We're currently hiring at Taught by Humans - two open roles (Customer Success and Generalist, if anyone's interested) with a mild time crunch due to live grant projects. Grant projects which are ironically about hiring and getting people into work. But more on those later.

Growing the team has been a process - from only interns on uni schemes, to letting go of my first permanent hire three months in, to what I would honestly call the current dream team. This week has been a tough one - I had offered two people roles who would have been fab additions. Unfortunately both fell through - one due to a much better counter offer, and the other to bad timing. It's tough, but we move on.

Always Learning

Living and breathing one of our values, Always Learning, I'm always on a mission to improve our hiring processes. From many angles:

Speeding it up as hiring is quite team consuming
Making it easy for candidates, and avoiding any biases in hiring
Avoiding getting lots of spam or unsuitable job applications

Something I Hate About AI

CV and cover letter used to work really well, especially for interns. Their CVs didn't really matter - although I did learn that having had random part time work in a bar or a shop was weirdly correlated with doing well in a small team. Their cover letters were written by them and showcased what they knew about Taught by Humans and why they were a good fit. I genuinely used to enjoy reading them.

Then GenAI came along. Now for every role I get 40+ nearly identical cover letters.

I am thrilled to apply for...

Sigh.

CVs are awful anyway. At this point a list of roles, companies, dates and responsibilities would be much better.

Avoiding Spam

To try and stop my email address getting spammed if we put job specs online, my wonderful Ops Manager suggested a Google Form. Simple but seemingly effective.

We added two blocking questions:

Right to work
One day a week in Bristol

Which seems to have prevented unsuitable applicants.

Then added three relevant questions per role. I have also added the ability to add links to LinkedIn, websites, anything people want to show off. No CVs in sight.

Analysing Answers

One issue with being a solo founder is having to make big decisions on your own, and sometimes questioning yourself and your own biases. Luckily the team, and my partner, are very good to check in with opinions. But sometimes I do like to rely on Claude.

When we had ten applications for the Generalist role, I saved them to a spreadsheet, anonymised them, wrote my thoughts and outcome, then shared only the responses with Claude. I wanted to confirm I wasn't overlooking something.

Claude initially identified four people to interview. Two I agreed with, although not as enthusiastically for one. Two I disagreed with - interestingly what I had picked up on that Claude didn't was their answers didn't demonstrate any collaboration or teamwork. Something super important to our team culture. I did interview the latter.

There were two maybes - one where two different Claude chats disagreed with basically the same prompt. I'd love to spend time diving into the reasoning behind this, as it's a real flaw of language models. I decided not to interview in this case. The other was referred by one of the team - so a maybe is always an interview.

The rest we both agreed were not right for the team or the role.

Interview Stage One

Now we haven't hired yet, so this is an incomplete story. But bear with, it does have a point. I have interviewed three people out of this list - the one me and Claude disagreed on (a very accomplished individual, but not right for this role or the team), Claude's top pick who interviewed well and is through to the next stage, and the maybe who was a referral - my current front runner for the role.

Rather than go on about how AI misses something important in anything human related, I want to look at something else. How the hell do we actually fairly assess people pre-interviewing?

I keep wondering - would we have interviewed the referral if they hadn't been referred? I'm actually not sure, probably but not definitely. And is this because while CVs and cover letters aren't right, is asking people to write responses about themselves the solution? It penalises people who don't shout about themselves, or who have interesting careers and too much to talk about. Videos and weird AI interview screening are also not the solution. Myself and my accent would never get a job that way.

We Might Have a Solution

As part of a TechFirst grant - which is, yes, a grant about hiring and getting people into work - we are going to use Dotly (our AI learning companion) to try to get to know people a bit better before matching them with jobs.

Our logic being:

People seem to enjoy chatting to Dotly (and other AIs)
Women often undersell themselves - and a conversation might help with that
Young people, people who have never worked, and people returning to work after breaks often don't make the link between their natural day to day skills and hireable skills

I see this as a solution, and 100% not a replacement for anything human - a replacement for CVs and little boxes.

But I am unsure if we're missing something - because our biggest signal for hiring is gut feeling. Mine, the team's, and people who know Taught by Humans and what it takes to work here (the two people offered roles this week were both referrals from customers).

In previous blogs about hiring, I've mentioned a phrase - earned nepotism. It's when someone vouches for you, ideally someone who has worked with you. Or in the gold standard case - we have worked with you as a partner or customer. It's not exactly unfair, well not in the way of the boss' nephew getting a management role with no experience is. And it's a big derisk from my side as the founder of a tiny company. But there is definitely something about doors and ladders not being equally accessible to everyone.

So rather than claiming we're revolutionising hiring, or replacing a certain stage in the process, perhaps this pilot is to see if Dotly can open doors for people who don't have that earned nepotism yet.

We're Hiring. And It's Made Me Question Everything About How Hiring Works

Want to put this into practice?