Healthcare providers are eager to use artificial intelligence for the flood of appointment reminders, test‑result notifications and prescription alerts that flood patients’ phones each day. Yet a recent Sinch survey of 470 global health‑tech leaders reveals a troubling pattern: 75% of organizations that have deployed AI agents have been forced to pull one back because of governance failures, most often the accidental exposure of personal health information.
Widespread rollbacks despite rising spend
Three in four providers running AI‑driven patient communications have already had to switch off an agent, the report says. While 98% of all industries plan to increase AI communications budgets in 2026, healthcare lags behind, with only 16% of respondents expecting a 50%+ spend boost—still above the overall 13% average.
Key use cases and expectations
Providers see the biggest opportunities in outbound voice agents for appointment reminders (22% of leaders rank this as the top AI use, five points above the 17% industry average) and identity verification before patients can view test results or confirm prescriptions (41% place this near the top of their AI priorities).
Governance gaps, not compliance complexity
When AI agents go live, healthcare matches other sectors on 19 compliance and privacy measures, with a failure rate only one point higher than the overall average. The real issue appears to be internal misalignment: senior executives report mature guardrails at a rate of 36%, while the directors tasked with building those safeguards report only 18% maturity. This disconnect leads to under‑funded risk‑mitigation work and, ultimately, data leaks.
Data leakage as a leading cause
One‑third of rollbacks were triggered by personal identifiable information (PII) exposure—such as a test result sent to the wrong patient or a medication record appearing in a simple appointment reminder. These incidents stem from infrastructure failures; the underlying platform failed to block the data from surfacing.
Investment focus shifts to infrastructure
Healthcare leaders are already prioritizing trust, security and compliance (75% of respondents) and are placing communications infrastructure above AI agent development (64% versus 58%). They seek partners that can deliver reliable, compliant and accountable platforms, recognizing that a solid foundation is essential to prevent future rollbacks.
Looking ahead
Despite the setbacks, 55% of healthcare organizations have AI agents in production, though this is still seven points below the overall industry average. Only 42% of health‑tech leaders say they are very confident about scaling AI at speed. As the sector continues to invest in the right infrastructure, the expectation is that AI‑driven patient engagement will become more reliable and secure, protecting both data and the vital trust patients place in their caregivers.
Original reporting: KTVZ (Central Oregon) — read the source article.
