The following commentary first appeared last week in the monthly update published by the Financial Information System for California (FI$Cal), the state’s accounting, budget and business portal. It was edited only lightly for style.

At FI$Cal, our customers stand at the forefront of what we do. I’m excited to share our evolution at FI$Cal — putting artificial intelligence (AI) and robotic process automation (RPA) into our systems to bring more help to our customers.

The implementation of FI$Bot marked one of our first uses of AI. A key upgrade to the chatbot last year used AI to allow it to learn from its experience, so that FI$Bot improves answers to user queries as it goes forward. I highly encourage you all to try FI$Bot.

We are also looking at AI to make use of our historical ticket data in the FI$Cal Service Center. The goal is to train the FSC system to help users faster and at a much higher level, starting by processing emails from the inbox and routing them directly to staff who can respond to them.

We are just getting into the deployment of RPA in a number of important tasks. RPA automates many of the more mundane, repetitive tasks that workers perform. The software, or robots, respond to data and communicate with other systems to perform a wide variety of tasks. They do it with 100 percent accuracy and can run 24/7. The robots do the tedious work, saving the more interesting and creative jobs for people.

At FI$Cal, we have used RPA in a number of key areas. Not long ago, Covered California and the California Department of Forestry and Fire Protection (CalFire) needed to add thousands of vendors to their systems. This is a very time-consuming process that would have taken months under the traditional method of adding each vendor individually. We deployed RPA to make these requests in big batches. The robots made the request, and our vendor management group created a special queue where all the robot-generated requests went.

In the case of Covered California, we used two robots to add more than 10,500 vendors. A task that would have taken months took 18 days.

Another RPA use we developed helped with a new State Controller’s Office functionality release that changed how deposits were entered into the system. This forced a couple of departments to re-enter hundreds of deposits. Using RPA reduced the end-to-end process of reversing and remitting a deposit to 4 minutes from a minimum of 30 minutes before.

It’s easy to see how the proper deployment of RPA can save hundreds or even thousands of hours for departments. It cannot be used for everything, however, and we realize that not all departments even need it. We are sensitive to the needs of our departments and are always working to avoid any disruption to their processes.

This is one reason we have formed a user group within our Customer Impact Committee. If departments are building new tools around FI$Cal, we want to hear from them as soon as possible. We need to be sure we’re working together as RPA becomes more mainstream.

Finally, I want to let users know that we are aware that many of the final steps they are taking for year-end close statement generation are manual and outside the FI$Cal system. We are looking to use RPA to automate those steps. Departments spend a lot of time on this, and it won’t be long before they are pulling statements together to close this fiscal year. We think it’s feasible to use RPA for this, and we are analyzing it with a goal of implementing something in the fall.