We recently won the Artificial Intelligence award at this year’s Housing Technology awards. Adam Cresser, director of data and AI tells us more.
We were recognised for the implementation of a disrepair prediction model that my team developed in collaboration with our repairs colleagues.
The aim was to understand which of our properties are most at risk of becoming a disrepair case. Proactively understanding this and taking action helps us reduce cost and improve customer experience.
Making improvements like this through the use of data and technology is a key part of our corporate strategy.
This award was not just for this initiative, but instead a recognition of the capability, approach and the team we’ve built to enable us to use data and technology to make improvements for our residents. In our awards submission we included other examples and demonstrated our ability to deliver future ideas too.
What we did
Our talented team, worked closely with our repairs team to really understand the problem they were trying to solve – a high number of disrepair cases, which we were often managing reactively. This is the worst of both worlds, as there’s a significant cost to us (millions of pounds), and customers are unhappy. Our challenge was to give them something of a crystal ball, to see where they could proactively take action.
We took a range of data sources from our housing management and asset management systems, and external open source data, to create a ranking of likelihood. Our work saw that a subset of around 18% of our homes were accountable for over 50% of disrepair cases. This gave our high risk repairs team a better place to start and concentrate their efforts. This new register of properties with an increased risk of needing repairs updates each month, based on activity so is a live insight into our homes.
The high risk repairs team use that data to proactively drive activities. We’re still working together to continually improve the accuracy (our target is to predict 70% of cases, and to identify which of those will happen over the next three months) and embed how we make the most impact from it.
This initiative represents the things I love about Notting Hill Genesis, when we’re working at our best. We had a clear purpose, focussed on residents and different teams pulling together behind that. Working on a bold and innovative aim, failing fast, learning, and continually improving. Talented people, empowered to make things happen, and just getting stuff done.
Ludovic Basse, high-risk repairs project lead, said, “The analytics and data science team show a drive for continuous improvement and enjoy getting stuck into new processes. By delivering this project and the live nature of the register, we can tailor services, visit residents in need of support during times of escalated risk, and manage damp and mould escalations in line with changing legislation.”
We built our own talented team, introducing new roles to give us the capability and used resourceful ways to attract people, who can use their skills for a real social purpose. Alongside this, we’ve grown a talented group from existing colleagues, our Analytics Community of Experts (ACE), who are taking these ideas and doing similar things in their local areas.
All of this is underpinned by having the data to do it. We’ve invested a lot of work in improving the quality of our data and the systems we capture it in in recent years.
We have a regular cycle of experiments – for example we’re currently working with colleagues in complaints and operations on how we can better identify customers likely to become complainants. Other initiatives have included exploration of using AI to grade and score conditions from pictures, as well as other ideas.
We look forward to sharing more, specifically the impact they’re having, and how we’re using data and technology to continually improve.
