Public finance turns to AI

Public finance turns to AI

Estonia’s Ministry of Economic Affairs and Communications has implemented an open source, AI assistant named Bürokratt

Microsoft’s Valentina Ion explains why AI plays a critical role in helping public finance organisations transform operations and deliver citizen-focused services while driving productivity and efficiency  

By Rebecca Gibson |


Reported fraud, ineligible claims and errors cost US taxpayers up to $521 billion per year, according to estimates from the US Government Accountability Office. Meanwhile, figures from the European Public Prosecutor’s Office suggests countries in the European Union lost around €13 billion ($15 billion) to public finance fraud in 2024. And the National Audit Office approximates losses were between £55 billion ($73 billion) and £81 billion ($108 billion) from 2023 to 2024 in the UK.   

However, the true cost of tax evasion is likely being vastly underestimated, as fraudulent behaviour often goes undetected and unreported. With public finance budgets being squeezed and the cost of living rising, it’s vital for public finance organisations to find new ways to mitigate fraud and corruption. Many are arming themselves with AI.    

“Organisations can use AI to collate data from traditionally siloed, disconnected sources to build a 360-degree view of individual taxpayers and analyse their behavioural patterns and historic transactions to create accurate risk profiles,” says Valentina Ion, global public finance and social services lead at Microsoft. “These insights also allow them to rapidly identify abnormalities that may indicate errors, fraud or tax evasion, so they can proactively address these issues and reduce revenue loss.”   

Valentina Ion

Valentina Ion is global public finance and social services lead at Microsoft

The Inter-American Center of Tax Administrations (CIAT), for example, has co-developed the e-Invoicing Anomaly Detection Solution Accelerator (e-IAD) with Microsoft to help national tax agencies in its 42 member countries enhance electronic invoicing and reduce tax fraud and evasion.   

Powered by Microsoft Azure, e-IAD uses AI, machine learning and advanced analytics to automatically scrutinise taxpayer behaviour and analyse data from multiple sources to detect anomalies, validation errors or fraud in tax declaration forms and the e-invoicing process. The solution also provides personalised guidance to auditors, allowing them to make faster and better-informed decisions and optimise resource allocation. e-IAD, which can be scaled as tax regulations change and agency needs evolve, was successfully piloted by Costa Rica’s Ministry of Finance and is being rolled out to national tax agencies in three additional CIAT member countries.   

Fighting fraud is just one of the ways AI can help public finance organisations transform their operational processes to overcome the myriad of challenges they face, says Ion.   

“These agencies are under immense pressure to maintain financial stability so local and national economies continue to thrive, despite converging geopolitical, societal and environmental crises,” she says. “Many organisations are dealing with challenges caused by budget cuts, the ageing workforce, skills shortages and other factors, so they’re trying to decrease costs, increase operational efficiencies and achieve more with fewer resources. They can achieve all of this with AI tools.”  

Solutions such as Microsoft 365 Copilot for Finance, for example, use generative AI to automate many of the routine, time-consuming administrative tasks employees typically carry out manually.  

“AI can automate processes for tasks such as data consolidation and analysis, revenue and tax collection, or procurement and grant management,” says Ion. “This saves time and also removes the administrative burden from employees, freeing them up to focus on higher-value tasks.”  

Organisations can also use solutions like Microsoft Azure OpenAI Service to develop AI agents or virtual assistants, which can quickly extract key information from multiple structured and unstructured data sources to equip employees with the detailed insights they need to complete tasks more efficiently.   

“An AI agent could help an employee to analyse a company’s financial data alongside public finance policies and legislation to expedite the auditing process and ensure compliance,” suggests Ion. “Meanwhile, AI agents in contact centres can automatically summarise insights from conversations, determine citizen sentiment and recommend the best actions for employees to take in real time, enabling them to hyper-personalise interactions and increase first contact resolution rates.”  

Public finance agencies can also build self-service portals for citizens, who can interact with AI agents to get quick answers to basic queries or assistance with completing processes such as paying for permits or submitting tax declaration forms.    

