← All case studies
Elin.ai · 2019–2020 · Berkeley SkyDeck

Customer discovery as product strategy

An early-stage AI company in the Berkeley SkyDeck accelerator with a thesis, a prototype, and the wrong business model. The job was to interview the market into clarity.

250+
Discovery interviews
Structured, worldwide
1
Business-model pivot
Validated by data
First
Partnerships & POCs
Signed post-pivot
01 · Problem

The original go-to-market was anchored to assumptions, not signal. Without a structured discovery system, the team risked scaling an MVP that the market wouldn't buy.

02 · Approach
  • Designed a repeatable discovery framework: targeting, scripts, scoring, synthesis.
  • Ran 250+ interviews across geographies and segments personally.
  • Turned synthesis into a weekly strategic decision document for the founders.
03 · Outcome

The discovery system surfaced a clearly more promising business model and ICP. The MVP was rescoped, the GTM was rewritten, and the first partnerships and POCs were secured against the new positioning.

04 · Lessons learned
  • 1.Discovery is a system, not a phase. Run it every week or you'll guess every quarter.
  • 2.250 interviews are a moat. Most teams stop at 20.
  • 3.A pivot is cheap if the evidence is overwhelming.