Before buying AI, ask what it will do for your practice's bottom line

Silicon Valley VC Vinod Khosla recently told Fortune that “every hospital system is racing to adopt AI.” It’s the kind of sweeping claim that could apply to nearly any corner of health care — pharmacies, physical therapy clinics, durable medical equipment suppliers — all chasing the same trend. AI has clearly reached peak hype.

AI is indeed powerful. But I am a provider of AI-powered technology solutions for the health care industry and here is the truth: Not everyone needs AI or stands to extract any value from its use. While the potential impact of AI appears to be a gamechanger for everyone, most specialty practices and practice groups can’t justify adopting a technology simply because venture capitalists insist everyone else is doing it. Health care doesn’t have an AI problem, it has an operational capacity problem. And leaders aren’t really racing to adopt AI, they’re racing to eliminate bottlenecks that thwart successful operations.

A healthy dose of skepticism

Consider a parallel case of hyped technology. Jeremy Clarkson, famed British auto enthusiast and star of Clarkson’s Farm, has long been skeptical of self-driving cars. To paraphrase his argument: A “driverless” car can conceivably take you down the road and park itself at the shopping center. But you have to go along for the ride anyway and the car can’t walk into the store and buy your groceries when you get there. So why not just drive yourself? It may be cool tech, but it doesn’t save you any time and you don’t really gain anything by using it.

Yet, Clarkson feels differently about a self-driving tractor that autonomously plows and cultivates his fields. In this case, the technology expertly conquers repetitive and time consuming tasks while freeing up hours of Clarkson’s day for other work. He’s a raving fan, not because of the novelty, but because its usefulness is so clearly apparent.

This is the type of “what’s in it for me?” approach that healthcare practices should take with AI adoption for their organizations.

Five questions every practice leader should ask before buying AI

If staff are frustrated by work that technology should already be doing, then AI may present a solution. To evaluate whether an AI tool is worth adopting, practices should weigh five considerations:

1. Does this remove work — or simply change who does it? Many AI tools create another dashboard. The best eliminate steps.
2. Does it fit existing workflows? Less work, not more. Health care doesn’t need yet another login.
3. Can staff trust it? Human oversight. Confidence scores. Exception handling. Audit trails. All necessary!
4. Can you measure value in days — not months? Ask for real-world case studies and evidence of results with the tool or service.
5. Will this still matter after the AI hype fades? Subscription costs and fees cannot outweigh real financial value or time savings.

Real-world example

I know of a multi-location ophthalmology and optometry group practice in the Southwest that wanted to replace a failing on-premise fax server. Their existing setup had become an operational liability. Faxes were silently failing. Confirmations were unreliable. Eight staff members rotated through manual fax sorting, with two employees spending entire days on the task. Every inbound document required someone to open it, identify its type, search for the matching patient and file it in the EHR by hand — a repetitive process that took roughly four minutes per page.

The practice wasn’t shopping for an AI solution per se. They were simply trying to fix a bothersome referrals process. They selected tooling that automated much of the process, and the results were substantial. Implementation took only four weeks and per-document handling time pretty much halved. In the first full month live, the team processed 3,700 referrals — a 25% increase in volume with no added staff, largely because previously hidden faxes were now visible and searchable. Modeled across the practice’s referral volume and locations, the change represented an estimated $3.6 million in annual economic impact through recovered revenue and labor savings.

Staff also gained the ability to instantly locate any inbound fax, even one not yet fully processed, which meant patients calling to check on a referral could get an answer on the spot instead of a “we’ll get back to you.” Support tickets dropped sharply, and rollout itself was fast: Most staff needed only about 30 minutes of training.

The fact that AI technology is incorporated in the new tooling is practically beside the point. What matters is that it addresses their problem, and it’s actually useful.

The takeaway

The practice didn’t set out to adopt AI because everyone else seems to be, and it didn’t try to shoehorn in some cumbersome new platform that would disrupt their daily routines. It just decided to fix a broken process, and AI happens to power the tool that solved it — quickly, measurably, and without disrupting existing workflows. That’s the model specialty practices should be looking for: Technology judged not by how advanced it sounds, but by what it actually returns to the bottom line.

Denis Whelan is the CEO of Documo, a leading provider of cloud fax and Intelligent Document Processing (IDP) solutions for health care organizations. With a deep understanding of health care workflows and compliance requirements, he helps organizations modernize document processes, reduce administrative burden, and accelerate patient care. Denis is passionate about bridging the gap between legacy systems and modern automation, transforming fax-driven and manual processes into secure, streamlined, and actionable workflows. His work focuses on advancing AI-driven solutions that improve efficiency, reduce errors, and give healthcare organizations greater operational visibility-so care teams can focus on what matters most: delivering better patient outcomes.