Ten years ago, being good at spreadsheets was a hirable skill. Being a VLOOKUP pro got me a part-time job in an accountancy firm at uni. My highest ever paid job was basically to manage the messiest Google Sheet (which I did not make) for an edtech startup during COVID.
For a while, Taught by Humans' bread and butter was running spreadsheet courses for women - because spreadsheet skills unlocked better-paying jobs and more responsibility. That was real, and it worked.
Now, being decent at spreadsheets is expected for pretty much any office job. People do not even put Excel on their CVs anymore. It is still important though - Nicole, our Founders Associate, is currently learning data analysis in Google Sheets rather than jumping straight to Python, because I think it helps you feel the data before you automate it. (Yes, I am a data geek. Luckily, so is she)
So Where Does AI Fit Into All Of This?
I deeply believe AI tools (whether that is ChatGPT, Claude, Copilot, or automation tools like Zapier, Make or n8n) should also become required workplace skills. At varying levels, suited to roles, company needs and individuals.
Which is where I am going to have a little whinge. I fail to see how apprenticeships in AI are solving this problem. We have been through this before - web development apprenticeships for everyone, data apprenticeships for everyone, regardless of their actual job. Apprenticeships are, in my opinion, being oddly misused here, and the nice government funding that comes with them is probably why. They should be for opening up new roles to existing staff, or new jobs to new people.
I do think the roles listed in the AI and automation practitioner standard are really interesting, and how the apprenticeship is structured is amazing for them. But what I am actually seeing is managers, product people, and ops leads being funnelled onto 18-month programmes - people for whom AI is part of their role and needs to be learnt, but who do not need an 18-month funded programme to do it.
Multiverse are doing genuinely interesting work, and I respect what they are trying to build. But their own data makes my point for me. With a 52.6% completion rate - against a sector average of 65.4% - and Ofsted finding apprentices enrolled on courses that did not match their actual jobs (funeral planners and security guards on data and AI programmes), the wrong people are ending up on these programmes. Not because Multiverse is bad at what they do. Because the incentive structure of apprenticeship funding pushes volume over fit. Not because Multiverse is bad at what they do. Because the incentive structure of apprenticeship funding pushes volume over fit.
AI Skills For AI Roles
Advanced spreadsheet and data skills are still jobs in their own right - but we have landed on expecting everyone to have some ability relative to their role. As a former data person, it was always better to work with data-literate teams. They came to you with more interesting problems.
AI should land in the same place - everyone should be able to prompt and use it safely, relative to what their job actually requires. Not everyone needs to be building automations and agents, and that actually feels like a fantastic new entry-level job that companies can hire for, train into, and grow. A perfect use of apprenticeships.