How government agencies can harness generative AI
A message from ServiceNow
Generative AI has the potential to solve some of government agencies' biggest digital transformation challenges — but many don't know where to start.
ServiceNow held its annual customer and partner event, Knowledge 2024, in May in Las Vegas with over 20,000 customers and partners in attendance.
- At Knowledge, harnessing generative AI for the public sector and seeing real benefit from its capabilities was the topic du jour.
Now, Justin Herman, the Vice President of Mission Success, Global Public Sector at ServiceNow and the former Artificial Intelligence for Citizen Services lead for the U.S. federal government, is continuing the conversation as he answers some of the most pressing questions around what's possible in government with generative AI.
1. First things first: What do you feel are the implications for generative AI for government agencies?
Herman: Over the past decade, we've seen government agencies embark on a campaign to harness more data to support their operations.
- Then, about three years ago, these agencies recognized a problem. It wasn't that they didn't have enough data — it's that they had too much data.
This is where the power of AI started to play a much bigger role. Without AI, there would absolutely be no way for these agencies to sort through vast volumes of data to make it more actionable.
- Federal government agencies can now look for predictive actions and serve their constituents and stakeholders with this data in brand new ways.
- Large language models trained on customer-specific domain data present a tremendous opportunity to help agencies respond rapidly to their constituents. Imagine if government procurement departments could anticipate a requisition order before the employee asks? Or an IRS that can use already submitted data history to help complete tax forms?
At ServiceNow, we're seeing many federal agencies adopt AI in these ways.
2. How it's done: What should government technology managers and CIOs consider before implementing a generative AI strategy?
Herman: There are many areas where governments can be more proactive with how they deliver services to citizens.
The first step is figuring out the use cases that will give them the biggest bang for their buck.
- This is explicitly mentioned in the White House's January 2024 Executive Order on AI where agencies were tasked with coming up with use cases for generative AI.
When not applied to a specific use case or pain point for an agency, it generally will not get enough traction or funding to build and scale.
3. Some examples: How have agencies benefitted from AI?
Herman: I previously served as the U.S. Federal lead of Artificial Intelligence for Citizen Services at the U.S. General Services Administration's Technology Transformation Service, where I worked with agencies where huge volumes of data were collected.
- From health care modernization and emergency management, to defense and public benefits, across departments they aimed to take that data and enable teams at the front lines of public service to make better decisions.
- With the volume of data, it was clear this was not something human beings could effectively analyze and rapidly make decisions. AI became the critical tool that helped filter the data and help managers make decisions.
Other examples of how government agencies can put AI to use include:
🔒 Cybersecurity: AI can help security specialists identify network vulnerabilities before they are exploited, defending IT systems and networks faster and more comprehensibly.
💰 Fraud, waste and abuse: Finding repetitive or non-productive behavior patterns can help optimize taxpayer dollars and identify funding gaps where there shouldn't be.
🪖Access: AI can help bridge the gap intuitively between the customer, in this case, citizens or government employees, and the complex systems designed to serve them.
4. Okay, but: What challenges should government agencies be aware of as they look to implement a generative AI strategy?
Herman: It goes back to the basics. It's important to ensure high-quality data is available for AI to produce high-quality results.
To ensure high-quality data, look at the processes you have in place to generate data and start to find opportunities to clean up the data as part of an existing workflow.
- This is where ServiceNow has invested in our platform and our capability. Once you start to digitize your workflows, you have good, clean data. Then, you can actually start to leverage that data for decision-making and for AI.
- We should also remember that AI is a tool, a powerful tool, but it still needs the human element to work. AI can help with decision-making, but humans need to be in the mix to act on the right path forward.
At Knowledge 2024, ServiceNow also announced that our generative AI capabilities will be integrated into the ServiceNow Government Cloud Community (GCC) to enable government agencies to leverage key use cases like code generation, case summarization, incident summarization, and GenAI‑powered search to deliver services quickly and effectively.
- GCC is one of ServiceNow's dedicated clouds designed to meet the full spectrum of regulatory, compliance and security and opens up our generative AI capabilities to more than 700 public sector customers. Through this new development, governments can harness GenAI to improve their mission customer experiences and best serve the public needs.
5. The impact: How can AI benefit citizens, constituents and businesses?
Herman: The government workforce will become a lot more productive with AI.
For example, right now, responding to citizen inquiries and issues is a very manual, painstaking process that touches multiple systems.
📈 Using generative AI, our internal data shows workforce productivity increases dramatically — an approximately 40% to 50% improvement.
Citizens can also now get many of their requests fulfilled through self-service portals that leverage generative AI. That means they don't even need to call the contact center or the help desk to get an issue resolved.
6. Looking ahead: What will the role of generative AI in government agencies look like in five years?
Herman: I fully expect that AI and generative AI will be a part of everything the government does.
- The speed at which we are seeing the adoption of both technologies today in the federal government far exceeds anything else we've seen in the past.
Government agencies will have to reimagine their processes. They'll need to consider their workforce and ensure they have the right people, suppliers and contractors — with the right skill sets — in place.
We'll see rapid adoption if AI isn't treated as an afterthought or a nice-to-have but instead as core to every part of the government's operations.