Microsoft 365 Copilot could save every civil servant working for the UK Government almost two weeks per year, according to findings from the 2025 Microsoft 365 Copilot Experiment: Cross-Government Findings Report.
During a trial involving 20,000 civil servants between September and December 2024, the government found that each individual using the AI tool to assist with administrative tasks such as drafting documents, updating records and preparing reports saved an average of 26 minutes per day. Over the course of a year, that’s nearly two weeks saved per person, as well as significant cost savings. In fact, the UK Government believes it can save £45 billion ($61 billion) by investing in more generative AI technology to automate the delivery of services across the public sector, migrate service processing to cheaper online channels, and reduce fraud and error with digital compliance solutions.
The results of a pilot study in the UK suggests civil servants could save two hours per week every year by using Microsoft 365 Copilot to assist with basic administrative tasks
“Generative and agentic AI technology is ideal for empowering employees to work more productively and efficiently, all while cutting costs and improving service delivery,” says Kirk Arthur, worldwide government solutions lead at Microsoft. “Government organisations worldwide are facing geopolitical instability, increasing bureaucracy, economic austerity, staff shortages and many other challenges, so AI is quickly becoming the solution they need to continue delivering high-quality health, social care, law enforcement, transportation and other services to the public.”
Governments have many opportunities to automate administrative tasks that staff currently do manually, says Arthur.
“Tasks such as finding information, analysing data, organising Microsoft Teams meetings, transcribing notes and writing emails, presentations, surveys, reports, policy documents, training materials and other pieces of content are essential but tedious, time consuming and difficult for some individuals,” says Arthur. “Today, employees spend hours on these tasks, which prevents them from concentrating on higher-value work that urgently needs their attention. This causes employee dissatisfaction and burnout, and it negatively impacts service delivery and the citizen experience. However, using AI tools like Microsoft Copilot to automate part or all of these tasks is a game-changer.”
Employees of Somerset Council in the UK, for instance, are saving an average of 10 hours per week by using Microsoft 365 Copilot for administrative tasks. This has not only boosted productivity and efficiency but has also improved the quality and consistency of employees’ work and freed them up to carry out more in-depth tasks. Consequently, the council can now deliver better services to the public, even as it navigates operational challenges from reduced funding and policy changes.
The Australian Government reported similar benefits following a six-month whole-of-government trial of Microsoft 365 Copilot. Participants estimated they saved up to one hour when summarising information, drafting documents or searching for information, with 40 per cent of them saying they were able to reallocate time to higher-value activities like staff engagement and strategic planning. In addition, 40 per cent of managers perceived uplifts in efficiency and quality in their teams.
Government-led organisations elsewhere in the world are using solutions powered by Microsoft Azure OpenAI Service to alleviate the administrative burden on employees. “Many of our partners, such as Axon Enterprise (Axon), Esri and SAS, are embedding capabilities from Azure AI Foundry into their solutions to create new tools and solutions that are transforming how we work,” says Arthur.
Axon, for example, has used Azure OpenAI Service to develop Draft One, which automatically transcribes audio from police officers’ body-worn cameras and writes a draft incident report within minutes. The AI model is calibrated to strictly follow the audio to prevent embellishments or speculation, but officers are required to edit text in places and must approve the report to ensure accuracy.
According to Axon, most police officers using Draft One have halved the time it takes to write reports, with some agencies recording up to an 67 per cent decrease in time spent on this task. Draft One has also increased job satisfaction and boosted recruitment rates. In addition, in a double-blind study, 24 independent evaluators concluded that reports written with the help of Draft One performed equally well as those written solely by officers in three of five categories, and better in two.
“Law enforcement officers in the USA spend up to 40 per cent of their shifts drafting reports, but using AI to automate the administrative aspects of this work frees them up to carry out higher-impact tasks, such as engaging with the public or conducting investigations,” says Arthur.
Kirk Arthur is worldwide government solutions lead at Microsoft
US-based Microsoft partner Geographic Solutions has used Azure OpenAI Service to develop Virtual OneStop (VOS) Sapphire AI, which rapidly generates high-quality job descriptions, customises interview questions based on applicants’ resumes and creates accurate case management notes on prospective candidates. The Georgia Department of Labor adopted VOS Sapphire AI in December 2024 and has since recorded a significant increase in employee productivity and a 65 per cent rise in the number of AI-generated records.
Applicants can also use VOS Sapphire AI to write resumes and cover letters tailored to specific job openings within minutes, potentially saving them hours of work. It has produced more than 2,300 new resumes and enhanced 3,100 others since it was rolled out in the State of Texas in April 2025. This has led to a 70 per cent upsurge in the number of resumes being submitted, as well as a 74 per cent increase in the quality of these documents across the two US states. “VOS Sapphire AI is making it easier than ever for labour departments to quickly find the right candidates for the available roles, which is getting people into work more quickly and having a positive socioeconomic impact,” says Arthur.
