Negotiating for Better Terms on AI Vendor Contracts with David Peppard

This is part of our executive insights series where Elion CEO Bobby Guelich speaks with healthcare leaders about their tech priorities and learnings. For more, become a member and sign up for our email list.

Name: David Peppard

Role: Chief Information Officer

Organization: US Acute Care Solutions

To start off, can you give a quick overview of your remit at USACS?

I wear a few hats. I’m the CIO, so all of technology rolls up to me—core infrastructure, cybersecurity, facilities, and much of the work we do around AI. I also run revenue cycle. That means every visit—about 11 million a year—is coded, billed, appealed, and collected under my team. A lot of my time is spent on strategy: where our products and services go next and how we support 420 locations of emergency medicine and inpatient clinicians.

On the tech side, what’s at the top of your priority list?

Security is always top of mind. After that, three buckets dominate:

1. Clinical AI. This includes ambient clinical documentation, clinical summaries tools, and solutions that help extract structured and unstructured data to help us improve patient safety and clinical quality. Essentially anything that lets our clinicians spend less time in the EHR and more time with patients.
   

2. AI in revenue cycle. We’re already using autonomous coding, but we need to get the automation rates up. We’re also looking at AI to fight denials, work second-level claims, and shorten A/R.
   

3. Data, analytics, Workday. We’ve got a pretty robust data warehousing team and analytics team and have a steady pipeline of Workday financial and human capital improvement projects.
   

Let’s dig into ambient documentation. What’s been your experience so far?

Twelve months ago we short-listed ten vendors, knowing that over time we’d likely have three or four partners. We chose Freed as our first partner, and their model has improved radically over the last year, but the landscape keeps shifting—Ambience, Sayvant, Cleo, and others look promising too.

We work in large health systems, so we’re expecting that in the coming years, Epic’s partner of choice will become the default. A lot of the 80-plus ambient startups you see today won’t survive the consolidation wave or will need to work with smaller physician practices.

You also mentioned AI for revenue cycle. How has that been going?

It’s still relatively early for some of the “bots fighting bots” solutions. The tech looks interesting, but it’s expensive. Vendors price it on labor savings, but compared to the underlying platforms we’d use if we built it, they’re marking it up 30-40x. We can’t wait, so we either have to negotiate better or build it ourselves until pricing comes down.

I think we are seeing building become more viable for folks with how much easier software development is getting. What’s your current stance?

I would say actually over time we’re building less internally and trying to be more disciplined about what we build. I want to buy certain vertical products, but we’re actually building more of the core RCM stack ourselves using a rapid-application development process. 

Any lessons to share based on your recent AI vendor evaluation experiences?

First, every CIO needs a seat at the contract table—if you’re not vetting the spend and the legal terms yourself, you’ll end up hand-cuffed to volume commitments and exclusivity clauses you never intended to agree to. AI is moving too fast for long-term lock-ins. Right now many vendors’ investors push them to show big, sticky recurring-revenue numbers, so they’ll ask for your whole book of business before they’ve proven any value. Know your worth, be patient, and tie payment to real performance.