There is a business model in recruiting technology that we refuse to run: charging candidates for access to opportunity. It appears in soft forms — priority placement for a fee, paid visibility to employers, premium tiers that decide who gets seen — and every one of those forms turns a labour market into an auction that the people with the least money lose by default.
The rule we operate by
Candidates may pay for tools. They never pay for chances. A candidate can choose to use our workspace to prepare a profile, rehearse for a conversation or understand a market, and that is a service with a price like any other. But whether a candidate is surfaced to an employer, ranked, shortlisted or introduced is decided on evidence and fit alone. Money has no seat at that table, and the architecture makes sure it cannot buy one.
The same principle applies in the other direction. Employers pay for the work our digital workers do — the observation, the preparation, the evidence — not for the right to be shown candidates the candidates did not agree to be shown to. Consent is recorded before any profile moves, and it can be withdrawn.
Why fairness has to be built, not promised
Any system that ranks people can drift toward the biases of the data it learns from. We treat that as an engineering fact rather than a public-relations problem. Rankings are explainable: every shortlist arrives with the evidence that produced it, in language a hiring manager and a candidate could both read. Protected characteristics are not features. And the final call — the introduction, the interview, the offer — is always made by a person who is accountable for it.
That last point is the one we are asked about most. If the machine is good, why keep a human in the loop? Because hiring is a decision with consequences for a life, and accountability cannot be delegated to a model. The digital worker prepares the decision so that it is faster, better-evidenced and less influenced by whoever happened to shout loudest that week. It does not make the decision, and it is not allowed to.
What candidates can expect from us
Clear labelling of anything sponsored, so a paid placement is never mistaken for an editorial recommendation. No hidden ranking fees. A record of what was shared with whom, on request. And a route to a human when something looks wrong, staffed by people who can actually change the outcome.
Fair AI in hiring is not a feature you add later. It is the set of things you decide never to sell.
We publish this so that we can be held to it. If you ever see Meridian charging a candidate for a chance rather than a tool, that is a bug in the company, and we want to hear about it.