Other government-led organisations are capitalising on Azure OpenAI Service to provide self-service chatbots and virtual assistants that can provide 24/7 help to the public. The City of Burlington in Ontario, Canada, has collaborated with Microsoft partner MNP Digital to develop an assistant that expedites the building permit process to address high housing demand. The MyFiles portal has cut the average time to process permit approvals from 15 weeks to between five and seven weeks, and also enables citizens to easily track the status of their applications.
“AI chatbots can answer questions, provide information and guide the public through application forms to ensure they are completed correctly,” says Arthur. “The public no longer need to wait in long phone queues or make in-person appointments to complete tasks like renewing their driving licence or applying for financial aid, which is quicker and more convenient for them. Plus, employees only need to get involved in complex conversations or to review and approve applications, which reduces their workload significantly.”
Meanwhile, in Argentina, Buenos Aires City and Pi Data Strategy & Consulting have used Azure OpenAI Service to evolve its existing chatbot, Boti, which manages two million queries per month without human intervention. The upgraded Boti with ChatGPT chatbot personalises interactions in real time and can respond in both Spanish and English. To date, it has managed more than 17,000 queries autonomously, which has halved the operational workload for government employees.
Similarly, Montgomery County in Maryland, USA, worked with Microsoft partner Zammo.ai to introduce an AI agent built on Azure OpenAI Service that can conduct multi-turn conversations on more than 3,000 topics in 140 languages. Monty 2.0 also uses Montgomery County’s geographic information system to provide personalised responses based on the user’s address. Since it launched in late 2024, Monty 2.0 has facilitated more than 20,000 constituent conversations, achieving a 50 per cent customer satisfaction rate and reducing unanswered queries from between 35 and 45 per cent to just 10 to 15 per cent.
The public can use AI-powered chatbots like Buenos Aires City’s Boti with ChatGPT app to get immediate answers to their questions
“Immigrants or members of the public who don’t speak the government’s official language fluently are often excluded from accessing the services and support they need,” says Arthur. “However, chatbots powered by generative AI can translate both speech and text in real time, so everyone can communicate effectively in their preferred language. This eliminates communication barriers and increases service accessibility and equity.”
To reap the benefits of these AI tools, government organisations must carefully plan their adoption strategy. “They can start by defining the business outcomes they would like to achieve and building a dedicated team of internal stakeholders that can develop and implement a strategy for using AI to fulfil these goals,” advises Arthur. “Next, they must clean up their existing data and use grounding and retrieval augmented generation to ensure the AI tool can pull real-time data from multiple sources to deliver the most accurate and relevant responses.”
Crucially, organisations must set parameters to guarantee the AI tool can only access specific data that aligns to their governance policies to prevent security breaches, adds Arthur. “It’s essential to implement robust security measures to safeguard systems and data,” he says. “By putting Zero Trust principles at the core of their cybersecurity strategy, organisations can ensure everything is locked down by default, which stops employees from accidentally – or maliciously – creating vulnerabilities that will enable hackers to attack critical systems and steal data.”
In addition, organisations should establish clear guidelines to govern how employees interact with data and AI tools. Arthur recommends following the principles and best practices outlined in Microsoft’s AI governance framework, Responsible AI Standard and its Secure Future Initiative. “Microsoft is working with AI experts and policymakers around the world to help shape and share effective implementation practices to ensure our customers use AI ethically and compliantly,” he says. “We make all of this information available publicly – for example in our online Microsoft Trust Center – to help governmental organisations navigate the ever-changing regulatory landscape and protect their systems, employees and the public. All of Microsoft’s AI solutions have been designed following these standards, so they have built-in data security, governance and sovereignty capabilities to combat ever-evolving threats.”
Microsoft is also developing new cloud sovereignty solutions to alleviate organisations’ concerns about data security and privacy when introducing AI solutions. In June 2025, for example, it expanded the Microsoft Cloud for Sovereignty with the new Sovereign Public Cloud and Sovereign Private Cloud, which will be available to all European customers in all existing European data centre regions. This will ensure customer data stays in Europe, governed by European Law, with operations, access and encryption controlled by European personnel.
“The Sovereign Private Cloud is designed specifically for governments, critical industries and regulated sectors that need to meet the highest standards of data residency, operational autonomy and disconnected access,” says Arthur. “This is just one example of how Microsoft is working to increase digital sovereignty for governmental organisations and allay their fears about data protection so they can reap the benefits of AI.”
To further alleviate their concerns about the potential misuse of AI, governmental organisations can develop best practices to control how its employees and partners develop, implement and use such tools. Microsoft, for example, follows six key principles when developing its AI technology: inclusiveness, fairness, accountability, transparency, reliability and safety, and privacy and security.
“By following these core values, we can design, build and operate AI solutions that are trustworthy, secure, reliable and deliver accurate, non-biased results that uphold societal and ethical principles,” says Arthur. “Organisations should take a similar approach and also develop ethical decision-making frameworks to ensure they retain human oversight over their AI tools and comply with legal, ethical and regulatory requirements.”
