Validate the idea before you build it

This is stage two of the five-stage founder journey: you arrive with a written wedge from stage one, and you leave with evidence — or with a dead idea that cost you a month instead of a year. Validation is the work of finding out whether anyone will pay before you spend months building. It feels slower than starting to code, and it is almost always faster, because the most expensive way to test an idea is to ship it.

Ask about the past, not the future

Start with conversations, and have them with people who live the problem, not friends who will be kind to you. The list of twenty or more people and places you wrote in stage one is where those conversations come from. The mistake that ruins most of them is asking about the future — "would you use this" invites politeness, and politeness is worthless data. Ask about the past instead, because past behavior is the only thing a person cannot exaggerate to spare your feelings.

Concretely, questions like these do the work:

• "Walk me through the last time this problem cost you an afternoon. What exactly happened?"
• "What do you do about it today, and what have you already tried before that?"
• "What has dealing with this cost you so far, in money or in hours?"
• "When it went wrong last time, what broke downstream?"

Someone who has cobbled together a spreadsheet, hired help, or paid for a tool that disappoints them is showing you demand. Someone who says the idea sounds great but has never lifted a finger against the problem is showing you the opposite, however warm the conversation felt. Task Machine's "Discovery Interview Kit" playbook keeps this discipline on rails: it installs the interview guides, the outreach drafts that get conversations booked, and the note templates that make each interview comparable to the last.

Sell it before you build it, one rung at a time

Interest becomes evidence only when something is at stake, so put a price in front of people early and climb an escalation ladder where each rung raises the stakes:

1. A waitlist with a stated price. "Join the list — it will cost about two hundred a month at launch" tests far more than a bare email signup, because a stated price makes signing up a small commitment instead of a costless courtesy.

2. A preorder or a paid pilot. Ask the warmest conversations to prepay a discounted first month or schedule a paid pilot with a start date. Money moving, even a little of it, is a categorically stronger signal than any number of enthusiastic replies.

3. A concierge test. Deliver the outcome manually to a handful of paying customers before any product exists — you are the software, running the process by hand behind a thin front. This rung tests the most, because it earns revenue while teaching you exactly what the product must do, in what order, and what customers actually check when the work comes back.

If you cannot get anyone to commit money, time, or a scheduled call, more building will not fix that. The "Experiment & Assumption Sprint" playbook helps you run this ladder deliberately: it turns the beliefs your wedge depends on into a ranked list of assumptions, pairs each with the cheapest experiment that would test it, and tracks what each experiment actually showed.

Write the kill criterion before the first conversation

Ideas rarely die of bad results. They die of drift, months after the results were already in, because there was never a moment when anyone had to look at the evidence and decide. Guard against that by writing a kill criterion before you start: a concrete threshold and a date. For example — "by the last Friday of the month, twenty problem conversations done and at least three people who prepaid or scheduled a paid pilot, or the idea changes." If the threshold is not met on the date, you stop or change the idea. Deciding the bar while you are still objective protects you from the sunk-cost version of yourself who will exist four weeks from now.

Run validation as a four-week sprint

Validation without a shape sprawls, so give it one. A month is enough for a wedge this narrow:

1. Week one — book the conversations. Turn the stage-one list into scheduled calls, aiming for ten booked by Friday. This is pure outreach volume, and it is where the "Idea Validation Sprint" playbook earns its place: agents build the prospect list, draft each outreach message, and draft the follow-ups for everyone who goes quiet.

2. Weeks two and three — get to twenty conversations, then make the offer. Run the past-behavior interviews, keep notes in one place, and by the end of week three put the concierge offer or paid pilot in front of the warmest people you spoke with. Selling to people who just described the problem to you is the fairest test the idea will ever get.

3. Week four — decide against the kill criterion. Tally the evidence on the date you wrote down, with no extensions and no "one more week." The idea clears the bar, changes shape based on what you heard, or dies. All three outcomes are wins compared to building in the dark.

Through all four weeks, the drafting and chasing run as supervised agent work. Agents in Task Machine start supervised, so every outbound message an agent drafts waits in your inbox for approval before anything is sent — your name stays on every word while the pipeline-keeping stops consuming your evenings. And as interview notes pile up, the "Customer Discovery Synthesizer" playbook does the job founders always skip: it reads across all the conversations and surfaces the repeated phrases, the workarounds mentioned more than once, and the moments people leaned in, so week four's decision rests on patterns rather than on whichever call you remember most vividly.

Evidence is the exit, not enthusiasm

The temptation at the end of a good sprint is to count compliments. Do not. Count commitments.

You are ready for the next stage when twenty problem conversations are done, at least three people have prepaid or scheduled a paid pilot, and the kill criterion was evaluated on its written date rather than quietly postponed.

Those first committed people are more than evidence — they are your first customers, and the conversations that found them are the same motion that finds the next ones. Stage three, get your first ten customers, scales that motion from three to ten.