Government agencies have always lagged behind the private sector when it comes to tech adoption. But artificial intelligence could help close the gap.
Reality check: Nearly two‑thirds of U.S. businesses implemented some form of AI in 2017, according to a survey by the National Business Research Institute. Yet just 20% of government agencies are even in the planning stages of AI adoption.
While agencies are showing more interest in AI and automation, “I wouldn’t say it’s pervasive,” says Paul Seckar, a principal in the public sector practice at consultancy Grant Thornton.
Citizens stand to gain from this nascent trend. In the United States, the average wait time for a Medicare appeal decision is about two years. Intelligent chatbots and robotic process automation could cut that time and take much of the pain out of tedious tasks like applying for benefits, renewing a driver’s license or correcting a tax return.
Studies suggest that’s not just blue‑sky thinking. Within five to seven years, U.S. government agencies could save as much as 1.2 billion hours per year and $41.1 billion in annual costs, according to Deloitte.
While they lag behind businesses in AI adoption, public agencies do have one advantage compared to many companies: Most sit on large data reservoirs that AI and automation apps need to thrive. That’s a significant factor, even though government data is often not structured in the consistent formats that automated systems require.
The war on paperwork
Facial recognition technology, primarily for security applications, is attracting much of the investment in government AI today. The Department of Defense, the Transportation Security Administration and Immigration and Customs Enforcement are all exploring ways to use facial recognition and other AI tools for security applications.
The broader opportunity is improving government services by using AI not just to cut red tape but to kill it off for good. One reason so many government processes turn into painful citizen experiences is manual data entry and filing. Over half of all state and federal government workers can’t get their work done in a typical week because of excessive paperwork burdens, according to a survey by Governing, a magazine that covers public policy and management issues.
AI‑powered document processing could save government agencies time and money, says Rob Buhrman, another public‑sector principal at Grant Thornton who is experimenting with a machine learning prototype to process citizen applications for government services.
Chatbots can run call centers
Using AI in government call centers is another compelling use case awaiting investment, says Buhrman. The Internal Revenue Service, for example, manages 116 million calls from taxpayers per year, at an approximate cost of $41 per call. AI chatbots, Buhrman says, can handle responses to many of these questions for as little as little as 40 cents per query.
Similarly, AI voice assistants in DMV and post offices could answer questions and gather information from callers before connecting them with a service rep. “Hold times could be slashed and people would be quickly directed to the department best suited to help them,” says Nasser Aftab, director of business development at Phonexa, a call center technology firm.
The General Services Administration is currently piloting an initiative to help government agencies answer simple citizen queries by phone, web, or through voice platforms such as Alexa.
Digital processes > robotic process automation
Government agencies are also exploring the use of robotic process automation (RPA) and similar tools to overhaul back office processes, says Seckar. Tasks such as invoicing, claims processing, and purchasing can all be automated, freeing employees to work on higher‑level tasks.
Intelligent software bots can open email and attachments, fill in forms, connect to system APIs, read and write to databases, move files and folders, and extract structured data from documents.
Public sector healthcare is a good example of a government service that could be improved using intelligent automation. For example, patient outcomes at VA hospitals could benefit from predictive analysis, AI‑assisted diagnosis, and health monitoring, says Jason Odden, director at tech consulting firm Cask.
“The opportunity for patient experience improvement is significant, from getting patients in the door, with an appointment with the right person the first time, all the way through treatment to a successful outcome,” Odden adds.
Most of those examples are still speculative, of course. But the rapid rise of chatbots and AI tools like RPA suggests that government agencies will face intense pressure to move from planning to implementation in the near future.