BaseData BI is a micro business that makes sense of social housing data. This month we did something that, for a company our size, counts as a big move: we took on our first apprentice, on the brand new Level 4 AI and Automation Practitioner standard. This is me explaining the decision, and why I think it should matter to the councils and housing providers we work with as much as it does to us.
A fresh pair of eyes
Our apprentice is fresh out of college, and I’ll be honest about why that appeals to me. Social housing is, on the whole, a world built and run by an older generation. The systems are old, the people who really understand them have been at it for decades, and a lot of hard-won knowledge sits in the heads of people closer to retiring than to starting out. Bringing in someone at the very beginning of their working life, with none of that baggage and a completely different relationship with technology, feels like the right kind of shake-up.
What drew them to us wasn’t the AI buzzwords. It was the idea that this stuff could take the boring, repetitive grind off people rather than replace them. They’d already got fed up doing things the slow way and gone looking for a faster one, which is more or less the whole job. You can’t teach that. You can teach the tools.
I wanted someone from the next generation looking at a sector the last one built.
Why a micro business does this
There’s a myth that apprentices are something only big employers can afford, with their training teams and their slack in the system. For a micro business it feels riskier, and honestly it is. A large company can take someone on, spend a year or two developing them, and barely notice if it doesn’t pan out. We don’t have that cushion. When you’re small, every hire is visible and every mistake is expensive.
We decided it was worth it anyway. If you only ever buy in ready-made skills, you’re permanently at the mercy of who’s available and what they cost. Growing someone yourself is slower and less certain, but you end up with capability shaped to fit your business, and a business that isn’t one resignation away from losing something critical.
A few things beyond that made the maths work:
- The obvious one is time. Routine work can shift to someone else, which frees up the people whose time is better spent on the harder problems.
- A newcomer sees the friction you’ve gone blind to. Run the same process for years and you stop noticing how clunky it is. Someone learning it from scratch asks why on earth you still export that by hand every Monday.
- You build the specialist you’d struggle to hire. There is no ready supply of people who understand both social housing systems and modern automation. It’s easier to grow one than to find one.
What it costs
The money is less frightening than people expect. The AI and Automation Practitioner standard sits in an £18,000 funding band. For a small employer that doesn’t pay the apprenticeship levy, the government covers the large majority of that, with the employer putting in a small share, historically five per cent. The smallest employers taking on younger apprentices can end up paying nothing towards the training at all, and from August 2026 the government has said it will fund the full training and assessment cost for every apprentice aged 16 to 24. There are incentive payments on top for taking on eligible young people.
So the cash outlay is modest. What it actually costs you is attention: the time to sit with someone, review their work, and let them learn on real jobs rather than made-up exercises. That’s the part no funding claims back, and it’s the part that matters. The exact figures depend on the apprentice’s age and situation and on the funding rules for the year, so check the detail with a training provider before you bank on any of it.
Why this apprenticeship
The standard is brand new. The Level 4 AI and Automation Practitioner (reference ST1512) was only approved at the end of 2025, so our apprentice is on one of the first cohorts in the country. It isn’t a computer science degree wearing a different hat, and it doesn’t ask for one. It starts with visual, no-code tools and builds up from there.
Strip out the jargon and the job is this. You learn to look at how an organisation really works and find the manual, repetitive effort that’s quietly eating hours. You learn to build something that takes that work over, using low-code and no-code tools and, increasingly, AI that can be trusted to collect data, process it, and act on it without someone hovering over a spreadsheet. Then you learn the harder parts: testing it on real cases so it doesn’t just work in a demo, keeping it the right side of the law on data protection and bias, documenting it so the next person can follow it, and bringing colleagues along rather than springing it on them.
The assessment reflects that. It runs to around eighteen months, and it ends not with an exam but with a real project that solved a real problem in the business, plus a discussion with an independent assessor, graded pass or distinction. You earn it by making an actual organisation work better.
That last bit, the governance and data protection, isn’t filler, and it’s a big reason this standard suits housing. When the data is about people’s homes and tenancies, “move fast and break things” is exactly the wrong attitude. Building things carefully, and being able to show your working, is the job.
Where they’ll start
Two areas, to begin with. The first is the housing data itself. Social housing runs on big, complicated systems, and an enormous amount of the reporting that sits on top of them is still built by hand: repairs and works orders, voids and re-lets, tenancies, arrears, all pulled together every week or month so someone can see the state of play. It’s slow, it breaks easily, and it eats the time of people who should be doing something better. It’s also the perfect thing to learn on. Take a report someone rebuilds from scratch every Monday, work out what it actually needs to say, and turn it into something that produces itself, on time, every time.
The second is our own back office. Small businesses seize up when too much of the day is manual routine, and we’re no exception. Automating the repetitive parts of how we take on, deliver for, and support clients is how we free up capacity without either dropping the ball or just working longer hours. Every hour we save ourselves is an hour we can give a client.
None of this is fetching coffee and watching over someone’s shoulder. It’s real work with something at stake, from the first month, because that’s the only way anyone learns this properly.
Why it’s good for housing too
That’s the case for us. What about the people we work for?
Housing teams are stretched, and the amount they’re asked to report keeps climbing: Tenant Satisfaction Measures, the consumer standards, Awaab’s Law and the clock it puts on dealing with hazards. All of it needs data that’s accurate and that you can stand behind. And a lot of the work behind that data is still depressingly manual, exports and copy-paste and spreadsheets stitched together late on a Friday. Getting rid of that grind is what this apprenticeship trains someone to do, which means housing staff can spend more of the week on tenants and homes and less of it fighting a spreadsheet.
There’s a bigger point about the supply chain. Councils and housing associations lean on a whole network of specialist suppliers, and that network is only as capable as the people in it. When a small supplier puts real effort into building new skills, automation included, the clients feel it directly. And there’s the social value angle, which councils are obliged to weigh when they buy. A genuine apprenticeship, a real job with real training in the local economy, is about as concrete as social value gets. It’s one thing to write about it in a bid. It’s another to actually be doing it.
The bigger picture
I’m not going to pretend this is altruism. We did it because it’s good for us: more capacity, less fragility, and skills grown around the exact problems we spend our days on. What’s striking is that the selfish reasons and the good ones point the same way. The decision that strengthens a small supplier also helps the councils it serves, and hands someone at the start of their career a genuinely interesting place to build one.
For a business our size it’s a bigger gamble than it would be for a company ten times over. I’d argue that’s the point. If skills only ever get developed inside large organisations, they stay there. It’s smaller firms taking the risk that pushes capability out into the wider economy, and that’s the economy local authorities actually rely on.
If you run a small business and have quietly written off the idea of an apprentice, too small, too busy, too risky, it’s worth another look, especially at some of the newer standards like this one. The help is real, and the cost is lower than you’d guess. And if you’re on the other side of the fence, in a council or a housing association, wondering whether your data really has to stay this manual: it doesn’t. That’s the future we’re building towards, and we’ve just hired the first person to help us build it.
We’ll let you know how it goes. If nothing else, there’s a long queue of manual reports around here waiting to be put out of their misery.
