SKILL FILE

Discovery Plan with AI

Map a 4-week customer research sprint before you start building. Turns a fuzzy "we should talk to users" instinct into a week-by-week plan with goals, methods, sample sizes, and decision points.

Download Skill File ↓

What this skill file teaches Claude

Drop one markdown file into your repo. Claude Code learns how to run this entire workflow.

1

Pick the 5 right people

A short brief on who to talk to, why, and the 1-line screener that filters time-wasters out. The hardest part is picking; everything else flows.

2

A 5-question interview guide

Behaviour-based prompts, not opinion questions. Designed to fit a 30-minute call with room to follow the thread.

3

Find + book the 5 in a day or two

Where to DM, what to say, how to qualify fast. Usually 20 DMs nets 5 interviews. The skill includes the DM templates.

4

Synthesise live, not later

Patterns emerge inside the conversations. A short doc per interview + a 30-minute synthesis at the end is plenty.

5

Decision point built in

Three pre-committed outcomes: build the thing, change the offer, change the audience. No "more research" loophole.

What you can build with this

Validate a new offer before building

5 calls in 3 days will tell you if the problem you're solving is real. If they're polite but not pulling at the offer, change it.

Pre-MVP signal check

Before writing any code, run 5 conversations + show a Figma or doc. If 3 of 5 want to buy, build it. If 1 of 5 does, pivot.

New segment exploration

Considering a different audience? Talk to 5 of them this week. Decide on Friday.

Get the full skill file

Everything above is 80% of the skill file. Download the complete version with full implementation details, agent prompts, and ready-to-run scripts.

Common questions

5 is the threshold where you start hearing the same things back. Jakob Nielsen's UX research and Rob Fitzpatrick's Mom Test both land on the same number for similar reasons. After 5 you're usually collecting confirmation, not new patterns. If two segments matter, do 5 per segment.
If you push it: 3 days. Day 1: write the screener + DM 20 people. Day 2: line up + run the 5 calls. Day 3: 30-minute synthesis + decide. If you're juggling a job, stretch it across a week. Either way you're not waiting 4 weeks.
If 20 well-targeted DMs can't get you 5 conversations, that's already a signal. Either the audience is wrong or the topic doesn't pull. Treat the recruiting itself as data.
No. Surveys answer "how many" questions; interviews answer "why" questions. Early validation is a why question. Skip the survey, save the week.
Yes if the guide is tight. 5 behaviour-based questions, time to follow up on the most interesting thread. Longer interviews don't produce more signal, just more transcript.

Ready to automate?

SoloStack gives you every skill pre-installed — scraping, marketing, sales, CRM, and more. One repo. Every department.

Book a Call →