# Disqualification pass

From **Discovery that proves something: a field guide for founders** (Atlas, field guide no. 01).

The first pass over a scraped lead list. It screens out competitors and companies on a
collision course with what you are building, before you spend an enrichment credit on them.

Fill in every `[BRACKET]` before you run it, paste your extraction output at the end, and keep
the rows marked `KEEP`.

Two things about how it is written, because they are the parts that matter:

- **It defaults to KEEP.** A competitor who slips through costs you one awkward call, and you
  find out inside two minutes. A prospect wrongly dropped never shows up again and you never
  learn you lost them.
- **It returns a confidence rating**, so you can review the low-confidence rows yourself
  instead of trusting the whole pass.

---

## The prompt

```text
You are screening a list of companies for a customer discovery outreach
campaign. My company is building [WHAT YOU ARE BUILDING, one sentence].

For each company below, decide whether to KEEP or DROP it, using these rules.

DROP if any of the following are true:
- The company builds or sells something that directly competes with what
  we are building.
- The company builds in an adjacent space and their public roadmap,
  recent hiring, or recent funding suggests they are moving toward our
  space. Include this even if the overlap is a year or more away.
- The company's core product would be materially cannibalized if our
  product worked as intended.
- [ANY OTHER DISQUALIFIER SPECIFIC TO YOU: e.g. agencies, consultancies,
  resellers, companies below or above a size threshold]

Signals to look for when applying the rules above:
- Product descriptions containing [YOUR CATEGORY KEYWORDS]
- Job postings for [ROLES THAT WOULD BUILD A COMPETING PRODUCT]
- Funding announcements framed around [YOUR PROBLEM SPACE]
- Positioning language that overlaps with [HOW YOU DESCRIBE YOURSELF]

KEEP everything else, including companies you are unsure about. When the
evidence is thin, KEEP and flag it rather than dropping it.

For each company return exactly this, one row per company:
COMPANY | KEEP or DROP | ONE-LINE REASON | CONFIDENCE (high/medium/low)

Do not infer anything you cannot support from the information provided.
If a company's description is too sparse to judge, return KEEP with
CONFIDENCE: low.

Companies:
[PASTE LIST]
```