Taction Software launches AI governance guide as states tighten healthcare oversight

States are moving to require human review of AI-driven coverage and clinical decisions, and Taction Software has released a free governance guide aimed at helping healthcare organizations adapt. The guide arrives as new laws, litigation, and weak internal controls raise compliance and liability risks for hospitals, payers, and digital health vendors.

Why it matters: - State laws are now forcing healthcare organizations to prove that humans, not just models, are making or reviewing coverage and clinical decisions. - The new rules raise the bar for auditability, transparency, and liability in systems that use AI for prior authorization and medical necessity reviews. - Organizations that cannot show how an AI recommendation was reached may face regulatory scrutiny, operational disruption, and legal exposure.

What happened: - Taction Software, a Chicago-based healthcare IT and custom software development firm, released a practical guide to responsible AI governance for healthcare organizations. - The guide is meant to help hospitals, digital health companies, and payers translate new state AI requirements into software architecture and workflows. - Taction Software is offering the guide at no cost to healthcare organizations evaluating their exposure to the new rules. More information

The details: - As of 2026, 37 states have enacted or introduced legislation governing AI in healthcare. - Roughly a dozen states now require a qualified human to own any medical necessity denial. - Washington's Senate Bill 5395 took effect June 11, 2026, and bars insurers from relying solely on AI to deny prior authorization based on medical necessity. - Washington also requires a licensed physician or health professional to make that denial. - The law requires the reviewer to examine the requesting provider's recommendation, the enrollee's medical history, and the enrollee's individual clinical circumstances. - Washington also requires AI review criteria to reflect the individual patient rather than group data alone. - Maryland's HB 1563 took effect June 1, 2026, and requires health insurers to report quarterly to the Insurance Commissioner the number of adverse decisions issued, the type of service involved, and whether AI was used. - Maryland's law gives the commissioner authority to investigate insurers with significant increases in adverse determinations, especially emergency department denials. - Georgia's SB 444 takes effect Jan. 1, 2027, and allows insurers to use AI in prior authorization, but only after a qualified human conducts utilization review with a clinical peer participating. - Taction Software says the guide covers human-in-the-loop review workflows, decision traceability, HIPAA-aligned audit logging, model documentation, bias monitoring, and vendor oversight. - The guide maps those technical controls to obligations that state laws are beginning to impose. - Taction Software says the guide is vendor neutral and focused on architecture rather than any single product.

Between the lines: - The release is responding to a readiness gap: AI adoption in healthcare has moved faster than governance. - A January 2026 Health Affairs study led by Stanford researcher Michelle Mello found many insurers lack robust governance processes to monitor the accuracy and bias of AI tools used in utilization review. - A 2024 survey of 93 large health insurers found 84% were already using AI for some operational purpose. - Public trust has not kept pace, with two-thirds of U.S. adults saying they have little trust that AI will be used responsibly in healthcare. - Litigation is starting to test whether automated decisions meet the individualized review promised in insurance contracts. - Humana is being sued in Kentucky over its use of the nH Predict tool, with patients arguing the AI ignored specific clinical circumstances. - Taction Software says the new rules make traceability and human oversight an engineering problem as much as a policy problem.

What's next: - Healthcare organizations building or buying AI for coverage and clinical workflows will need to tighten governance now, not after regulators intervene. - The company says decision traceability, tamper-evident logs, continuous bias monitoring, and vendor oversight should be built in from the start. - Organizations that already operate under HIPAA and FDA constraints may have an easier path to compliance, but the legal baseline is moving upward. - Taction Software says the guide is designed for technical and compliance leaders who need to make those changes work in practice.

The bottom line: - State law is turning AI oversight in healthcare from a best practice into a requirement, and the winners will be the organizations that can prove a human was truly in the loop.