Readers Write: Chatbots Are Repeating a Familiar Healthcare Mistake

Chatbots Are Repeating a Familiar Healthcare Mistake
By Robin Monks

Robin Monks is chief technology officer of Praia Health.

I have five healthcare applications on my phone: multiple patient portals, a lab app, a health tracker, and a record repository from a prior provider. None of them communicate with each other. I work in healthcare technology and build patient experiences for a living, and this is still the reality. That should be a red flag.

Health systems have spent years layering digital tools onto already complex environments. The latest additions are chatbots and AI assistants, each of which are designed to solve a narrow problem. Individually, they make sense. Collectively, they recreate the same fragmentation that digital transformation was supposed to fix.

This pattern has played out before. Mobile strategies in the 2010s produced a flood of standalone apps that patients downloaded once and then abandoned. The industry burned time, money, and credibility getting them into market.

A similar cycle is now underway with AI. Tools are being deployed quickly, often without a coherent strategy. The World Economic Forum has already warned that most health systems lack a defined generative AI strategy despite rapid adoption. That gap between enthusiasm and discipline rarely ends well.

Multiple chatbots now exist within the same health system, each tied to a specific function. From a patient perspective, distinctions between vendors or capabilities are meaningless. Every interaction reflects on the organization. When one tool fails, trust erodes across all of them.

Many of these tools cannot complete the jobs they start. A scheduling request ends with instructions to call an office, and a refill request redirects patients to another app. Each handoff introduces friction, and each point of friction increases abandonment. The industry has built an impressive number of ways to begin a task and very few reliable ways to finish one.

Identity is another weak point. Many tools operate within a single channel, often SMS or a standalone app, without a consistent identity layer. Context is lost as soon as a patient moves between touchpoints. The system does not recognize the same individual across channels, so the experience resets.

The result is confusion about how to complete even basic care tasks. The ONC reports that most individuals now juggle multiple patient portals, yet almost none use tools to manage them. When the path forward is unclear, engagement drops, preventive care is delayed, and follow-ups are missed. The gap between intent and action widens.

Clinical systems learned this lesson years ago. Fragmentation in clinician workflows was treated as a patient safety issue, driving consolidation into unified records. The same logic applies to the patient experience. Fragmentation on the front end produces the same outcome: missed steps, incomplete information, and avoidable risk.

Adding more tools will not fix this. The problem isn’t a lack of functionality, it’s a lack of orchestration.

A workable approach starts with data that spans the full care journey rather than being trapped in individual systems. National efforts such as TEFCA and standards like FHIR are making this increasingly feasible. The infrastructure is emerging, but the industry’s track record suggests it may still find a way to misuse it.

Identity must also be treated as foundational rather than optional. A consistent, portable identity allows continuity across channels and services. Without it, every integration is shallow, and every experience is brittle.

Most importantly, digital experiences must enable action. Many current solutions are little more than read-only interfaces. They show information, but cannot do much with it. A useful system allows scheduling, making payments, obtaining referrals, and performing follow-through without forcing patients to navigate a maze of disconnected tools.

None of this is conceptually difficult. The challenge is discipline. Health systems continue to approve new tools faster than maximizing existing ones. Vendors continue to sell point solutions that solve isolated problems while ignoring the broader experience. The result is more complexity, more fragmentation, and diminishing returns.

Healthcare is approaching another familiar fork in the road. One path continues the current trajectory: deploy more AI tools, watch adoption plateau, and quietly move on to the next trend. The other path requires doing the harder work of integration and orchestration, using emerging interoperability infrastructure to build experiences that actually hold together.

The technology is no longer the limiting factor. The limiting factor is the willingness to stop adding digital point solutions and start designing systems that function as a whole. Without that shift, the industry will repeat a cycle it should have already outgrown, replacing one generation of digital clutter with another.