Skip to Content

Perform AI for Public Sector: Public goes AI!

Anne-Laure Thieullent
March 4, 2020

By Pierre-Adrien Hanania and Anne-Laure Thieullent

The public sector is ready to embrace AI!

At the crossroads between citizenship, political action, and common values, artificial intelligence shows great promise for the public sector, and ultimately, citizens.

Whether it be detecting tax evasion, preventing crime, or accelerating the attribution of social benefits, artificial intelligence is a powerful companion for the spectrum of public services. What’s at stake, though, and how can the use of AI lead to better governmental action towards the citizen and society.

Governments across the globe have been emphasizing one key aspect of the technology – that there is a “democratic requirement”[1] to serve all within the AI journey; that it should act “for humanity,”[2] “for all,”[3] and that it should constitute an approach that “benefits people and society as a whole.”[4]

Enhancing the full potential of data for citizen services

To fulfill the potential of augmented government, it is essential to first master data, intelligently automating administrative processes and gaining insights that provide a better overview and enable better decisions for the benefit of the citizen.

At the same time, to embrace AI, it is essential to keep up with new societal, industrial, and judiciary standards relying on complex sets of data.

We see four main fields of the public sector that AI will empower:

For all of these use-case categories, there is one key aspect of our vision of AI, linked to the government’s strategies: the technology shall augment and assist humans, but not replace them. The power to make decisions remains truly embedded in human governance, in a world where AI provides the best insights possible.

The work has already begun:

While several nations have already published their AI strategies, specific elements of the public sector have also started work on concrete missions, together with Capgemini:

Guiding the citizen regarding public services: In many countries, public services struggle to answer the influx of questions regarding social benefits in a short time period. The large number of different services, in addition to the lack of resources and the complexity of administrative processes, make it difficult to answer all enquiries in real time. For a social insurance organization, Capgemini and Sogeti implemented a fully functioning AI-based chatbot that answers all types of insurance-related questions. It channels human characteristics such as kindness, honesty, and empathy, it is available 24/7, and it builds on AI to keep learning and thereby improve its case management. After just one month, it was able to help 75% of the time, mastering the most trivial questions that previously created a bottleneck within public administrations.

Detecting fraud: Fraud causes billions of money losses every year that governments struggle to retrieve. For the tax authority of a European country, Capgemini implemented an AI solution to combat VAT carousel fraud. While this complex type of VAT fraud made it difficult to apply siloed applications and manual, individual inspections, advanced analytics provided decisive insights within networks of people and data. Hence, links between factors were identified, making it possible to detect patterns, resulting in a return on investment of a factor of 20.

Optimizing the use of CT scans: The Norwegian Akershus University Hospital, which provides healthcare to approximately 500,000 inhabitants around Oslo, wanted to improve the use of Computer-assisted tomography (CT) examinations in emergencies.[5] Together with Capgemini, the hospital developed an approach enabling more than 5,000 anonymous CT examinations to constitute the unstructured data being analyzed. Thanks to a combination of machine learning and natural language processing, the project attained outcomes impossible to achieve with a manual analysis of a limited number of randomly selected patients. The results of this project are now used for quality assurance and optimization of indications and procedures for CT scanning in critical situations, helping the hospital serve its patients better.

Toward a digital social contract:

Public institutions and governments will be at the frontline of this technological journey towards improvement while being guardians of its ethical and lawful use.

Furthermore, the following aspects of the AI-powered transformation journey towards mastering data must be considered by public sector organizations:

  • Working on sustainable data governance, bringing an enabled data pool to life
  • Relying on partners where it makes sense, be it established software partners or startups
  • Bringing together the power of academia, universities, and research institutions
  • Proactively addressing the question of trust and ethics, nurturing change management and data-centric culture within organizations.

By addressing the problems most citizens face at some point, with the right partners and a continuous effort towards the ethical and scalable use of technology, AI can transform the way public services are developed and delivered.

Read more:

[1] See Nationale KI Strategie (German AI Strategy), on:, p. 4.
[2] See French national Strategy, on:
[3] See Indian national strategy, on:
[4] See communication of the EU on Artificial Intelligence, on:

This article has been co-authored by the following experts.
Please get in touch with Pierre-Adrien Hanania to discuss further the positive impact AI can make.