About 30% of artificial intelligence (AI) pilots intended for health care applications reach the production phase of the software development lifecycle, based on a recent survey conducted by Bessemer Venture Partners, Bain & Company, and Amazon Web Services. The survey respondents said that security, data readiness, integration costs, and limited in-house expertise affected whether a pilot reached production. However, 80% of surveyed leaders across the more than 400 included health care payers, provider organizations, and pharmaceutical companies said they believe that generative AI (GenAI) will transform clinical decision-making in the near future.
GenAI is an AI platform that creates new text, images, and videos based on patterns it assembles from existing data. GenAI platforms use advanced algorithms and large datasets to generate new outputs in response to user prompts.
About 32% of surveyed executives believe that GenAI solutions from startups are superior to those from existing large tech companies. About 48% of the surveyed executives said that they prefer working with startups rather than existing tech companies. Two-thirds (64%) of executives were willing to co-develop AI strategies with newer companies, especially those that show clear return on investment.
As of the time of the survey, about 15% of AI projects involved vertical AI applications specifically designed for particular industries or use cases, as opposed to general-purpose AI that could be adapted for various industries. The majority of AI projects are developed internally or in partnership with horizontal AI labs or large technology companies that develop general-purpose AI applications. About one-third of companies develop their AI projects in-house.
These findings were presented in The Healthcare AI Adoption Index, by Sofia Guerra, Steve Kraus, and additional editors from Atlas. The goal was to understand how health care buyers are approaching AI applications, including related jobs that they prioritize, how they are making adoption decisions, and the status of their AI projects.
The full text ofThe Healthcare AI Adoption Index was published on April 15, 2025, by Bessemer Venture Partners. A free copy is available (accessed June 19, 2025).
For more information, contact: Bessemer Venture Partners, 889 Winslow Street, Suite 500, Redwood City, California 94063; 650-853-7000; Website: https://www.bvp.com/
August 2025 00US25EUA0023