GSA sets AI pilot as test for procurement process

GSA has released a solicitation for an artificial intelligence pilot to test a new procurement process for commercial solutions.

The General Services Administration has released a solicitation for an artificial intelligence pilot as part of its commercial solutions opening program to streamline how it procures innovative and commercial solutions.

This CSO process falls outside of the Federal Acquisition Regulation and was authorized as part of the 2017 National Defense Authorization Act.

The solicitation this week is for the use of AI for regulatory streamlining.

GSA’s FEDSIM office is managing the program. The pilots are intended to help GSA learn how to streamline the acquisition process, fast-track vendor selection timelines, simplify contract administration and find ways to let vendors retain their core intellectual property.

Proposals are due April 20 and the AI pilot will not be worth more than $10 million, according to the solicitation. The pilot will likely last a year.

FEDSIM is working with GSA’s Technology Transformation Service and the Centers for Medicare and Medicaid Services on the pilot. CMS is the sponsoring customer.

They are looking for a software-as-a-service solution that CMS regulatory staff can use to modernize workflows as they try to meet several executive orders that call for streamlining regulations.

The current regulatory process is “burdensome,” according to the solicitation because it is largely analytic and the applications the workforce uses are antiquated.

Those applications do not support workflow approval and routing, identify outdated regulations and duplicative language. They also do not help with cost estimates of regulatory changes.

Goals of the pilot include to increase collaboration, document history and track regulatory changes. Some of the specific AI features CMS is looking for include predictive analytics, natural language processing, and clustering and grouping.

The pilot should accommodate 20 users and include 10 terabytes of data capacity.

In their submissions, companies need to explain how their solution pushes the state of the art and need to submit their commercial price list.

Evaluation criteria FEDSIM and CMS are using includes responsiveness, transition approach, capabilities and related experience, and technical merit.

After evaluating submissions, FEDSIM and CMS may ask for one or more companies to submit a pull proposal.