Our AI Policy Buddy: Transforming Children's Services in North Yorkshire

Engine’s AI Policy Buddy: Transforming Children's Services in North Yorkshire

In our last blog, we reminded ourselves that AI adoption won’t just happen. It will require us to demonstrate how AI can make lives better, inspiring everyone’s involvement, and integrating new tools seamlessly into daily routines. We also allowed ourselves to raise a cynical eyebrow, looking askance at how well/badly Team AI are doing when it comes to getting people to join them on their journey. To cut a long story short: shock and awe aren’t the way to get everyone onboard so we need to do better.

In the spirit of inspiration and usefulness, this week we’re thrilled to share a very real example of successful AI adoption: North Yorkshire Council’s implementation of Engine's RAG AI engine, Policy Buddy, for their children and families social care teams.

I like it. What is it?

We built North Yorkshire Council a RAG AI engine: a private GPT trained on a repository of all the local and national legislation, procedures and guidance relating to children’s services. It has a simple, accessible front end, enabling staff to find information, check the sources and generate summaries of new material. We talked about it over a couple of months, but we built it in less than two weeks. It’s multi-lingual, secure, easy to use, and best of all it only costs about the same to run for one year as buying six and a half new iPhones. So far, we’ve enrolled nearly 1000 council staff onto the platform and they’re finding it genuinely transformational. But don’t just take our word for it...

A vision for change

From the outset, the North Yorkshire team had a clear vision for change. They set specific goals: reducing workload, empowering staff, and enabling a greater focus on direct work with children and families. Jonny Hoyle, who has led the change along with Cath Ritchie, has been instrumental in driving this vision forward. They communicated their vision with passion, but also remained practical and focused on how the new tool would help staff. Jonny shared a memorable moment on LinkedIn:

"Pinch me moment yesterday!! Following our small trial of Engine's Policy Buddy, designed specifically for Children's Services, we've decided to roll it out to the whole workforce. I spent 5 minutes showing a Team Manager what it is and what it could do... she then stood up off her chair, jumped up and down with excitement and called for people to come and look!"

Assessing impact and achieving quick wins

The team assessed the impact of Policy Buddy with surveys before and after implementation. They carefully examined the benefits and value, but avoided getting bogged down in process. Instead, they focused on achieving quick wins and maintaining momentum. Jonny highlighted more positive feedback from a Social Work manager at their midpoint review:

"I feel like the Policy Buddy has changed my life already."

A diverse and engaged team

A diverse team has been crucial to the success of this transformation. They have kept the momentum going, engaging staff at every level. Jonny's social media posts reflect the excitement and engagement of the team:

"💥 Blown away yesterday! 💥I attended our Disabled Children's Services team away day to talk to them about technology and innovation and to show them our Policy Buddy with Engine. The result was outstanding!!"

Building trust in tech

The Policy Buddy has not only impressed the team but also built trust in technology. It provides answers specific to North Yorkshire Council, helping social workers trust the tool because they can verify every source it uses. Jonny also tested the mobile version:

"The layout is easy to use, the response is quick...the answer was comprehensive, reassuring, and contained references so I could check the AI."

Looking ahead

North Yorkshire's successful implementation of Policy Buddy is a testament to the power of clear vision, passionate communication, and practical focus. The team's dedication to assessing impact, achieving quick wins as well as long-term objectives, and maintaining momentum has made this transformation possible. To quote Jonny one last time:

“Great work Julia Yong, Kieron White, Anthony Sherry, and to everyone here in North Yorkshire Council who have made access to innovation as seamless as possible. Especially the AI Ethics Guru Catherine Ritchie!!”

Thanks, Jonny. We’re chuffed to bits.

Why are we in love with North Yorkshire? (Aside from the stunning scenery)

Team Engine all enjoy reading rave reviews from our customers, and North Yorkshire is no exception, but this one is special for a few reasons:

Our partners in the council really get innovation: Yes, we’ve had to lean in to some checks and balances on how public money is used – as citizens you’d hope that’s how things work – but there’s also been ahigh tolerance for adapting where there is no precedent, finding a way, and having to adjust as we uncover issues. Jonny and Cath know that innovation isn’t ‘business as usual’ – but it can still be well-managed. We love that attitude, and it means dedicating resource to innovation, as North Yorkshire have done, if you want to realise the benefits down the line. Have a transformation lead where that is their main job, not their involuntary corporate objective.

It’s a tough crowd: Actually, it’s a lovely crowd –have you ever been in a room full of social workers? They’re amazing: kind and inclusive. But in some ways it’s not an obvious target for early tech adoption. Care teams are thoughtful, outcome-focussed and expert but also time-poor, highly alert to risks, and diverse. They aren’t hanging out at tech conferences waiting for someone to pitch them a new app; they’re busy. And more than almost any other profession they witness first hand, on a day-to-day basis, the harmful effects of things that happen online to the children they work with. But once you provide them with an AI solution that’s both relevant to their vocation AND makes their day easier, they will embrace it because…

It matters. I hope our other customers won’t mind when I say that while it’s great when our AI helps grow a business – because that’s about peoples’ lives and livelihoods too –nothing comes close to the feeling you get when you know you’ve done something to make things easier for asocial worker, who can then more easily help a vulnerable child. And that’s what’s happening. Social work experts – rightly – talk about reflective practice and evidence-led decisions, but reflection and information gathering take time. And our AI tools can help find evidence and best practice more quickly, giving you back even more of that precious thinking and practice time.

It proves you can make good tech accessible. Our tools are simple, they work on phones, you can use speech to text – and they comply with the highest web accessibility standards. They work in different languages and adapt to the style you want – and they are right-sized for the job, making them a little kinder to the planet and putting them within reach of non-profit organisations and SMEs. And they’re secure, so you don’t have to risk a free online platform gobbling up your data. Cutting edge AI doesn’t have to be the preserve of big business – nor does it have to have so many bells and whistles that you start to picture the data centres guzzling the last drop of energy and fresh water. Democratised AI makes us happy.

Not just accessible, but affordable and easy to set up

Good generative AI doesn’t have to cost the earth or lock you down to one vendor or one solution. For context, enterprise-grade service from ChatGPT is around ten times the price of a Policy Buddy: Open AI ask £100k per year for a fixed number of users. Copilot is £20 per person per month –and that’s before you start looking at the cost of the time your own staff could spend developing their own tools - and then making sure they keep working every time Microsoft tweak their cloud infrastructure. If you make that kind of outlay it gets hard to justify looking for other, smaller, specialist options if your expensive silver AI bullet isn’t quite doing what you need. That’s putting the best tech out of reach for a lot of charities, public sector bodies and SMEs, even though those solutions are designed not for specialist teams but for general use.

Simple, tailor-made solutions can be found from smaller companies and they can be proven and trusted. Sure, we had to navigate a bit of local data and tech governance, but there was no platform reengineering or serious headaches for council staff or their IT support teams, so even implementation and training costs are near zero. And we won’t mind if you tell us you want to buy another tool from someone else to work alongside it. We’re nice like that.

So, what did we learn?

Well, Jerry Springer, we’ve learned that if you follow the golden rules of transformation then AI adoption will happen. If you don’t, it won’t. But there’s one tiny health warning: it still has to be the right AI at the right time. In another universe someone is writing a blog about the rise and fall of Segway, proving you can do everything right – vision, marketing, design… investment - but still fall short if it doesn’t meet a market need, isn’t affordable, or just doesn’t suit your customers’ behaviour. But you can pivot and survive.

Interesting times lie ahead for AI companies as they look to grow the market.

‍