Training employees how to use AI tools and follow responsible use principles is vital. “Governments can’t assume everyone will happily adopt AI solutions or instantly know how to use them,” says Arthur. “Instead, they must explain how AI will benefit individuals in their specific role and train them to use it properly, for example by teaching them how to write effective prompts so they get the result they want. This will encourage widespread adoption and ensure organisations maximise their AI investments. We always advocate for humans to be kept in the loop whenever AI is deployed too.”
The Public Investment Corporation (PIC), an asset management firm fully owned by the South African government that mainly serves public sector entities, took this approach when it deployed Copilot. The organisation initially rolled out Microsoft 365 Copilot to 50 users, who reported that, on average, they were able to rapidly analyse high volumes of documents to process and approve investment opportunities in six months rather than 12. To drive adoption, PIC built a ‘champions’ network’ and used its knowledge platform, Tech Lounge, to share Copilot insights and success stories from business users. Since then, PIC has experienced a 1,000 per cent surge in demand for Copilot licences.
“Microsoft offers various training courses and resources – such as the Public Sector Center of Expertise and Digital Skills – to help governments teach their employees how to optimise their AI investments,” says Arthur.
Microsoft’s 2025 Work Trend Index Annual Report indicates that 80 per cent of the global workforce currently lacks the time or energy to do their jobs, not least because they are interrupted by a meeting, email or notification every two minutes. To overcome this issue, 82 per cent of leaders expect to use digital labour to expand and support their workforce in the next 12 to 18 months, and 46 per cent of organisations are already using agents to fully automate workstreams or business processes. In addition, the report suggests there is a new type of organisation emerging – the Frontier Firm, which is built around intelligence on tap and human-agent teams. Currently, 71 per cent of workers at these firms say that their company is thriving.
“AI is rewriting the rules of business and driving a seismic shift in the way we work,” says Arthur. “Many people have used AI in their personal lives and they’re excited by what it can do, so they want to use it at work too. If governments don’t equip employees with AI tools, they will likely use commercially available AI tools that may unintentionally expose sensitive personally identifiable information – not out of malice, but just because they are trying to do their jobs better. It’s incumbent on them to find a way to introduce secure enterprise AI tools to support their employees to work more productively and free them up to focus on the main reason they joined the public sector – making a real difference to the lives of others.
“Digital transformation has been a buzzword for at least the last decade, but the growing number of successful AI implementations worldwide shows the technology is fundamentally altering how organisations operate. AI is the technology zeitgeist of our era. We’re not just talking about it in the technology and business worlds; we’re seeing it referenced on TV and in movies, talking about it with friends, and using it in our daily lives. Not only is AI boosting employee productivity and satisfaction, but it is also empowering governments to reimagine their processes, develop services that would previously have been impossible and improve the lives of their citizens.”
Partner perspectives
We asked selected Microsoft partners how they are using Microsoft technology to help government employees boost their productivity, collaborate effectively and deliver high-quality services to the public
“Smart cities thrive on a connected, informed citizenry,” says Gary Wong, global segment leader of power, utilities and infrastructure at AVEVA. “In response, infrastructure players are now linking up previously siloed domains to make better use of finite resources for everyone’s benefit. By exchanging electricity and water data, for example, both operators gain better situational awareness, enabling them to anticipate and address residents’ needs ahead of potential emergency events.”
“Our newest solution, Axon Assistant, is an AI voice companion built into Axon Body 4 to offer an eyes-up experience and a range of capabilities for officers in the field,” says Ran Mokady, general manager of Axon Assistant at Axon Enterprise. “For example, Real-Time Translation works with more than 50 languages, helping officers to communicate clearly during traffic stops and calls for service. Policy Chat on mobile, desktop and voice lets officers ask policy questions and get clear, verifiable answers anytime, anywhere. Powered by a range of Microsoft tools, Axon Assistant capabilities let officers stay focused, get quick answers to everyday questions – whether about agencies policies or general knowledge from the internet – and work more efficiently to keep communities safe.”
“Delivering world-leading GIS requires world-class geo-enabled cloud infrastructure and business productivity tools,” says Christian Carlson, senior director of global business development for state, local and provincial government at Esri. “Esri leverages both through our partnership with Microsoft. From Azure to Fabric, Microsoft technology helps to power Esri’s government solutions. These capabilities enable leaders worldwide to make better-informed decisions, improve resilience and public safety, and plan more prosperous and sustainable communities.”
“Our collaboration with Microsoft empowers government employees to use both SAS and Microsoft open APIs for seamless integration,” says Gordon Mullin, head of government research and development at SAS. “Microsoft Graph API, Power Automate and Azure Generative AI complement our own government solutions. Microsoft Teams, Outlook and SharePoint can be integrated into our intelligence solutions for accessibility, increased collaboration and efficiency across departments.”
Discover more from these partners and others, including Coretek and Simpson Associates, in the Summer 2025 issue of Technology Record. Don’t miss out – subscribe for free today and get future issues delivered straight to your inbox.