Why we removed the worker marketplace, and what Koora is becoming

We have removed the part of Koora where a provider could browse care workers, and we are not building the job posting that was meant to follow it. This is a note about why, because the reasoning changed what we think the company is.

What we built, and why it looked right

The original shape was reasonable. Care has a workforce shortage, providers struggle to fill shifts, and Koora was already assembling verified, current credentials for workers. If you hold a set of people whose checks are current, letting a provider search that set looks like the obvious next move. Post a job, find someone already cleared, onboard them in a day instead of three weeks. We built the browse surface and had job posting queued behind it.

The logic was that compliance gets you in the door and hiring is where the value compounds. Compliance is a burden, and burdens are a hard thing to build a company on. Nobody wakes up excited about credential expiry. So we wanted to be more than that, and hiring was the answer we reached for first.

What providers actually said

We launched in July. Since then we have spent most of our time in front of providers, and one thing came up repeatedly. It was not about features, and it came up at the same point every time: when a provider was working out whether to put their workforce in.

Providers were uneasy about putting their workforce into a system where other providers could see it. Not their compliance data, which nobody could see anyway without the worker granting access. The plain fact of the roster. In a sector where everyone is short-staffed, a searchable list of care workers with current credentials, sorted by location, is a recruitment list. We had built one and then asked providers to fill it with their own staff.

Job posting would have made that sharper, not softer. The clearest signal that a service is worth approaching is a service advertising that it is short.

The part we got wrong

The mistake was not the marketplace on its own. It was not noticing what the marketplace did to everything underneath it.

The value of a shared credential record comes from it being complete. The entire proposition is that the workforce is in one place with current checks, so nobody has to assemble it again from scratch. Anything that gives an organisation a reason to hold part of its workforce back does not reduce that value at the margin. It removes the reason the thing exists at all.

So a feature we added to make the product worth more was quietly making the core of it harder to adopt. We were asking for the whole workforce while running the one feature that made handing over the whole workforce feel risky.

What it makes possible

Two things open up once nobody is shopping for staff.

A provider can put their entire workforce in without it being a recruitment risk. That is the difference between a compliance tool used for the awkward cases and a record that actually covers everyone, which is the only version that does the job.

And staffing agencies can use it properly. An agency's book is its business, and a platform where client organisations could browse workers is close to the last place an agency would ever upload it. With browsing gone, an agency can put its whole book in and use Koora for the thing it is genuinely good at, which is keeping credentials current across people who work across many sites for many clients. Agencies do not pay for passport access here. Placements work the way they always have, with an agency sharing a specific worker with a specific client, which is a deliberate act rather than a shop window.

What we think Koora is now

We have started describing Koora as the bureau for Australia's care workforce, and the word needs defining because it can be read two unhelpful ways.

It does not mean a government body. And it does not mean a credit bureau, which is worth being precise about, because a credit bureau holds a file you do not own and cannot see without asking. This is the other way around. The worker holds their Career Passport and decides who sees it. The sources are government registers rather than employers reporting on each other.

What we take from the bureau idea is narrower. One maintained record that many organisations can rely on, instead of the same person being checked from scratch by every service they work for. Checked once, kept current, and readable by whoever the worker allows.

That reframes the burden problem too. Compliance is genuinely a burden, and we were right that a compliance tool alone is thin. We were wrong about where the extra value comes from. It comes from the same work stopping being repeated across the sector: one worker, checked once, current everywhere they work, rather than every service they pick up shifts with running the same checks over again in the same year.

What we are not claiming

Much of what "bureau" implies is not built yet, and the distance is worth naming rather than glossing. We are early and we have no track record. This is a decision and the reasoning behind it, not evidence that the decision was right.

We also gave something up, though it was never a customer request. Nobody asked us for job posting. It was our own bet on where the value would come from, and we had built the browse surface and queued the posting behind it before we tested that bet. We removed the code rather than hiding it behind a setting, because a feature kept alive on the off chance is a feature that keeps shaping the product. If we are wrong about this we will rebuild it deliberately, and we would rather make that decision again from scratch than drift back into it.

The limit on the rest is the same as it has always been. Where an authoritative register exists we verify against it: identity through the Document Verification Service, AHPRA, the public working with children check lookups in Victoria, Queensland, WA, Tasmania and the NT, ban registers, and the police checks we will run ourselves soon. Everything else we review, which is a different and weaker claim, and we will keep saying which is which. The provider's own team reviews in Koora, or Crew does it for them, and the provider keeps the legal responsibility for who they engage.

What this commits us to

• Providers do not browse workers. Not as a setting, not as an enterprise feature, not later.
• The worker holds the Passport and controls who sees it. The bureau holds the verification history.
• Agencies do not pay for passport access.
• We keep "verified by Koora" and "reviewed" distinct, including when the looser word would read better.

None of this makes compliance exciting. It does mean the work only has to happen once, which is the closest thing to a real answer we have found.