can't—do for discovery

AI transforms the drudgework of discovery:

✅ Where AI helps:

Transcription & tagging: No more hours with spreadsheets
Pattern detection: Surface themes across 100 interviews
Quote extraction: Find the exact user pain point in seconds
Synthesis: Generate insights summaries to share with stakeholders

❌ Where AI doesn't (yet) help:

Strategic judgment: Which opportunities to pursue
Context understanding: The unspoken "why" behind behavior
Trade-off decisions: Balancing user needs vs. business goals
Creative leaps: The breakthrough idea that reframes the problem

Think of AI as your research assistant, not your strategist.

The weekly discovery ritual: keeping it lightweight

Discovery fails when it becomes a quarterly project instead of a weekly habit. Jeanette recommends a simple weekly cadence:

Monday (30 min):

Review last week's learnings. Pick one burning question for this week.

Tuesday-Thursday:

One customer conversation (interview, support call review, or usability test).

Friday (30 min):

Synthesize: What did we learn? What changed our thinking? Share a 3-bullet summary with the team.

This isn't about doing more research—it's about making research a predictable input to decisions.

Chapters & show notes

[0:00] Welcome and intros
[0:30] What is continuous discovery and why it matters
[9:44] The 3-lens framework: Desirability, Feasibility, Viability
[17:16] Qual vs. Quant: When to use each method
[30:12] How AI transforms discovery workflows
[38:45] Common discovery anti-patterns to avoid
[45:20] Building a lightweight weekly discovery habit
[52:30] Tools and tactics for remote discovery
[58:15] Q&A: Stakeholder buy-in and time management
[1:05:40] Q&A: Measuring discovery impact
[1:11:25] Closing thoughts and next steps

Key resources mentioned

Book: Continuous Discovery Habits by Teresa Torres - The foundational text on modern discovery practices
Framework: Jobs-to-Be-Done (JTBD) - For understanding what users "hire" products to do
Tool: Evermuse - AI-powered discovery platform for automating synthesis
Method: Opportunity Solution Tree - Visual framework for connecting user needs to solutions

Action items for teams

This week: Schedule one 30-minute customer conversation. No agenda needed—just listen.
Next sprint: Add a "discovery" column to your team board. Each week, add one learning.
This quarter: Run one qual study before your next major feature. 5 interviews can prevent months of waste.
Ongoing: Share one customer quote in every product review. Make the user's voice impossible to ignore.

About the speakers

Jeanette Mellinger - Former Research Leader at Uber Eats, now consulting on discovery practices. 15+ years helping teams understand users deeply.

Eran Dror - Founder & CEO of Evermuse. Previously led product teams at enterprise SaaS companies. Passionate about making discovery accessible to all product teams.

Want to learn more?

Follow Evermuse on LinkedIn for monthly webinars on product discovery

Book a 1-on-1 discovery assessment with the Evermuse team

Subscribe to our YouTube channel for all webinar recordings

This post was adapted from the November 2024 Evermuse webinar on continuous discovery. Watch the full recording for deeper insights and Q&A.