Life is full of unexpected moments. On my first day with Raven Housing Trust, my commute took me past a sheltered housing scheme I hadn't visited in nearly 30 years. I smiled at memories of doing cartwheels on the green, the smell of paint and canvas in the communal art room and helping my grandad water the plants on his patio.

A few hours later, I discovered that that same scheme was owned by Raven.

You might be wondering what this has to do with artificial intelligence (AI). For me, it’s a beautiful reminder that social housing is personal. Behind every repair, ASB case and service charge is a real person living a range of human experiences. That perspective has shaped every technology decision I've made since joining the sector.

It’s both a privilege to do such impactful work, and a responsibility.

It was easy to get swept up in the hype when generative AI first emerged. Across every sector, organisations seemed to be adopting AI at pace, often without a clear plan. In housing, where we hold some of the most sensitive information about people and the decisions we make can affect their homes and lives in the most profound ways, that isn't good enough.

As AI adoption continues to outpace governance and regulation both within our sector and more broadly across business, the most important question is not "how do we use AI?" but "how do we use it responsibly?"

At Raven, we're embedding a layered governance framework around AI, combining policy, technical controls and oversight. But when it comes to how AI is used and the decisions that follow, our most important control is, and will continue to be, the human-in-the-loop. People remain accountable.

That consideration shaped how we used AI during our recent stock acquisition, with controls built into the process and one practical question: where can AI help most, without replacing human judgement?

Last month Raven completed the second largest stock acquisition our sector has seen, bringing more than 1,900 homes into our organisation. Each came with vast quantities of data and documentation that needed to be reviewed, transformed, validated and migrated into our system. We had to do this within challenging timescales to ensure we were ready to deliver the same high standard of service to our new customers from day one.

Our Digital, Data and Technology (DDaT) team used Microsoft Copilot within our secure enterprise environment to help build a data transformation pipeline. We extracted information from thousands of PDF and Word documents and reconciled hundreds of thousands of rows of data. Tasks that would previously have taken weeks of manual effort were completed in a fraction of that time. This helped a lean in-house team deliver a highly complex programme quickly.

Was it perfect first time? Absolutely not.

Like many organisations, we experienced our fair share of experimentation, learning and refinement. But the most important takeaway wasn't about the technology. It was about governance.

Every output was reviewed. Every result was validated. A member of the DDaT team approved every data transformation rule before it was applied. We used Copilot to run validation checks on its own work, but we also operated a human peer review process to assess the quality and accuracy of every output before it could be progressed. Accountability for every stage sign off remained with a person, and so did the inevitable decision to stop and try again when things weren’t going in the right direction.

Humans set the context and desired outcomes, designed the process, and signed off on the outputs. Copilot did the heavy lifting.

AI helped us work faster, but it didn't replace professional judgement. That human-in-the-loop approach gave us the confidence to expand its use across the programme, applying it where it added most value whilst maintaining ownership, accountability and quality throughout.

The challenges facing our sector are becoming increasingly complex and we need to think differently about how we deliver services. Used responsibly, AI can help us respond faster, make better use of our resources and free colleagues from repetitive administration. That way they can focus on work that delivers real value.

On 12 October 2026, Raven welcomed almost 5,000 new customers. The data and systems were ready. But most importantly, Raven’s people were ready and able to serve our new customers.

Could we have completed the acquisition without AI? Of course. Given enough time, enough resource and enough skill.

But that's not the real question. The real question is whether AI helped us deliver a better outcome for customers, more quickly, and while maintaining the controls and assurance such a complex data transformation required.

The answer is yes. In social housing, technology should never be about the technology itself. It should always be about the people it serves. It's personal.