None of us set out to be recruiters.

I wanted to be a stockbroker. I did the qualifications. If I had been any good at it I would be driving a G-Wagon and fiddling with a portfolio, not writing this.

Nobody grows up wanting to do this job, and a lot of it is dull. Really dull. In-house at a big energy company it was form after form after form, just to book one interview. And before anyone tells me those days are over — they bloody well are not. I sat an interview process at a very large bank this year. Two legacy systems, and Excel for most of it.

But the work underneath the admin matters. The people are the company. Finding the right ones is the whole job, and agency-side you need to find them fast, because it is dog eat dog out there.

And it is hard, in a way people outside it never see.

Put someone in, they do not work out, and you get it in the neck. Not the hiring manager. Not the six people who interviewed them and loved them. You.

We get put in the same bracket as snake oil salesmen and estate agents. Meanwhile the job is thankless, genuinely difficult, and the good part — ringing someone to tell them they have got it, and hearing what that does to them — is one of the best feelings there is in work.

So, why this

I did not build Hikari to revolutionise recruitment. I built it to get the boring half of my own day back.

I used to tell my boss we needed fewer recruiters and more coordinators, so the recruiters could do the part that actually needs a person — sourcing, closing, negotiating the offer — instead of chasing their own tail. Nobody did anything about it. So eventually I built the thing myself.

Now the honest part. Every other post on LinkedIn is another AI product, or another warning that AI is coming for your job. It is saturated, and it is dull. I am almost apologetic that we have built one. I would rather not say the letters at all and just call it software, because the moment anyone says AI, eyes roll. Ours included.

What I would say in its defence is this: it was not built by engineers who have read about recruitment. It was built by people who did the job. That counts for something.

What you can hold us to

We publish what our scoring does, including when it is wrong. This year we tested our own system for name bias. We found some. We took the names out, tested again, and the effect disappeared — and we will publish both sets of numbers in full, not the flattering one. People deserve to know what their AI is doing to their candidates.

No buzzwords. No song and dance about how wonderful we are. We are not going to charm you into buying anything.

Do I think it is good? Yes. I also think it will be better in six months, and better again after that.

Patagonia once ran an advert telling people not to buy their jacket. I have always liked that.

And if it is not for you, do not buy it. Honestly. We would rather you did not.

If it is — we are a small team, so you get us. We will build the parts your desk actually needs, because we are the same as you and we know what a good one looks like.

Launching a business is a bit like the first day at school. You want everyone to like you, so you become a social chameleon — agreeable, polished, whatever the room seems to want. We are not doing that.

We are what we are. We were just as knackered as you are. That is the whole of it.