One of the bets at the center of A Confident Company is that the right atomic unit for this class of product is not "an app" — it's a session engine. Once you have one, everything else is a matter of rubrics, personas, and brand.
The session is the product
When a user opens JobConfident and runs a practice interview, what's actually happening underneath is a highly structured sequence:
- A persona is selected based on role, level, and target employer.
- A rubric is loaded for the competencies being assessed.
- The model plays the evaluator role, generating questions that exercise the rubric axes.
- Each response is scored in real time against the rubric, with voice or video telemetry optional.
- A post-session summary stitches competency scores, gap analyses, and improved-answer synthesis into a durable artifact.
That sequence is not JobConfident's — it's Confident OS. JobConfident configures it for a specific job-interview context. DeposeConfident, in private beta, configures the same engine for cross-examination. SalesConfident (H2 2026) configures it for discovery calls and objection handling.
What reusing the platform actually means
"Shared platform" is one of those phrases that means everything and nothing. Concretely, for us, it means that when we stand up a new vertical:
- Persona engine, evaluation core, scoring, longitudinal tracking, billing, identity — all Confident OS.
- Rubric library, persona specs, brand theming, pricing, copy — per-brand.
We've been holding ourselves to a rough 70/30 split. If a candidate vertical needs more than 30% net-new infrastructure to ship, it's not a Confident Company product yet — either the infrastructure needs to graduate into OS, or the vertical needs to wait.
Why we don't put this on the product sites
Nothing on JobConfident.com mentions Confident OS. That's deliberate. A candidate preparing for a final round doesn't care that our session engine also powers a deposition tool. They care that we're focused on their room.
The platform is legible on the parent site because the audience here is different: partners, licensees, acquirers, senior hires. They care about leverage and repeatability. They should see the shape of the engine.
What this unlocks
The practical effect is that we can evaluate candidate verticals with something closer to a portfolio-manager lens than a product-manager lens. "Does this clear the four filters? Does Confident OS carry at least 70% of it?" If both are yes, we can afford to ship, price for a narrow audience, and let the brand quietly find its people.
That's how you end up with six brands on one roadmap.