Why Seven in Ten Healthcare Systems Can't Get AI Out of the Lab
A new Google Cloud survey reveals that cost unpredictability and integration hurdles are trapping healthcare AI in pilot purgatory, with the sector lagging 13 points behind other industries in enterprise deployment.

Illustration: AI-generated
Healthcare organizations are racing to adopt artificial intelligence, but most are hitting a wall when trying to move beyond small-scale experiments. According to a new Google Cloud report, only 30% of healthcare organizations have successfully scaled advanced AI across their operations—trailing the cross-industry benchmark by 13 percentage points and highlighting a sector-specific struggle to translate pilot success into enterprise reality.
The survey, conducted by Hypothesis Group for Google Cloud and polling more than 250 healthcare executives, paints a picture of an industry caught between enthusiasm for AI's potential and the practical challenges of making it work at scale. While healthcare providers have rapidly adopted AI tools for both administrative and clinical tasks in limited settings, the transition to organization-wide deployment is proving far more difficult than anticipated.
Cost Chaos Stalls Deployment
Financial unpredictability emerges as a primary culprit in the scaling crisis. Forty percent of healthcare executives cited unpredictable costs as a reason for pausing AI initiatives altogether, while more than 80% identified cost as a key barrier to reducing dependence on existing AI vendors. The variable, usage-based nature of AI expenses—often measured in "tokens" that represent units of computational work—doesn't fit neatly into traditional IT budgeting frameworks.
An Accenture report referenced in the findings suggests a potential solution: companies adopting "tokenomics" strategies avoided a projected 44% increase in token costs over two years. Aashima Gupta, Global Director of Healthcare at Google Cloud, is advocating for healthcare organizations to embrace these cost management approaches as AI becomes more complex and resource-intensive.
Integration Headaches and Data Security Concerns
Beyond cost, integration and deployment challenges plague nearly 60% of surveyed healthcare executives attempting to move AI from pilot programs to enterprise-wide systems. The complexity of connecting AI tools with existing electronic health records, clinical workflows, and administrative systems creates technical bottlenecks that pilot programs don't reveal.
Data security adds another layer of difficulty. Forty percent of healthcare respondents named protection of sensitive data as especially important—nine percentage points higher than the cross-industry benchmark. This heightened concern reflects healthcare's unique regulatory environment and the sensitivity of patient information, making organizations more cautious about widespread AI deployment.
The Multi-Model Solution
Google is positioning a multi-model approach as the path forward. The company recently launched Gemini for healthcare, a platform designed for providers and payers that enables cross-competitor model usage from Google, OpenAI, and Anthropic under unified governance. This strategy aims to give healthcare organizations flexibility and customization capabilities while maintaining the security and oversight requirements that make the sector particularly risk-averse.
The report's findings suggest healthcare organizations may need to create new organizational structures to manage AI at scale, with some experts pointing to the potential need for dedicated roles focused on AI cost management and governance. As the technology continues to evolve, the gap between pilot enthusiasm and enterprise execution remains healthcare's most pressing AI challenge.
Sources
Follow Health AI Central
Keep up with health AI where you already read
On LinkedIn
The day’s strongest health AI stories in your feed, one or two each weekday.
On Google
Choose Health AI Central as a preferred source, and Google is more likely to show our reporting in Top Stories.
Reporting & accountability
How we maintain accuracy
Health AI Central separates reported facts, analysis, and public-source data. Readers can review our standards, request a correction, or consult the relevant research methodology before relying on a claim.