Are Healthcare AI Investments Delivering Real Value or Just Hype?

The Impact of AI on Healthcare Investments

AI is making waves across healthcare, from nurse navigators using AI triage assistants to speed up case prioritization, to tools that help providers document visits and automate repetitive administrative tasks. As these applications multiply, investments in AI have surged. But healthcare systems are now questioning if the returns justify the costs.

Many organizations pay premium prices for AI tools that save only minimal time—sometimes just five minutes a day—making it tough to justify the investment. There’s a clear mismatch between what health systems pay, the limited value current AI tools deliver, and the more advanced systems expected in the near future.

Many adopt AI because it feels like a “must-try” technology, but it’s still early days. Numerous AI vendors may not survive because they can’t prove value when renewal time arrives. This leads to hesitation and fatigue among healthcare organizations toward certain AI providers.

Cost Concerns in Healthcare Technology

According to the 2024 Healthcare IT Spending Report from Bain & Company and KLAS Research, nearly half of healthcare providers identify cost as their biggest pain point with existing technology stacks. Expensive AI tools that show limited return only exacerbate this issue.

Often, AI tools perform well in pilot phases but struggle when scaled across entire health systems, especially where integration with complex workflows is needed. Some vendors may claim wide usage, but it might only be a single researcher in one department actually using the tool. This means such AI solutions aren’t yet proven or optimized for large-scale deployment.

Healthcare organizations should ask their current vendors about AI strategies before jumping to new companies offering single-solution tools. Established vendors that integrate AI into existing workflows tend to offer more sustainable value than point solutions that might not last long.

Barriers to Implementing Generative AI

The same report highlights regulatory and legal concerns as the top barriers to adopting generative AI, cited by 38–43% of respondents. These issues slow down implementation.

Many AI solutions are built on public models like ChatGPT, lightly customized and repackaged for healthcare. Though these tools may sound innovative, their actual impact is often modest.

Billions have already been invested in healthcare AI, with billions more on the way. Yet, it remains unclear how much of this will translate into better patient care or meaningful time savings. The World Economic Forum notes it’s too soon to say whether generative AI will help, harm, or simply waste resources without improving lives.

Questions to Consider Before AI Purchases

The promise of AI is real but many current tools aren’t ready for widespread use. This forces health system leaders to rethink their investments. Managing multiple point solutions—one for documentation, another for billing, a third for patient follow-up—adds cost and complexity over time. CIOs often spend as much time fixing integrations as they do evaluating new tech.

As a result, many health systems are refocusing on their core platforms, asking how AI is integrated there rather than chasing new vendors. These integrated solutions might not get as much buzz but often provide a more stable, less disruptive path.

One example is Kaiser Permanente, which applies system-wide AI governance across research, clinical operations, education, and administration. They emphasize AI as a tool to support clinician judgment, not replace it. When launching generative AI tools, oversight and deliberate design are key.

Healthcare organizations should pause and ask these foundational questions before investing in AI:

• What problem does this tool actually solve? Seek tools that target specific operational bottlenecks with measurable outcomes. Set baseline metrics before pilots to track improvements. If a tool only automates existing processes without improving outcomes or staff satisfaction, it’s likely not worth it.
• How much time and money does it realistically save? Include implementation, training, and support costs when calculating ROI. Spending over $1,000 per user annually to save less than 15 minutes a day usually isn’t sustainable. Focus on tools that remove entire workflow steps, not just speed them up.
• Is this a pilot, or proven to scale? Demand evidence of successful use across various organizational sizes and settings. Look for consistent results across broad patient groups and clinical environments before committing to full deployment.
• Will this fit into our existing system, or add another layer? Prioritize AI tools that integrate smoothly with your EHR and reduce system complexity. Be cautious of solutions requiring extra data entry, separate logins, or workflow disruptions.

Healthcare systems need a realistic approach to AI investments—one that distinguishes real value from tools that only shine in demos.