Calibration labs call script (Track 2)

This is a test of whether the segment is real, not a campaign. There are about nine named calibration & testing labs (plus accreditation directories). Nine is too few for an email test, so you phone, cheapest-to-reach first, and you listen for one specific sentence.

The success test, decided up front

• Segment is real if 2 or more of the first 3 labs say, in their own words, "our customers ask us for measurements we can't do today."
• Segment is not real if 0 of 3 say it. Stop; report it; move the effort back to ICP1.

Decide this on the first three calls. Do not keep dialing hoping for a yes.

Before you dial

• Claim the lab's row (one owner — page 12) and read the claims boundary. The four conversation-enders apply on the phone exactly as in writing.
• You are doing discovery, not a demo. You want to learn what their customers ask for that they cannot deliver.

The script

"Hi, I'm ⟨name⟩ with ApertureAPI. We host a reconstruction service that turns holographic microscope images into height maps with a re-computable provenance record. I'm not selling you anything today — I'm trying to learn whether this is useful to labs like yours. Do you have two minutes?"

Then, in order:

1. "What kinds of measurements do your customers send you work for most often?"
2. "Is there anything they ask for that you can't do today, or have to turn away?" ← this is the sentence you are listening for.
3. "When a customer needs to know exactly how a result was produced — software, settings, version — how do you give them that today?"
4. "If a customer asked you for a holographic reconstruction with a re-computable receipt, is that something you'd want to offer?"

What you may and may not say

• You may say: determinism (same input and config give the same result ID), the content-addressed provenance record, SHA-256 commitment to input and config, that software-defined calibration recovers pixel pitch very tightly on two committed closed-loop targets.
• You may not say: any accuracy or agreement number, any speed, "SI-traceable" or "certified," or that we replace an instrument. If asked "can this replace the measurement my customer depends on?" the answer is no, and our own report says so in writing — then pivot to the free validation and the Software Design Partnership (page 22).

What to log

Per lab: reached (y/n), the answer to question 2 verbatim, and your yes/no on the success sentence. After three calls, write the verdict (real / not real) in the weekly report.

Artifact when done: three logged calls with the success-sentence tally and a written real/not-real verdict.

Buyer pages to pair with this page: industrial QA · validation docket · workflow fit. Pipeline stages and KPIs: page 00.