Insight by KPMG

Low-code platforms, APIs democratizing intelligent automation

Kirke Everson, a principal and government intelligent automation leader at KPMG, said a lot of agencies are still in the early stages of applying intelligent au...

Shape

Intelligent Automation Trends

Many of these automation platforms are starting to integrate other more advanced capabilities. For example, some of the low-code platforms, many of whom are already in production at federal agencies are starting to incorporate robotics process automation (RPA) inherently in their platforms. The other things that the RPA vendors are doing is they're starting to enable application programming interfaces (APIs) to more advanced artificial intelligence and machine learning solutions within their platform. So it makes it much simpler for a government agency who's implementing one of these platforms to tap into what I'll call democratized AI.

Shape

Low-Code Platforms and Intelligent Automation

These [low-code] platforms allow you to stand up new capability in a matter of, and I will dare to say days versus months. One example was we worked with a large federal agency who had agency employees all over the world, and when the pandemic first hit, there was a big concern about the safety of their employees overseas. We actually helped within a matter of days stand up a capability within a low-code platform to start to track where those employees were the process of getting those employees back to the country and making sure that they were safe.

A recent report from the Robotics Process Automation Community of Practice in the government highlights the impact of this technology over the last few years.

In 2020 alone, the report says RPA program maturity increased significantly with the number of automations deployed across the government increasing by 110% and the number of annualized hours of capacity created increasing by 195%. Last year agencies deployed 460 automations, which is expected to save more than 848,000 hours.

The paper says this growth demonstrates that programs have matured and increased their functional capacity, which meant automation tools became more impactful and therefore increased the demand for these software solutions.

Overwhelmingly, the CFO office is using automation to save time, some 49% of all implementations came from that group. But acquisition, administrative and IT were all in the double digits, showing how success travels.

Kirke Everson, a principal and government intelligent automation leader at KPMG, said a lot of agencies are still in the early stages of applying intelligent automation to their business processes.

“Some of the trends that we’re seeing are agencies are less inclined to just to do, one or two proofs of concept to prove out the technology. Now that they know it works, a lot of agencies are looking to others for lessons learned and implementing RPA programs that are a little bit more robust. And by robust, I mean, more enterprisewide,” Everson said on the Modern Government: Emerging Trends in Intelligent Automation in a Time of Rapid Change show sponsored by KPMG. “I think agencies are looking beyond RPA as well. The whole idea of hyper automation is starting to come into the vernacular of a lot of agencies. What I mean by hyper automation is RPA is definitely a stepping stone to artificial intelligence (AI).”

These advanced capabilities using AI and machine learning can only happen if agencies create a structure to manage the processes and data.

“Many of these automation platforms are starting to integrate other more advanced capabilities. For example, some of the low-code platforms, many of whom are already in production at federal agencies are starting to incorporate robotics process automation (RPA) inherently in their platforms,” Everson said. “The other things that the RPA vendors are doing is they’re starting to enable application programming interfaces (APIs) to more advanced artificial intelligence and machine learning solutions within their platform. So it makes it much simpler for a government agency who’s implementing one of these platforms to tap into what I’ll call democratized AI.”

Everson said an important step to democratizing AI is to make sure employees understand the technology and processes, but they don’t have to be experts.

“With these APIs, I can pull in a very quick machine learning algorithm just based upon what’s already been pre-determined from the vendor and allow, for example, natural language processing of a contract. I can pull in a natural language processing algorithm to read a document and extract certain things from that document, without leaving the low code or the platform,” he said. “These [low-code] platforms allow you to stand up new capability in a matter of, and I will dare to say days versus months. One example was we worked with a large federal agency who had agency employees all over the world, and when the pandemic first hit, there was a big concern about the safety of their employees overseas. We actually helped within a matter of days stand up a capability within a low-code platform to start to track where those employees were the process of getting those employees back to the country and making sure that they were safe.”

Everson added the 2020 memo from the Office of Management and Budget saying bots are non-person entities when it comes to identity management, and agencies can continue to take advantage of intelligent automation as they move more systems and data to the cloud.

“A lot of agencies recognize that developing these capabilities from scratch isn’t the most efficient way to do it. A lot of the large cloud providers are allowing agencies to tap into these capabilities as part of their infrastructure solutions. So if you want to pull in a machine learning algorithm that’s going to do something for you, you can actually pull that directly from, for lack of a better term, an API store, it’s very much been kind of parsed out,” he said. “For certain capabilities, you can use those capabilities by the drink, get the license cost every time you hit that API pay the fee. And it’s a very minor fee. What’s happening is all these things are starting to converge, where you’ve got the infrastructure, the cloud providers, allowing you access to some of these more advanced capabilities through APIs. You’re also having the low code vendors implementing some of these API’s to access some of those capabilities. Then you’ve got also the RPA providers implementing low code and an API into their platform, so there’s this convergence among the software-as-a-service, infrastructure-as-a-service and platform-as-a-service to basically give the customer choices. So there’s really no need to develop these standalone machine learning algorithms unless it’s for a very specific purpose, that may have a mission need.”

Copyright © 2024 Federal News Network. All rights reserved. This website is not intended for users located within the European Economic Area.

Related Stories