What is HMRC’s Connect AI System?

Connect is a data analysis IT system which compares multiple data sets to provide HMRC teams with profiles of taxpayers who are at higher risk of tax evasion. It has been described as a ‘digital detective’ by using big data and AI to help HMRC make proactive targeting decisions around its anti-evasion activity. 

What is HMRC’s Connect system? 

HMRC’s Connect system was introduced in 2010 but its functionality has increased incrementally since then. It initially cost £45 million to implement but the intelligence it has provided has yielded billions of pounds of otherwise uncollected tax since then. In the 2024/25 tax year, leads provided by Connect helped to generate an additional £4.6 billion in tax revenue, providing a significant inroad into the tax gap.  

The overall purpose of HMRC’s Connect system is to detect tax fraud and tax evasion and disrupt criminal activity. Connect’s algorithms make links between different activities and taxpayers from the separate data sources it interrogates, identifying patterns outside the expected norms. This gives HMRC’s compliance officers taxpayers to proactively target and investigate. The AI-powered system provides risk profiles and analysis in a fraction of the time that humans would need to do the same research. 

A byproduct of HMRC’s use of Connect is that with HMRC staff undertaking more targeted investigation, compliant taxpayers should be less likely to face mistaken or random enquiries into their tax affairs. 

How does the Connect system work? 

By cross-referencing 22 billion lines of data, Connect provides HMRC officers with 500,000 taxpayers to enquire into each year. Developed alongside BAE Systems Applied Intelligence, HMRC also collaborates with Capgemini and the SAS Institute to get the most from the system as societal, AI and taxpayer lifestyle trends develop over time. 

There are two main layers to Connect’s functionality: 

• mapping connections across different data sets 
• statistical analysis of the data to spot anomalies from the ‘norm’ 

Analysis picks up on factors like bank interest received in bank accounts far exceeding the bank interest reported on a tax return, or living a more lavish lifestyle than reported income could allow. 

A key factor in Connect’s success is the data sharing agreements HMRC has with other organisations, allowing it to analyse huge volumes of third-party data alongside its own. 

What data sources does HMRC’s Connect system use? 

There are five key categories of data which HMRC’s Connect system uses:  

1. HMRC’s own data e.g. tax returns 
2. UK government databases e.g. Land Registry, Companies House, Department for Work and Pensions (DWP) 
3. Banks and financial institutions 
4. Global exchanges i.e. Common Reporting Standard (CRS) data e.g. data from OECD partner countries on offshore bank accounts held by British taxpayers 
5. Open-source, social and e-commerce data e.g. etsy, ebay, Amazon, Zoopla users’ transactions 

These data sources are used to show inconsistencies or attempts to hide income or transactions which should be subject to UK taxes. The anomalies then form the high-risk cases which HMRC can investigate further.  

How does HMRC use Connect as an anti-evasion tool? 

HMRC uses Connect’s big data analysis to highlight risk areas and drive tax compliance campaigns, as well as preventing tax fraud before it occurs. 

Risk profiling and targeted campaigns 

Connect analysis has helped HMRC to focus on areas of tax where avoidance and evasion may be harder to detect or more widespread. Campaigns so far have included: 

• Let Property Campaign – an amnesty for buy-to-let landlords who haven’t declared their rental income, alongside toolkits to help landlords get their tax affairs correct in future 
• Help for Hustles –  information about tax obligations to people using online selling platforms, providing services such as tutoring, dog walking and gig economy roles like delivery riders, as well as social media content creators  
• R&D tax relief claims – increased compliance activity producing reduced error and fraud levels; warning letters to mail order retail businesses being targeted wrongly by R&D claim companies 
• Wealthy individuals - using CRS data to understand the on- and offshore income of high-net worth individuals, to increase the tax yield from this group 

Tax repayment fraud prevention  

Historically VAT repayments due to businesses have been risk assessed in real time by the Transaction Risking Upstream in a Connect Environment (TRUCE) system. TRUCE looked for irregular or unusual patterns of repayment request and put a hold on the repayment process until further checks could be made that there was no fraudulent intent. This has now been replaced by the Investigation and Detection Risking Service (IDRS) for VAT and self-assessment income tax repayment requests. This proactively prevents fraud rather than working to detect tax fraud after it has occurred. 

What does this mean for taxpayers? 

While Connect routinely analyses publicly available data, HMRC say they only use Connect’s AI to trawl the private social media accounts or emails of taxpayers as part of criminal investigations, and not for the general public. However as Connect has this ability, coupled with HMRC’s power to use it (within its responsibilities under the data protection rules), it creates a deterrent effect for would-be tax evaders. HMRC is accountable to the Information Commissioner’s Office as a data controller. 

Mitigating Risk 

• There are steps you can take to reduce the risk of coming to the attention of HMRC via Connect. 
• Get your house in order – ensure you are reporting all your income in a timely and accurate fashion 
• Maintain a clear paper trail – or better still a cloud-based digital bookkeeping solution 
• TaxAssist Accountants can review your tax affairs and advise you of any actions needed 

Need help? 

Call TaxAssist Accountants on 01494 718 777 or via our contact form here.

First published 4 Nov 2025

Share

This article is intended to inform rather than advise and is based on legislation and practice at the time. Taxpayer’s circumstances do vary and if you feel that the information provided is beneficial it is important that you contact us before implementation. If you take, or do not take action as a result of reading this article, before receiving our written endorsement, we will accept no responsibility for any financial loss incurred.

---