Best Practices

Activity Metrics Won't Tell You If Your AI Is Working

Activity Metrics Won't Tell You If Your AI Is Working

Rick Scorzetti, Chief Commercial Officer

Rick Scorzetti, Chief Commercial Officer

Healthcare AI should be measured by revenue impact, not operational activity.

The Wrong Scoreboard

There's a default way most healthcare organizations evaluate their AI investments right now. They look at operational KPIs. Percentage of calls answered. Average handle time. Response rate. Interactions completed.
Those metrics tell you the platform is running. They don't tell you whether it's creating value.
I think about this a lot, because I've spent my career watching healthcare organizations invest in technology that looks productive on a dashboard but never connects to the financial outcome that actually matters: did more patients get on the schedule, and did the practice capture more revenue because of it?
The answer is usually buried. Or it's never asked in the first place.
When you evaluate AI through operational KPIs, you end up optimizing for activity. More calls handled. Faster response times. Higher interaction counts. The dashboard turns green. Everyone feels good. But the schedule still has gaps, referrals still sit unprocessed, and the revenue that was leaking before the platform was deployed is still leaking.
The organizations getting the most from their AI investments have shifted to a different lens entirely. They're measuring value creation. And the difference is significant.


What Value Creation Actually Means

Value creation in patient access comes down to two things: recovering revenue you're currently losing, and generating revenue you're not currently capturing.
Revenue recovery means filling the gaps that already exist in your operations. Cancellations that get rebooked. No-shows that get re-engaged. Open slots that get filled before the provider's day starts with empty chairs. This is money already in the pipeline. The patients were scheduled. The demand existed. It just leaked.
Revenue generation means activating demand that isn't on the schedule at all. Lapsed patients who haven't been seen in 18 months. Care gaps that have no future appointment attached. Referred patients who were sent to you but never contacted. This is incremental revenue that most practices don't even realize they're missing because it never shows up as a line item. You can't see what you're not measuring.
When you combine recovery and generation, the impact flows through at high incremental margins. The providers are already on the schedule. The locations are already open. The infrastructure is already in place. You're not acquiring new patients from scratch. You're converting demand that already existed within your own four walls.
That's a fundamentally different value proposition than "we answer more of your calls."


What This Looks Like in Practice

I want to ground this in real numbers, because value creation only matters if the results are measurable and sustained.
We now partner with 6 of the top 10 dermatology practices in the US, supporting over 2,000 providers across 800+ locations. At this scale, the data is meaningful.
At one enterprise practice, our outbound AI campaigns are generating incremental revenue lift measured in tens of millions annually compared to the incumbent platforms they were running before. The outreach resolution rates tell the story: 14% blended resolution across all campaign types, which is 2.4x to 7x what the legacy tools were delivering. When you're reaching over a million patients a year, that multiplier changes the economics of the entire operation.
On the referral side, one national practice went from converting roughly half of inbound referrals to over 85% using fully autonomous processing. Every point of improvement represents patients already referred, already in the system, who would have been lost without proactive outreach.
More than three-quarters of the total revenue impact across our largest deployment is pure recovery. Not new patient acquisition. Captured demand. Revenue that would have walked out the door.
And the cost side holds up too. 60% reduction in call center operating costs. Patient acquisition cost cut by more than half. 360% average return on the investment within six months. Over $1M in operational savings within the first 90 days at one enterprise deployment.
These results are why we price on performance. When your outcomes are this measurable, tying your revenue to your client's results isn't a risk. It's accountability.

The Shift That Matters

The technology in this space is converging. The platforms are getting better every quarter. The demos are all impressive. Operational KPIs will keep improving across the board regardless of which vendor you choose.
But value creation won't converge. The ability to recover tens of millions in leaked revenue, to close referral gaps that have been open for years, to generate EBITDA impact from demand that was already sitting inside the practice's own patient base... that requires a depth of integration, operational expertise, and accountability that most platforms aren't built for.
If your AI investment is measured by percentage of calls answered, you'll optimize for a green dashboard. If it's measured by revenue recovered and generated, you'll optimize for the P&L.
The practices seeing the strongest results right now have made that shift. They stopped asking "is the platform performing?" and started asking "is the platform creating value?" The answer to the first question is almost always yes. The answer to the second one is where the real separation happens.

Rick Scorzetti is the Chief Commercial Officer at Parakeet Health, an AI-powered patient access platform for large specialty practices. He has 25 years of experience in healthcare technology sales leadership, including roles at Epocrates, WebMD/Medscape, and Jumo Health. Connect with him on LinkedIn.

Crafted in San Francisco 🌉

© 2026 Parakeet Health, Inc.

Crafted in San Francisco 🌉

© 2026 Parakeet Health, Inc.

Crafted in San Francisco 🌉

© 2026 Parakeet Health, Inc.