How We Use AI — and How We Don't
My Signal Path is an AI-powered interview. If you're a parent or a counselor, you should be asking hard questions about that. Here are straight answers.
The methodology is human. The interviewer is AI.
Every question in the interview — and the framework used to find the pattern underneath the answers — was designed by me, drawn from 30 years of interviewing people for a living. The AI is Claude, built by Anthropic, a company known for its safety research. Claude conducts the interview and writes the Snapshot, but it follows my methodology. It's not making it up as it goes; it's running the interview I would run.
What the AI never sees
The interview doesn't ask about race, ethnicity, family income, address, or anything demographic. A student provides a first name and a grade level, and then the conversation is about how they think, what gives them energy, and how they naturally operate. The Signal Snapshot is built from their words — nothing else.
Advisory, never gatekeeping
My Signal Path makes no decisions about students. There are no scores, no rankings, and no comparisons between students — every Snapshot stands alone. No school uses it to admit, reject, track, or place anyone, and it isn't designed to be used that way. It's a mirror, not a gate. Results are one input among many for a student and family thinking about what's next.
Guardrails built in
The interview only runs for students in grade 11 and above — the product enforces this at the start. If a student discloses something during the interview suggesting they may be at risk, we have a defined process for alerting the school or program that issued their access code — by code number only, so the adults who actually know the student can follow up. And the interview is built to stay on purpose: it's a career self-discovery conversation, not a therapy session, not a crisis resource, and it says so.
We check our own work
I personally review Snapshots for quality and safety on an ongoing basis. And because the Snapshot is built entirely from a student's own words — no demographic inputs, no scoring, no comparisons — there's a simple standard it has to meet: the same answers should produce the same read, no matter whose name is on them. That's not a promise that AI is perfect; it's the standard we hold ours to.
Look for yourself
Schools and programs can inspect the complete interview question set before any student takes it — just ask. Our full data practices are in our Privacy Policy. Questions about any of this: brian@mysignalpath.com.