Lirio

What it is

Lirio is a healthcare personalization platform built around Precision Nudging: behavioral interventions tuned to an individual's specific barriers and delivered through a proprietary Large Behavior Model (LBM), a model class that predicts a person's next health behavior the way an LLM predicts a next word, but trained on health behavior data instead of language. Rather than sending one generic reminder, Lirio treats every nudge as four independent decisions — content and framing, channel, time of day, and frequency — and sometimes routes the message to a caregiver instead of the patient when that person is the more effective intervention target. The platform is headless: it sits behind the CRM, data warehouse, marketing automation, and engagement tools a health system already owns rather than asking anyone to adopt new software or download an app. Lirio runs a comprehensive risk and compliance framework (HITRUST CSF, SOC 2 Type II, NIST, plus country-specific controls) that its own AI-assisted development work has to operate inside. It delivers more than 400 million nudges a year and reports outcomes including 4x more effective messaging versus standard health communication, a 60% reactivation rate on patients lost for three-plus years, and 50% HbA1c improvement in under six months.

Who it's for

Health systems, health plans, retail pharmacies, and life sciences companies that already have a CRM and engagement stack but are losing patients and members in the 8,760 hours a year outside of clinical contact, and want that existing stack to actually change behavior instead of just sending reminders.

Why it's better

• •Precision Nudging tunes four variables per message — content, channel, timing, frequency — instead of the one-size-fits-all reminder most systems send.
• •Headless architecture integrates with the CRM and engagement tools a health system already bought, so the pitch is 'your stack, now effective,' not 'buy more software.'
• •Caregiver-aware targeting routes the nudge to whoever is actually positioned to act, not reflexively the patient.
• •Built and audited inside a regulated-industry compliance framework (HITRUST, SOC 2 Type II, NIST) from day one, not retrofitted.
• •400M+ nudges a year gives the Large Behavior Model a scale of real-world behavioral data most health tech vendors don't have.