Are you staying ahead of the curve?
COVID-19 has accelerated AI adoption in the Public Sector. In March 2020, processes, people and services went online, forcing governments to adapt to the new normal and become more digital. According to a report by EY (1), in a few short months, governments have:
Digitized on a scale that might have seemed impossible before
Learnt to manage a remote workforce
Worked with the private sector to close skills gaps and develop innovative solutions
Used AI as a key weapon in the fight against the virus – from educating the public and screening patients to tracking and tracing contacts.
This article will focus on challenges and opportunities for governments to use AI to supercharge their processes, policies, and communications to bring value to their employees and citizens. AI can deliver benefits beyond process optimisation in the long run, helping to tackle long-term global challenges such as climate change.
Is the Public Sector ahead of the curve?
While AI is in governments’ minds, how real is this progress? Certainly, there have been impressive advancements in the past few years. For example, in 2018, the European Parliament and Council published a directive that citizens should have access to information, procedures and assistance through a "Single Digital Gateway". The AI solutions that private companies have been developing in recent years to better address customer needs have set an example for governments to make their information accessible through virtual assistants.
Government services in the UK have started to accept techniques like anomaly detection and computer vision to keep track of the quality of the country’s energy infrastructure (2).
However, despite significant benefits, attaining them is not an easy task. In the opinion of the OECD (3), government use of AI trails that of the private sector; the field is complex and has a steep learning curve.
We will zoom in on two key challenges that form a barrier to AI adoption in the Public Sector, offering guidance on how to overcome them. Finally, we will explore three cases where governments have successfully implemented AI to inspire the reader in undertaking their own AI projects.
What are two critical challenges to AI adoption in the Public Sector?
An all-round AI mindset
While Public Sector organizations see digital as a top priority, many are yet to embark on an AI journey. According to a WEF report, in some cases, public sector officials may lack the appropriate knowledge and in-house expertise to create an AI strategy or choose the best AI tools (3).
How can you solve this?
Since Superlinear was founded in 2018, we have worked in a capacity of “trusted partner” with the AI teams of many Public Sector organizations, including VDAB, FIT and FOD BOSA. Collaborating with partners gives organizations a fresh outside perspective on issues and opportunities. An AI partner is also able to leverage their expertise across clients and industries. While this has been the preferred route for our clients, there are also two other routes the Public Sector can consider when starting an AI project:
You outsource it. You work with expert partners that help you accelerate adoption by bringing in deep expertise and experience. This is often the most cost-efficient option in the short term.
You do it yourself. You internalize the project and build up a capability which will be useful for future projects. You also have greater control over your setup.
If you are interested in deepening your knowledge on this topic, read our article Five keys to set your AI project up for success.
A perceived lack of data
The second challenge for the Public Sector is widespread across the board: lack of data. While in many cases, the data needed for AI solutions to be developed and deployed is often neither accessible nor discoverable (4). More often than not, it is a perceived lack of data blocking AI advancement. Uncertainty about ethical considerations adds further layers of complexity.
How can you solve this?
In the last decade, the field transitioned from academic to practical applications of AI. One of this transition’s outcomes is the increased capability to tackle AI use cases with less, or sometimes no data. For example, there have been significant successes in the field of unsupervised learning and transfer learning. A model can be trained based on other datasets, often open-source or even simulated data, and then fine-tuned on smaller amounts (10s to 100s of data points) of data specific for your use case.
These successes lead to a lower data threshold to start investigating your AI use case. If you’re unsure where to start, an AI agency such as Superlinear can help you find creative solutions to circumvent data challenges and use open data. What’s more, thinking about AI use cases is often the first step to decide which data to start capturing and how.
In summary, don’t let the (illusion of a) lack of data block you from being innovative, but start working towards solving your challenge.
When it comes to data and ethical considerations, we have an in-depth article on fairness and bias in AI and how organizations can approach it at this link.
Due to these challenges, Public Sector organizations tend to delay buying decisions or reduce their scope when it comes to AI projects (3). In the next section, we explore three examples of government organizations that overcame these challenges, bringing value to their employees and their citizens with AI.
Three AI use cases for the Public Sector: inspiration to be ahead of the curve
These cases are collaborations between governments’ AI teams and Superlinear over the past year.
1) How AI-assisted Question Answering helps Flanders Investment and Trade attract foreign investors
Context: Flanders Investment and Trade (FIT) is a Flemish public agency that promotes Flanders’ international enterprise by supporting Flemish companies’ international trade activities and attracting foreign investors to Flanders. Every week, Flemish and foreign companies contact FIT with many questions on how to expand their business in or from Flanders. FIT’s answers and advice on these questions are key to the agency’s added value for its clients.
Goal of the project: Reshaping FIT’s question-answering process through artificial intelligence, allowing FIT’s advisors to give clients accurate, actionable and timely answers to their questions. The project aimed to help advisors free up time to conduct more in-depth research, improving the process’s overall quality.
Solution: An AI solution that automates the question-answering process at FIT. Superlinear’s model allows FIT and its advisors to save time and resources while improving user satisfaction.
Impact: We empower FIT’s advisors who can, in turn, help Flemish companies and foreign investors. AI implementation further positions the agency – and Flanders – as a high-profile market for businesses all around the world, as well as an innovative, data-driven public agency. Superlinear’s solution reduces the time spent on cases by 27%, and FIT’s advisors can help 36% more clients.
Read the full case study here.
2) Superlinear and VDAB reinvent orientation test with Artificial Intelligence.
Context: VDAB, the Flemish public employment service, worked together with Superlinear to update and streamline its orientation test using machine learning.
Goal of the Project: Reduce the time it takes job seekers to complete the orientation test, improve the user experience, and simplify the analysis of the responses to result in better career recommendations.
Solution: Orient 2.0, a digital and AI-supported tool that creates a personalized orientation test and uses machine learning to predict professional interests.
Impact: Time saving of almost 80% for job seekers. 10.000 users had already completed the orientation test just two weeks after launch.
Read the full case study here.
3) FOD BOSA: Improve the accessibility of governmental websites with AI
Context: The European Directive on the accessibility of government websites and mobile applications has been in force since December 2016. In collaboration with other partners, FOD BOSA developed an Accessibility Checker to meet the European Directive, which was “rule-based”.
Goal of the project: with AI, expand and improve the existing FOD BOSA Accessibility Checker, a tool which scans websites for accessibility violations and reports these violations back to the user.
Solution: an AI “superpowered” Accessibility Checker.
Impact: coming February 2021.
Conclusion
Despite challenges, AI adoption in governments is going strong. The right partner can help governments bring use cases to life or even discover new ideas. At Superlinear, we call the latter “Fast Discovery”. In this workshop, we co-identify the most impactful areas where AI can create real value for your end-users and your business. We sit down with decision-makers and influencers, domain experts, and data experts that have their knees deep in the business to map out their key ambitions and challenges, near and long term. By the end of the workshop, you will be able to visualize what a working version of your AI solution delivered after just one sprint could look like. To learn more about the workshop and book an initial consultation, visit this page, or book a no-obligation 30-minute call with our Client Partner Kristien Verreydt.
Sources
