By Michael Palantoni, Chief Strategy and Corporate Development Officer, athenahealth
Software vs. Services
A lot of the talk about AI in healthcare frames it as a clean break from the past. I see it differently. AI has revived a decision provider organizations have faced since the early days of the EHR, and it has made that choice much harder to avoid: Are you buying software, or are you buying an outcome?
For most of the last two decades, the market answered “software” and defined the space as the “EHR market.” The regulations and incentives of the late 2000s and 2010s calcified this. Buyers paid for seats and licenses, and they compared vendors on features and roadmaps.
We at athenahealth have thought this was a little odd for some time. Our bias has always been to get providers paid for doing the right thing, and to focus on the service outcomes that improve practice performance. In fact, before Meaningful Use, our original clinical workflow concept was a sort of anti-EHR: a clinical work management service focused on documents, authorization, and coding. When Meaningful Use arrived, we changed course and built a cloud-scaled EHR.
However, we kept some of the concepts from this service, including embedded document processing and in-app benchmarking for key clinical and operational tasks.
This resulted in some (at the time) strange choices for an EHR. We included embedded service operations, and we took over clients’ fax lines. We called this a “software-enabled service.” Despite all this, and our extensive RCM offerings, we are referred to mostly as an EHR company.
Bringing service back
Lo and behold, now in the 2020s we have a new term: “service as software” (SaS). AI has renewed the way to think about technology – deploying it to deliver work, not just a workflow. Now, virtually all software and SaaS companies face this identity crisis, which is very familiar to us. It raises the question: Have software companies always secretly been service companies?
This line of thinking is as important as ever. Demand continues to climb, while staffing gaps persist. Administrative work also absorbs a huge share of resources; by some estimates, it accounts for roughly 30% of U.S. healthcare spending.
That is where AI has the most to offer today. I think of it as the “backstage” of healthcare: the preparation and follow-up that make a visit possible and make sure it gets paid. The work doesn’t disappear, but the automation empowers organizations to deliver it faster and at a lower cost. At athenahealth, we deployed AI into our existing services model to achieve results that weren't possible in our old software plus services model. The new mix has allowed us to expand our scope, offering automated services like data extraction and chart population, that dramatically streamline workflows.
At the same time, most people just want to feel cared for in their most vulnerable moments, and they notice when they are being processed instead. Organizations that deploy AI well will be deliberate about where efficiency helps and where a person needs to remain in the room.
Questions for providers
If service outcomes are the product, the evaluation process should reflect that. What an EHR is, and what providers want it to be, is an open question. In the meantime, here are the questions I would want answered before signing.
Who is accountable for the result? A services orientation means the vendor shares responsibility for outcomes as well as functionality. Ask how the partner thinks about measurable efficacy and operational savings. Features and functions still matter, but, on their own, they are no longer enough. Paired with that is the business and pricing model: Does it create the right incentives for the service provider?
What happens at the edges? Ask whether the AI offers a complete service across every use case, or whether it is backed by staff who handle the edge cases. Whether we like it or not, faxes and paper claims still flow freely in our healthcare system, and agents need oversight.
How much context does the agent have? Agents that lack a full view of the patient, or can’t transact across the systems they touch, will be limited and more likely to make mistakes. Different models are also built for different jobs, so ask how well the platform connects is connecting data from the broader health care ecosystem.
Will this decision still be sound in two years? The last 24 months have shown how quickly assumptions change and how easily a purchase can turn into a regret. Ask whether the platform integrates new capabilities natively without rebuilding from scratch. Point solutions are already consolidating into broader platforms, which makes an open ecosystem more valuable every year.
The bigger stakes
These questions matter because the system has very little slack left. Clinicians are absorbing administrative load the system can no longer carry, and patients feel the effects in access and experience.
The upside is real. AI-enabled services connected to a broad healthcare ecosystem could give physicians more choice in how and where they practice, with support that travels with them.
None of that is guaranteed. But the organizations that press their technology partners on accountability and flexibility will be far better positioned to protect access and support their clinicians – and communities – through what comes next.
The editorial staff had no role in this post's creation.