Your boiler’s on the blink, you’re behind on the rent and your housing officer’s phone line is engaged! These challenges could be a thing of the past as Artificial Intelligence (AI) becomes more of a reality.
Your boiler’s on the blink, you’re behind on the rent and your housing officer’s phone line is engaged! These challenges could be a thing of the past as Artificial Intelligence (AI) becomes more of a reality. In the future, it could be as simple as inviting a digital assistant into your home and leaving them to alert the housing association. This concept is not so futuristic! Already housing providers are making giant leaps with AI to improve the customer experience.
Back in 1997, IBM’s Deep Blue, one of the world’s first AI creations, defeated world champion, Garry Kasparov in a chess match. At the time, many refused to believe that a machine could play with such a powerful human touch and defeat humanity at its own game.
21 years after Deep Blue’s controversial victory, AI is poised to serve humanity, not defeat it, by transforming housing management. This may be a bold statement, and there’s certainly no shortage of hype around AI. But now that the technology has reached the age of maturity, we are seeing the positive impact of AI in a wide range of business sectors, from cancer diagnostics in medicine to fraud prevention for banks.
And in April 2018, the Government put its weight behind AI when it announced its AI Sector Deal, a move to invest in research, people, places and infrastructure to help lead the transformation of AI technology.
So we ignore AI at our peril, but the question remains, does AI really have the potential to transform housing?
For some years now, public sector digital strategy has focused on driving customers towards online channels, and away from telephone and face-to-face interactions, in a quest to cut costs and improve the customer experience. One of the integral tools of this approach has been the chatbot.
Using natural language interfaces, chatbots enable customers to ask questions, give instructions and confirm information by speaking. The bots usually work best when these conversations have a pre-set order. So if your transaction with the housing provider is simple, paying a bill, say, or booking a repair, the interaction is straightforward and no human intervention is needed.
However, the problem arises when a customer has a more complex enquiry, needing to discuss their rent arrears or benefits payments for example. In these cases the bot interaction often has to be abandoned, the digital transaction reaches the end of the road, and the customer resorts to picking up the phone.
To overcome these limitations, some customer-facing organisations in sectors such as financial services and banking are turning to a new generation of chatbots, known as conversational AI.
Conversational AI provides another layer of AI altogether, which harnesses the power of machine learning, speech synthesis and natural language understanding to get to the heart of the customers’ needs and respond appropriately. It can even cope with the complex sentences and idiosyncrasies of human speech.
The game-changer with conversational AI, and the reason it has such a significant capacity to transform housing transactions, is that the technology has the capability to use, and learn from, all the content on the organisation’s website. This means that the AI actually builds a detailed knowledge bank, which it can add to on an ongoing basis, to help customers.
Customers frequently need to interact with their housing officer on a range of different topics, involving the condition of their property, repairs to appliances and payment details. They don’t want to have to navigate through pages of information that aren’t relevant to them before they find what they need.
Where an organisation has embedded conversational AI into its system, a customer can log on to the housing organisation’s website and ask questions using either text or voice, and the AI tool can pull together all the information which is buried deep in a website to provide the customer with an answer.
Of course, there will always be times where no machine can be a substitute for the human touch. But unlike the traditional chatbot which follows set pathways through an interaction, conversational AI can identify when a customer is vulnerable and may need the additional help of a real agent. The customer can then be helped with human intervention.
New innovations in technology don’t tend to come cheap. But if you consider the difference between types of customer transactions, it’s interesting to see that, taking the long view, an AI-based solution can help to drive savings.
According to SOCITM, in a digital investment advisory piece, it can cost as much as £14 to serve a customer in person, and £5 on the phone. Traditional web chats, where an agent interacts online with a customer, cost a similar amount. While a conversational AI chat can be as little as 20 pence per conversation.
With these cost benefits, it will be interesting to see how this trend in conversational AI will continue to develop in the housing sector, but its impact could be felt sooner rather than later. These intelligent conversational interfaces are already making waves in customer service led businesses because the technology is already available.
So in terms of customer service, conversational AI could provide some relatively quick wins. We may, however, need to take the long view when it comes to embedding AI into other areas of the housing sector, such as asset management.
It has long been recognised that AI has the potential to enable housing providers to manage their stock more efficiently. Neural networks, or machine learning, which falls under the umbrella of AI, can tell us a lot about the condition of a property without a human officer needing to go through the door.
Currently, housing organisations rely on planned maintenance programmes to ensure their properties are in good condition and everything is running smoothly. When problems arise, it is up to the customer to contact their housing officer and inform them that there is a defect, and a repair needs to be carried out.
This can mean that in some cases, defects are not reported and as a result, a property can quickly fall into disrepair.
However, AI has the capability to predict when a boiler is likely to fail, how long a household appliance will last and if a house is at risk of becoming damp. This takes the onus away from the tenant because, with this information, housing providers are able to plan repairs and maintenance much more effectively. Savings can be made by avoiding unnecessary repairs, while customers benefit from appliances that work well because they are being monitored for their performance.
Building AI into asset management will help housing providers make the shift from ‘defects and repair’ to ‘predict and prevent’. The result of this shift will be significant cost savings on planned maintenance, better information to inform the 30-year modelling process, and a reduction in the need for many repairs in the first instance.
While the application of AI in managing housing stock has enormous benefits, it’s a long term project because AI asset models take time to build and validate, and to connect to existing assets.
However, given time, this is an area where AI will ultimately be transformational.
Part of the challenge is knowing at what point investment in AI is likely to have the greatest returns. In some areas, such as customer service, developments such as conversational AI are ready to make an immediate impact, while in others, like asset management, we are taking a future view.
It may be 21 years ago since the world first realised the power of AI to defeat an opponent in chess; but now are we truly seeing what AI can do to solve problems, save resources and improve lives.
Find out more about NEC Housing.