In Estonia, the Ministry of Economic Affairs and Communications has implemented an open-source, AI assistant named Bürokratt to give citizens access to secure information about vital government services in seconds rather than days. Now, rather than filling out paperwork or visiting different government offices to resolve their queries in a process that could take up to two days, citizens simply ask a question via Bürokratt and it is automatically forwarded to the relevant state organisation. Bürokratt, which was developed for the ministry by the Information System Authority of Estonia, Microsoft Industry Solutions Engineering and Net Group, has been rolled out to six institutions and is likely to be expanded to others in future.  

Estonia

Citizens in Estonia can access any government service within seconds by using the AI-powered Bürokratt chatbot

“Citizens have one single point of contact for the entire government, making it quicker and easier for them to access the information or services they need,” says Ion. “This increases citizen engagement and satisfaction rates. This implementation also shows the crucial role AI plays in enabling fast and secure data sharing across different public finance and other government agencies.”  

Having access to AI-powered business intelligence and real-time data insights also improves forecasting and decision-making. This is particularly beneficial for budget and treasury agencies, which are responsible for disbursing public funds in ways that measurably drive economic development. “These organisations must allocate the right financial support to the right beneficiaries at the right time, and they can achieve this much more easily by using AI to automate and streamline payment, bid management, budget distribution and other processes,” says Ion.    

One organisation using AI to streamline grant management is the Rural Payments Agency (RPA), an executive agency of the UK’s Department for Environment, Food and Rural Affairs (Defra), which distributes agri-environment payments to thousands of farmers, traders and landowners in England every year.   

Working in partnership with Hitachi and Microsoft, Defra migrated RPA to the cloud-based Microsoft Dynamics 365 Finance platform and integrated Power BI and Power Apps to create a customer portal, automate the claims entry process and facilitate real-time reporting. Now, RPA can handle up to 50,000 claims per day and within the first three months of the new platform going live, it had processed 111,000 payments totalling nearly £2.5 billion ($3.3 billion). This amounts to more than 40 per cent of its annual 250,000 payments.     

“Moving to Dynamics 365 for Finance has made it easier for farmers and others to access public funds, while cutting RPA’s operational costs and significantly improving productivity and efficiency,” says Ion. “The solution can be easily scaled to incorporate future payment schemes, which will drive economic development.”   

Farming

AI tools are helping the UK Government’s Rural Payments Agency to rapidly distribute agri-environment payments to thousands of farmers, landowners and others every year

To capitalise on the many advantages of AI, organisations must modernise their data estates.  

“They need to migrate from legacy systems to the cloud, integrate multiple internal and external data sources to eliminate silos, and classify their data to ensure it is high quality,” explains Ion. “Once they have the right foundation in place, organisations can easily leverage AI to re-engineer and automate processes, democratise business intelligence and analyse their data to derive real-time insights and operate more productively, efficiently, cost-effectively and proactively.”  

Microsoft and its “strong” ecosystem of partners continually develop innovative solutions to support organisations with everything from tax revenue collection to customs and trade facilitation, grants management, procurement and more, says Ion. “Partners like Quantexa help organisations build a 360-degree view of taxpayers to improve risk profiling, compliance and hyperpersonalisation. Meanwhile, partners such as Accela are making it faster and simpler for taxpayers to access licensing and permitting processes, while those like Icertis are simplifying contract and procurement management. Other examples include Capgemini, which can assist with integrating multiple data sources and manage complex technology implementations, and EY, which can advise on how to update policies to cover new technologies.”  

In addition, Microsoft and its partners provide advanced cybersecurity and governance capabilities designed to support organisations in protecting systems and managing AI data access securely. 

“Public finance organisations face significant cybersecurity risks, making strong data governance and security critical components of any AI implementation,” says Ion. “Some solutions from Microsoft and our partners are designed to minimise vulnerabilities across an organisation’s IT environment, while others empower them to rapidly detect threats or attacks, isolate the targeted asset and resolve the issue to prevent widespread damage.”  

