HHS combines AI, automation, blockchain for acquisition project

Much of the chatter around IT modernization this year has centered on the potential benefits of emerging technologies like artificial intelligence, automation or blockchain could bring to federal agencies.

But the Department of Health and Human Services hopes to demonstrate how all three together could streamline its buying power by November.

Jose Arrieta, HHS’s associate deputy assistant secretary for acquisition, said Thursday that the agency is projecting that its ReImagine HHS initiative could obtain a full authority to operate before Thanksgiving, providing a chance to test the system with live data and potentially transform the acquisition process.

“We think the model that we’ve built is the operational model for administrative management functions across the department,” Arrieta said on a panel at the Red Hat OPEN FIRST Road Tour. “We are currently going to start to implement it in other areas. So this just isn’t a model for acquisition.”

ReImagine HHS aims to make acquisition more efficient by leveraging AI, automation and blockchain technologies in collaboration to craft a series of microservices that use procurement data to find cost savings, design better contracts and slash buying time.

Here’s how it works: Machine learning algorithms cleanse the data of five separate contract-writing systems and put it in a structured data taxonomy to track through blockchain technology. The blockchain automatically updates transaction information in a secure, compartmentalized cloud infrastructure.

From there, the project uses microservices to prepopulate acquisition and market research documents from the structured data layer. The microservices also can analyze historical prices-paid data from the agency’s product categories to find better price comparisons for contract negotiations. The system is coupled with a human-centered-designed interface HHS tested with employees to make sure it was easy to navigate.

The target is to trim the average acquisition’s lifecycle down from 200 days to just 90 and provide enterprisewide visibility into a spending process that operates across the federated agency.

Arrieta told FedScoop that the project has been using a static data set to test the potential solutions it can provide acquisition officials. If the project secures a full ATO, it can then take live data from the contract-writing systems to see if it can scale up its services.

“That’s really exciting for us,” he said, crediting Oki Mek, HHS’s emerging technology strategist, for the progress the project has made after starting as a $150,000 prototype.

As for the project’s potential to impact other management operations, Arrieta said the strategy combines the benefits of the three popular emerging technologies and can be deployed across multiple use cases.

“I could be wrong, but it gives us the ability to take the best of AI, the best of data management and an immutable record, blockchain, and apply them,” he said. “What is administration and management? It’s making an accurate dataset and using that dataset for reporting, for analytics, for processing.”

For agencies looking to innovate without a lot of capital, Arrieta said the relatively small cost of the project’s prototype, which the team designed alongside a business plan, allowed it to generate buy-in and manage risk while remaining flexible.

“Start small,” he said. “If it doesn’t work and you spend $50,000 to $100,000, kill it. If it works, maybe scale it a little bit and test it again. But my number one piece of advice is start small. Rebuilding an entire system costs too much money.”