NewMulti-agent job simulations for enterprise hiring

Your best candidateis still unfound

Every talent decision runs on skills data nobody has verified. Unfound is the source of truth: multi-agent simulations of real work that prove what people can actually do.

Plugs into your ATS AI screening built in Every score tied to evidence
11,000 LinkedIn applications a minute 35% of managers saw a stand-in at a virtual interview 67% of HR leaders slowed by AI applications 1 in 4 candidate profiles fake by 2028 65% say AI résumés make skills harder to verify
The shift

Everything an interview measures is now free to fake.

Applications, résumés, answers, even the candidate. Three numbers describing one event.

{{stv.a}}applications hit LinkedIn every minute, up 45% in a year.LinkedIn via The New York Times
{{stv.b}}%of HR leaders say AI-generated applications have slowed their hiring.Robert Half, 2,000+ hiring managers
{{stv.c}}%of hiring managers say someone other than the applicant took part in a virtual interview.Checkr survey of 3,000 managers, 2025
The open position

The system of record for skills.

Workday is the system of record for people. Nobody is the system of record for what people can actually do.

For five years the industry has preached the skills-based organisation: hire on skills, promote on skills, plan the workforce on skills. HR systems and talent platforms all sell skills graphs, and every one is built on unverified inputs: self-ratings, parsed résumés, course completions, endorsements. The decisions are skills-based; the data underneath them isn’t.

That record was never claimable, because nothing could measure demonstrated capability at scale and cost. Multi-agent simulation is that instrument, the first one. A new measurement technology is what creates the right to a new system of record: thermometers before medicine, telescopes before astronomy, simulation before verified skills data.

Unfound is the source of truth for human capability: the verified skills record the skills-based organisation has been missing.

Inside the simulation

Don’t ask how they’d handle it. Watch them handle it.

A real run forks at every decision the candidate makes, so there’s nothing to leak and nothing to rehearse.

Decision 1 of 5 · Product lead scenarioLive
AN
Analyst · agent

The Q3 data contradicts the launch plan. Launch is Thursday.

APause the plan and re-pull the dataVerifies before committing
BKeep the date and flag it after launchTrades certainty for speed
32 possible worlds from here. The first decision closes half of them.
The report

What the interview leaves unfound.

Same candidate, same week. One record you have to take on trust, and one you can replay.

The interview hears
Unfound sees
The interview hears“I’m very data-driven.”
UnverifiedAI-writable in 2 sec
Unfound sees
07:15Replayable

Hit contradictory data and paused the launch to re-pull it before committing a date.

Verification 91

Illustrative sample report

Use cases

One record of skills, every talent decision.

Hiring is where it starts. The same simulations verify skills wherever your team makes a call about people.

Hiring

Screen candidates at scale on today’s skills and future potential, not résumés and past accolades.

Hiring for sales, customer success and engineering
Onboarding

New hires run the same simulations as practice before they touch real customers. Their gaps are known from the hiring run, so day-one development plans write themselves.

L&D and skill certification

Teams pay for practice, not content: reps rehearse a new pitch against AI buyers, managers practise difficult conversations, support drills a launch. A flight simulator, not videos and quizzes.

Internal mobility and workforce planning

Match people to internal roles on demonstrated skills, not skills inferred from résumés and HR data. As roles are redesigned around AI, re-verify what your people can do, not only new hires.

Performance and promotion calibration

Simulation-based promotion assessments: what enterprises already pay assessment centres for, at a scale they can run for everyone.

Plugs into the ATS you already run
GreenhouseLeverAshbyWorkday
Unfound recommends. Your team decides. Every score links to the decision behind it, so no hire rests on a number nobody can check.
Early access partners
Questions

Hiring on evidence, explained.

What is an AI job simulation?

A live, AI-run version of the actual job. Unfound builds a scenario from the role, with its stakeholders, data and pressure points, and autonomous AI agents play the client, the engineers and the VP. The candidate makes real calls, the agents react, and every decision is recorded. You see how someone does the work, not how well they describe it. How the simulation works

How is it different from a skills test or a work-sample test?

Skills tests check knowledge in isolation, and a static work sample has one right answer that can be rehearsed or leaked. An Unfound simulation forks at every decision, so each candidate gets a new scenario that reacts to what they just did. It measures judgment under pressure: what they verify, what they trade, and how they handle pushback. See two runs fork

Does Unfound replace our ATS or our interviewers?

No. Unfound plugs into your existing ATS and keeps your sourcing as it is. Your interviewers stay too: they get an AI copilot that replays the simulation, so they walk into the interview knowing what to probe. How Unfound works

Your best candidate is still

found

Unfound finds them. Join the waitlist for first access to multi-agent job simulations.

You’re on the list.We’ll email you when your invite is ready.

Work email only. No spam, just your invite.