AI Week: Automation makes work less taxing for HMRC
In the last few years, the UK tax agency has automated millions of processes and it has ambitions to grow its use of AI into new areas. Tim East, CIO of customer compliance, tells PublicTechnology about the department's plans
There are probably few organisations with as much to gain from automation and artificial intelligence technologies as HM Revenue and Customs.
The department employs 65,000 people to collect an annual tally of £600bn in tax from citizens and businesses across the UK. All the while, it is in the midst of a 10-year transformation programme during which, by 2025, its workers will be redistributed around 13 regional hubs throughout the country.
All of which, surely, adds up to a lot of processes to automate.
And the tax agency has a clear – and long-standing – willingness to embrace AI and robotics.
HMRC’s Automation Delivery Centre was set up in 2016 and, within two years, had achieved its goal of processing 10 million automated transactions.
It has grown exponentially since then; the department’s annual report for 2018/19 revealed that, over the course of the year, some 15.7 million transactions had been completed across 78 separate automated processes.
The enthusiasm for AI stems from the top down. In an interview with PublicTechnology earlier this year, HMRC’s chief digital and information officer Jacky Wright picked out AI and machine learning as technologies that she wanted to see the department explore further.
“I think we need to focus more on AI and machine learning and maturing that,” she said. “Because that will unleash the power of the data we have… so we have more insight, and we're able to anticipate the types of things we need to think about: better citizen-centric services to collect taxes and that fund the public services that we need to provide.”
Risk and compliance is an area where the tax authority has already deployed AI. PublicTechnology caught up with Tim East, chief information officer for HMRC’s customer compliance group to discuss the benefits it has seen so far, and the department’s use of the technology might develop in the future.
PublicTechnology: How did HMRC first come to believe that artificial intelligence could play a role in its risk work?
Tim East: Artificial intelligence isn't new. HMRC has been using the advanced analytics technology that AI brings for at least a decade. However, the technology is maturing and, in particular, cloud computing is making analytics tools more versatile now than ever before.
There is a growing conversation in industry around the ethical adoption of AI and what that means. In HMRC, we set up a working group across our organisation to build greater awareness around the ethics issues and consider the governance we need.
Which operations or tasks were identified as those that might benefit from the introduction of AI?
We use AI to support a number of things we do. Examples include: identifying risks on some of our large-scale transactional services, such as on repayment claims for VAT and self-assessment; using analytics to help us identify risks that need attention and building case packages that are passed to our teams of investigators; and assimilating large amounts of data – this is a newer implementation, important for our compliance casework where we’re using AI alongside other tools like geo-mapping.
How did the department go about assessing the available technology, and deciding what would be a good fit for your needs?
I think our approach is probably quite standard, and similar to larger commercial organisations. Within our IT function, we have an Architecture and Innovation team that horizon-scans for emerging technologies. As well as this, we monitor trends and engage with vendors about solutions that look interesting. We're also exploring and making more use of open-source products too where we can. Typically, we'd run one or more pilots to make sure we've fully tested our thinking before any larger-scale implementations.
Were there any barriers – whether technical, cultural, or procedural – that needed to be overcome?
From a technical perspective, cloud computing is removing many of the barriers. However, there is a growing conversation in industry around the ethical adoption of AI and what that means. In HMRC, we set up a working group across our organisation to build greater awareness around the ethics issues and consider the governance we need. We recognise that being able to explain how we are using AI is very important in terms of maintaining the trust of our customers. And, as AI technology matures further, it will undoubtedly bring different ways of working, which will bring different cultural and educational challenges.
Describe how the technology has worked in practice, and what the biggest benefits have been thus far?
Our social networking analysis system, Connect, was ground-breaking across industry when we implemented it 10 years ago and we're now refreshing it. Connect underpins HMRC's compliance activity. For example, it helps us secure more than £1.2bn additional yield per year, and contributes to further reducing the tax gap.
How has it impacted the working lives of staff?
As already mentioned, this isn't a new area for HMRC as we've been using AI for a while. But, as the technology is maturing, it’s becoming easier to analyse larger amounts of data and pull out insights. Therefore, our analysts are increasingly able to do more and to work with larger amounts of data than ever before.
Will AI continue to play a role in HMRC’s risk and compliance work and, if so, how will this evolve?
Yes, AI will definitely continue to have a role. As ways of doing business develop, the trend is for more and more electronic connection and distribution of data, which moves us towards other interesting areas like distributed ledger technology. Understanding that, and understanding how our customers and their advisors and agents are using AI, potentially means there is new territory for us to get to grips with. As tax agencies and accountancy businesses begin to use AI more frequently with their clients, then it follows that HMRC needs to be able to understand how it is being used to make sure the tax outcomes are right.
How might the rest of the department make use of AI, and does HMRC have any plans to trial the use of the technology in other areas?
We're already exploring the use of AI in other areas of our work and, over time, we can expect to see more integration of the various tools we’re using. One key area is improving customer experience, where we are looking at using AI to improve how we: recognise why a customer is making contact and direct them to the most appropriate channel; predict ‘triggers’ for customers contacting us, for example a tax code change, and provide interventions so that they no longer need to; and make it easier for our teams – for example, our contact centre advisers shouldn’t need to open multiple IT systems when dealing with calls, and instead be able to input comments or process changes once into all of them through more automated routines.
This article forms part of PublicTechnology’s dedicated AI Week, in association with UiPath. Look out over the coming days for lots more content. Tomorrow, an exclusive webinar discussion – in which a panel of private and public sector experts will debate all the major issues related to government's use of AI – will be available to view on demand. Click here to register to do so – free of charge.
Tax agency head salutes colleagues’ work in building new platforms at speed
Cabinet Office annual report shows digital agency also brought in more than £2m in extra revenue
Dominic Cummings’ ambitions realised as prime minister’s office seeks to recruit crack squad of data scientists for No. 10 skunkworks
Cabinet Office minister said that, despite the controversy that often surrounds the PM’s top adviser, ‘people are interested in Dominic and his ideas’