Ion advises organisations to follow Microsoft’s lead by adopting the Zero Trust framework, which enables them to lock systems down by default and prevent employees from unintentionally – or intentionally – introducing vulnerabilities cybercriminals can exploit. “They should also establish clear guidelines to govern how employees interact with data and AI tools according to their specific roles,” she adds. “The principles and best practices outlined in Microsoft’s Responsible AI Standard, AI governance framework and Secure Future Initiative can help with this.  

“Furthermore, organisations should reskill employees so they understand how AI tools work and how to write effective prompts to get the best output. Similarly, it’s beneficial to educate citizens on how to use self-service portals and AI tools. There are multiple examples of governments that have leveraged Microsoft’s training and educational courses for this purpose.”   

With the public finance sector now recognising the benefits of AI – and how to successfully and securely implement it – Ion predicts adoption rates and use cases will increase in 2026.    

“We’ll see more organisations rolling out AI for the first time, while those that have already implemented the technology will begin scaling it across different departments and fully integrating it within their systems and teams,” she says. “We anticipate a future where humans collaborate with multiple AI agents across agencies to improve efficiency and citizen experiences.   

“Together with our partners, Microsoft will continue to empower public finance agencies with the technologies, solutions, knowledge and skills they need to optimise AI and transform into citizen-centric organisations that can easily and securely collaborate and exchange digital data with stakeholders across the wider financial ecosystem in real time. This will help the sector overcome the challenges it is facing, introduce innovative new services and ultimately lead to thriving economies and societies.”  

Partner perspectives 

Three Microsoft partners share how they are using Microsoft’s AI and cloud technologies to help drive greater efficiencies within public finance 

“According to the Tax Justice Network, countries lose more than $450 billion in yearly tax revenue to international corporate tax abuse and private evasion,” says Gabriel Bellenger, consulting and global revenue industry lead at Accenture. “Strengthening digital infrastructures is key to reversing that trend. In Asia, we helped governments modernise their IT systems by implementing eComply and Verify – transforming VAT reporting and compliance and enabling tax authorities to detect issues faster and act with confidence. As digital ecosystems mature, compliance will move upstream, supported by traditional and emerging technologies like agentic AI. This is essential to ensure a resilient tax environment amid the wave of agent-based interactions currently reshaping administrative processes.” 

“Tax and customs administrations are among the most digitally advanced public institutions, on par with innovation in the private financial sector,” says Karl Heinz Krug, vice president and industry lead for public finance in Europe at Capgemini. “Their operations generate vast amounts of data through tax filings, customs declarations and risk assessments. Leveraging advanced AI and cloud technologies, governments can enhance citizen-facing services, simplify compliance and gain comprehensive insights into taxpayer behaviour. Microsoft’s Azure platform provides secure, scalable solutions that empower public finance authorities to improve agility, transparency and operational efficiency, which supports more effective revenue collection and responsible financial management. These capabilities are essential for modernising public finance and delivering better outcomes for society.” 

“Quantexa helps public finance organisations make better decisions by connecting and understanding their data,” says Jerome Bryssinck, vice president of government solutions at Quantexa. “Using advanced analytics and AI, governments can strengthen financial integrity, modernise outdated systems and deliver services that are more efficient and responsive to citizens’ needs. Quantexa’s Decision Intelligence Platform promotes responsible and transparent AI use, supporting compliance with global standards such as the European Union AI Act. By creating a clearer picture of risk and opportunity across departments, Quantexa enables collaboration and smarter policymaking, helping public institutions build the trust, agility and resilience needed to navigate an increasingly complex global environment.” 

Discover more insights from these partners and others, in the Winter 2025 issue of Technology Record. Don’t miss out – subscribe for free today and get future issues delivered straight to your inbox.   

Subscribe to the Technology Record newsletter


  • ©2025 Tudor Rose. All Rights Reserved. Technology Record is published by Tudor Rose with the support and guidance of Microsoft.