At Social Care Wales, we’re always looking for ways to improve how we support the people who rely on our services. Like many public sector organisations, we’ve been exploring how artificial intelligence (AI) can help us do just that – more efficiently, more accessibly, and more transparently.
One of our first steps into this space has been the development of a chatbot to support users with qualifications-related queries. It’s a tool designed to help people find the information they need, when they need it, without having to wait for phone lines to open or emails to be answered. But just as important as the technology itself is the way we’ve built it: with safety, fairness, and transparency at its core. That’s where the Algorithmic Transparency Record (ATR) came in.
Meeting real needs with smarter tools
The need for this chatbot became clear as we looked at the challenges facing our enquiries service. Since the pandemic, our phone lines have been open for fewer hours, and during busy periods, it’s been difficult for users to get through.
At the same time, our team was spending a lot of time responding to repetitive queries, many of which could be automated. We estimated that around half of all qualifications-related questions could be handled by a digital tool. That’s time we could be using to support more complex or sensitive enquiries.
We also knew that not everyone prefers to pick up the phone. Some users find it easier to search online or use digital tools, especially outside of standard working hours. Our chatbot was designed to meet those needs, offering a more flexible, accessible way to get answers.
A chatbot that’s inclusive, accessible, and built for the sector
The chatbot is now live on the “Find a Qualification” pages of our website. It’s designed to support people working in social care, those looking to join the sector, and other stakeholders by helping them quickly find information about qualifications, registration, and career pathways.
It works by searching a structured database of job roles and qualifications that are part of the Qualifications Framework for social care and early years. All the information it uses is already publicly available on our website, so there’s no risk of it sharing anything sensitive or confidential.
What makes it especially user-friendly is its ability to understand everyday language. Whether someone is asking about funding, where to complete a qualification, or whether their current qualification is valid for registration in Wales, the chatbot can guide them to the right information instantly.
We’ve also made sure the chatbot is inclusive and accessible. It supports both Welsh and English and includes a speech-to-text function so users can speak their questions if they prefer. After each interaction, users are invited to leave feedback, which our Digital Team reviews to help us keep improving the service.
The role of the Algorithmic Transparency Recording Standard
While the technology is exciting, we knew from the start that building trust was just as important as building functionality. That’s why we used the UK Government’s Algorithmic Transparency Record as a framework throughout the development process.
The ATRS helped us ensure that our supplier had the right credentials and could clearly explain how the chatbot works. It guided us in structuring the chatbot’s responses, making sure users are informed about how the tool operates and where it might influence decision-making.
It also helped us get our internal processes right, from assigning accountability and day-to-day management, to making sure users know how to raise concerns if they’re unhappy with a response. The ATRS gave us a central place to document all our impact assessments, including Welsh language and equality, diversity and inclusion (EDI) considerations.
In short, it gave us confidence that we were doing things properly, and it gives our users the same reassurance.
Looking ahead
The chatbot is part of a wider pilot to explore how new technologies can enhance our customer service. By making it available to a broader audience, we’re gathering valuable feedback that will help us refine its features and better understand how people want to interact with us.
We’re optimistic about what this means for the future. By reducing the number of queries that require human intervention, we can free up our team to focus on more complex cases. And by offering a more accessible, responsive service, we can make it easier for people to get the support they need, when and how they need it.
The Algorithmic Transparency Record has been a key part of that journey. It’s helped us build a tool that’s not just smart, but also safe, fair, and open. And in a world where public trust in AI is still being earned, that feels more important than ever.
