14th May 2025
Data, AI and digital strategies to support the future of housing
By Andy Baker, customer solutions consultant – Data and AI, Civica
In 2024, social housing providers in the UK spent more on repairs and maintenance than ever before. £8.8 billion in fact, which was 13% higher than the previous year and 55% higher than five years ago. Of course, a significant amount of this has been invested in important upgrades around fire remediation, building safety and energy efficiency. It’s also a time when housing associations are in need of developing better financial resilience, so value for money is being scrutinised closer than ever.
Research from LocalDigital, part of the Ministry of Housing, Communities and Local Government, estimates that as much as £400 million is wasted every single year on repairs and allocations that are not delivered correctly. Perhaps there were challenges accessing a property if residents were unavailable, the wrong parts or equipment were brought to the job, insufficient time allocated to complete the work, or the wrong skillsets assigned. The research suggests that these factors can result in 40-50% of responsive repairs not being fixed first time.
Finding ways to reduce these avoidable failures and the costs they carry is an essential part of improving services and delivering value for money in social housing. So, what stands in the way?
AI is often touted as something of a silver bullet, but does it have a role to play here? If so, how? A good place to start is to look at the private sector to see how similar challenges are solved and where AI fits in.
Lessons in predictive maintenance from the private sector
Experiencing a level of failure where up to half of maintenance jobs are not completed first time would quickly set alarm bells ringing in the private sector – both metaphorically and probably literally too. In fact, the alarms would ring at a much lower threshold and long before the issues had a chance to materialise.
Take manufacturing, for example. This sector is reliant on operations continuing without disruption or delay. Keeping assets running – machinery, equipment and vehicles – can mean the difference between success and failure.
Industry 4.0 technologies such as the Internet of Things ensure that a manufacturer’s assets are all connected digitally so that real-time data on their performance can be collected. This data is fed into digital twins that closely monitor operations and use AI to anticipate when maintenance should be completed. It’s a way to optimise the asset and extend its lifetime.
The gains from predictive maintenance like this are huge: 15% reductions in downtime; 20% increases in productivity; 30% savings from carrying lower inventory; 5% reductions in the costs of new equipment – these are just some examples of benefits from adopting predictive maintenance, in this case highlighted by Deloitte.
OK, social housing isn’t going to start adopting these advanced maintenance processes – not yet, at least – so what can we learn, and what takeaways are relevant?
No matter the level of technology available, taking a more predictive approach to maintenance rather than a reactive or periodically planned approach is driven by having reliable asset data. It needs effective data standards in place so that the data is consistent and usable. This doesn’t require state-of-the-art technologies, it can be managed via much more simple software, but this is exactly where the social housing sector is struggling.
Housing data challenges
The LocalDigital research suggests that less than half (45.5%) of landlords trust that the data they have access to is reliable. 54.1% of landlords say that they have been made aware of poor data quality within their organisation over the last twelve months.
One of the most common data issues is having lots of disparate, unconnected records that live in silos. There might be different standards for recording information from one department to the next, or even between one person’s Excel spreadsheet that’s saved on their desktop and the colleague that sits next to them. This makes the data incompatible for sharing, cross-referencing and gathering insights, and almost impossible for AI.
A lot of the time, landlords and social housing providers simply won’t have the tools for collecting and inputting data. Instead, it’s collected manually or using cumbersome legacy technology, which leaves it prone to error and ultimately not fit for purpose. There is also a lack of data skills within the sector and sometimes no training available to change that.
This all leads to poor data quality and prevents providers from unlocking the potential of data.
Data as an asset in social housing
Instead, data should be considered as an invaluable asset for informing decisions in social housing.
We’ve highlighted the role of predictive maintenance above, but its role in improving operational efficiency extends from there. Landlords can use data to optimise staff allocation and resources, leading to greater cost-efficiency. It can be used to ensure compliance around things like health and safety checks or energy efficiency standards by tracking and scheduling essential inspections.
It’s also useful in strategic decision making, such as using demographics, location and tenant data to make smarter decisions about where to invest in new housing developments or refurbishments, ensuring they meet local demand.
This is all only possible with reliable, standardised data collection and analysis.
Aside from operational improvements, the other main objective of data management is improving the tenant experience.
By analysing tenant data, landlords can identify individual needs such as disability requirements or rent payment patterns to provide tailored support. They can predict potential arrears in rent payments or risks of eviction and develop early intervention strategies. They can take a proactive approach to managing tenant enquiries or complaints.
It’s about being able to offer far more personalised services by having a consistent and accurate view of the tenants’ interactions with their landlord and a reliable history of the property.
A single view of the tenant
To borrow insights from another part of the private sector, retail has been a trailblazer in building a ‘single view of the customer’. With this, retailers can deliver consistent and efficient multi-channel customer experiences.
You can order an item on your mobile, turn up to collect it instore and then receive a tailored discount for your next purchase a few days later as an email. If you’ve got a problem with the purchase, AI-powered customer service is available 24 hours a day through online instant messaging services to resolve it. This level of data integration – where the retailer knows it’s you at every touchpoint – has completely transformed the sector. We probably all now take it for granted!
Of course, there is a natural desire that these kinds of seamless experiences are found in all aspects of life – interactions with housing associations and landlords included. So how about creating a single view of the tenant, or a single view of the property? Again, while a joined-up multi-channel approach to delivering personalised services in social housing is not unrealistic, first there are some fundamental data management practices that need to be built.
This means removing siloes and creating a single version of the truth. It means building data literacy and skills to improve self-reliance. It means developing capabilities for internal data governance, data sharing and opening up to integrations. Completing a data maturity assessment should be the first step, and from there creating a data strategy and roadmap.
There is no silver bullet, unfortunately, but part of the later roadmap will probably include AI integrations to support service delivery. Since AI is only as good as the data that feeds it, cleaning up your data is the essential first step.