TRIMEDX Chief Data & AI Officer Srilekha Akula discussed the impact of AI on healthcare technology management (HTM) at the 2026 Healthcare Information and Management Systems Society (HIMSS) global health conference and exhibition. AAMI News recently wrote an article about Srilekha’s presentation, detailing five principles for successfully implementing AI in health systems.
What does it take to maximize the impact of AI in HTM and deliver real results to HTM professionals and patients alike? At HIMSS 2026, TRIMEDX Chief Data & AI Officer Srilekha Akula discussed TRIMEDX’s path to successfully integrating AI into its clinical asset management business.
Today, the health system faces a sector-wide transformation, with research from the Harvard School of Public Health indicating that AI could lead to a dramatic reduction in treatment costs and improvements in health outcomes. For a country like the United States, which lags behind its peers in health care outcomes and ROI, this is good news.
Akula said that TRIMEDX has met the moment with three core tenets: First, innovate with intention. Second, maintain a client-centered approach. Third, balance speed with patient impact. “‘Move fast, break things’ does not work” when patients are involved.
Five principles for successful implementation
According to Akula, “data matters more than algorithms.” The TRIMEDX view is that, as foundational models improve, there is no need to build an in-house model. Also, given that the health sector is subject to serious risk and regulation, this creates the need for “extreme accuracy, reliability, and trust.” Akula said, “Focus on your data, get your data right.”
Second, AI should meet your data where it lives. Distance and data manipulation both degrade the output quality of a given AI model. Instead, running application logic as close as possible to the source data maintains clinical context, improves accuracy, enables effective automation, and builds end-user trust. This also reduces the deterioration of inference.
Third, AI may reveal unexpected value. Using a large, longitudinal data set strengthens model reasoning and can allow an organization to generate an idea of what ‘good looks like.’ Some interesting findings TRIMEDX has made during its AI implementation include that OR inefficiency is driven more by device readiness issues than by scheduling difficulties.
Fourth, “AI should blend in, not stand out.” Artificial intelligence should fit into existing flows, for her, and also for customers. Ideally, it should be built into the workflows and tools that already exist, and “improve outcomes not just dashboards.”
Finally, culture will determine AI success. According to Akula, “Without culture, things are not going to move.” To successfully implement AI, you will need to “Meet your technical teams where they are”, involve experts and your vendors, and tap into your super-users while planning for detractors.
Navigating implementation pitfalls
According to Akula, AI’s potential to improve BMET productivity is a boon amid an aging workforce and a wave of retirements. AI also has the potential to drive cost savings and revenue generation from a given HTM department using advanced data analytics. She also sees potential in reducing cybersecurity risks through intelligent threat detection and in preventing device failures before they occur.
How should organizations measure success? Operational gains, adoption and satisfaction, patient care impact, financial improvements, and data quality and effectiveness are all good metrics to adopt. Challenges remain, including cost, data quality and availability, and stakeholder misalignment, but to date, TRIMEDX has enjoyed significant success in utilizing 20 years of longitudinal data on more than 6.1 million medical devices. The future for AI in HTM is bright, and successful implementation can improve capital planning, help predict downtime, and speed up supply chain processes across the entire field.