The U.S. government’s journey in artificial intelligence (AI), much like its journey to cloud services, is bound to be a cautious one, according to Accenture Federal Services executives who spoke at a roundtable in D.C. on Thursday.

As reported by Washington Business Journal, Accenture Federal’s chief technology officer Dominic Delmolino told reporters that federal agencies’ approach to artificial intelligence has “in many ways mimicked its move into cloud computing” in that the least complex workloads will move over first.

The private sector of course is already using artificial intelligence and machine learning to solve problems like how many servers are needed to keep AWS cloud running.

So far the federal government has used AI to build virtual assistants into public-facing agency websites that help guide users to answers to common questions. Examples of those include Emma, a virtual assistant on the U.S. Citizenship and Immigration Services website or Sgt. Star on the U.S. Army’s website, according to the report.

If used correctly, AI should alleviate some of the burden of day-to-day administrative work from government employees, Delmolino said, including how AI can be used to clear backlogs of decisions that agencies have to make.

Accenture Federal said its own strategy consulting contracts with the federal government will help agencies come up with the right strategy for AI.  

For now, federal government employees are still grappling with cloud computing; according to a recent study by Deloitte, 40 percent of federal government employees don’t know if cloud computing has had a positive or a negative impact on their department or agency.