Your cohort, rehearsed
Every batch you run ends the same way: a room full of people giving the talk that decides something. AtsIg gives each of them a private rehearsal loop before that day: record a take, get every note pinned to the exact second it describes, fix one thing, take two.
What a cohort gets
- Rehearsal on the real talk. Founders rehearse the demo-day pitch; doctoral candidates rehearse the defense. Not generic exercises: their own material, timed against their own format.
- Evidence, not vibes. Every note carries a timestamp and survives a validation gate before anyone sees it; notes that fail the check are dropped, visibly.
- Their style survives. Reports open with what to keep, and rewrites stay in the speaker’s own voice. The goal is a sharper version of each person, not thirty copies of one ideal presenter.
How a batch works
You tell us the cohort and its date (demo day, defense window, competition). We invite the group as a block; each person signs in, consents individually, and rehearses on their own schedule. Speakers own their recordings and reports.
What your program does not get, on purpose
- No ranking. The product has no way to compare one person to another; that surface does not exist by design, and we refuse to build it. Progress is each speaker against their own last take.
- No dashboards over individuals. Program staff never see anyone’s recording, transcript, or report. What you get is happier, better-rehearsed speakers.
- No surveillance posture. Only the audio of a take is analyzed, consent is per session, and we name processors instead of promising blanket EU residency: our infrastructure runs in the EU, and the coaching analysis itself is an API call to Anthropic. The full picture: trust, plainly and the privacy policy.
For an institution, this is the point: you can offer real coaching without creating an assessment instrument someone could misuse later.
What it costs during the pilot
Nothing. Pilot batches are free; the price is feedback: fifteen minutes with you after the cohort finishes, on what worked and what did not. The pilot is deliberately small, so batches are scheduled, not unlimited.
Who builds this
Built by an engineer trained at EPFL and École Polytechnique, now at MIT. Those are the schools the founder studied at, not partners or sponsors of this product.
Bring a cohort
One email starts it: hello@atsig.ai. Tell us the program, the cohort size, and the date that matters. You’ll get an answer from the person who wrote the code.