Rearchitecting how communities access care in the AI era
Byline: Erin Reilly , Chief Social Impact Officer at Twilio
Abstract
Healthcare is fundamentally built on trusted conversations. Increasingly, AI agents are joining these conversations, and reshaping how we deliver and access care. As agents move from an era of passive automation to autonomous task execution, we face new and necessary challenges around governance, security, and human-in-the-loop oversight. Realizing the true promise of agentic AI requires looking beyond surface-level applications in favor of foundational infrastructure that our communities deserve. And the leaders who are bold enough to build it.
Introduction
Nearly thirty years in social impact have taught me a simple truth: new technology forces us to ask better questions, and reframe the challenges we’re trying to solve. Instead of automating what we’ve always done, agentic AI invites us to reshape how work gets done itself.
While AI agents cannot provide or interpret medical treatment, they are powerful tools that can organize, summarize, and present information to assist care teams and streamline manual tasks. Industries with the most to gain sit at the intersection of high service volume and heavy administrative burden.
Community pharmacies, managing billions of patient touchpoints annually, live in this reality. For them, agentic AI shifts the question from “How do we staff around administrative bottlenecks?” to “Which patient interactions truly demand clinical judgment and empathy—and how do we ensure our people can focus on them?”
For healthcare organizations tackling the world’s biggest challenges, this shift is critical. It allows us to spend less time managing administrative burden, and more moments maximizing human impact.
Moving beyond passive automation
Agentic AI is beginning to disassemble the administrative maze between a prescription being written and a patient starting therapy. Instead of leaving clinicians and patients to navigate stalled prior authorizations or fragmented copay programs, intelligent agents can autonomously query insurer portals, verify inventory in real-time, route financial assistance, and coordinate delivery schedules behind the scenes.
Yet in regulated care, the capabilities that make agentic AI so compelling are exactly what make governance so complex. Every system component that’s connected to an autonomous agent expands the surface area for risk—from access control gaps to failed clinical handoffs.
Autonomous execution demands foundational control by design. One-off agentic workflows are great for pilots, but are ultimately a short-term fix for broader architectural challenges around interoperability, granular access controls, and human-escalation paths. When we build systems grounded in the needs of our communities, we can transform emergent AI capabilities into the resilient infrastructure our communities deserve.
Infrastructure in action
Last year, Twilio supported more than 30,000 social impact and healthcare organizations in reaching 803 million people globally. We’re proud to partner alongside leading health systems, community pharmacies, and health-tech innovators to power millions of trusted care conversations every day.
The University of California Davis Health demonstrates the power of an infrastructure-backed approach, serving over 6 million patients. To serve the needs of a diverse patient population, including rural, migrant, and elderly communities, they recognized that no off-the-shelf application could act as a true digital front door. So they built their own. Using Twilio as their foundation, they built UCNOW to engage patients on the channels they already depend on. This flexible framework gave them the operational control needed to create a “digital bridge” that expands care access and maintains clinical integrity.
New York University Langone Health highlights how foundational infrastructure scales over time. Anchored by six hospitals and over 300 outpatient facilities, they leveraged Twilio to rapidly scale patient outreach during the Covid-19 pandemic. They then used that same foundation to synthesize historical care data that powers automated, personalized follow-up workflows across their network.
While NYU Langone and UC Davis serve radically different populations—from NYC’s dense metro center to California’s rural agricultural valleys—both demonstrate how re-envisioning patient access starts with a foundational strategy. Neither relied on legacy architecture or singular offerings; both built a unified engagement layer that meets diverse community needs and streamlines clinician workflows.
Architecting for autonomous support
AI agents cannot replace human clinical decision making, but as they start executing autonomous workflows, establishing a foundational approach can’t be an afterthought. An AI model alone cannot safely navigate a patient’s care journey. Models reason, but infrastructure acts—and frontline care demands both.
For healthcare leaders operationalizing agentic workflows, bridging intelligence and execution requires four core architectural pillars:
❖ Identity & behavioral guardrails
➢ What is the agent authorized, and restricted, to execute?
❖ Persistent context across channels and systems
➢ How does the agent preserve context across channels and systems? ❖ Human-in-the-loop handoffs
➢ How does the agent route to human teams when risk emerges?
❖ Real-time observability & auditability
➢ How do we monitor system behavior and maintain regulatory compliance?
Identity & behavioral guardrails: Treat every AI agent as a credentialed digital team member with time-bounded permissions and strict scope-of-practice limits. Programmatic guardrails allow routine administrative and operational workflows—like patient intake or refill reminders—to execute autonomously, while automatically blocking unapproved actions before they create risk or exposure.
Persistent context across channels and systems: Autonomous care fails when infrastructure forgets. When an agent loses context the moment a patient switches channels or forces them to re-explain their situation each time they call in, it drives escalation costs and lowers care engagement. Maintaining a secure, real-time memory layer across communication channels and backend systems prevents repetitive questioning and outreach, and turns potential drop-off points into uninterrupted access to care support.
Human-in-the-loop handoffs: Clinical trust in agentic workflows depends on a seamless, deterministic escalation path to live care teams. Unlike traditional customer service frustration, healthcare conversations can carry implicit signals—like a patient underplaying severe side effects or quietly stretching a prescription—buried in daily dialogue. To detect these subtle nuances, healthcare AI agents must be able to evaluate real-time conversations against persistent context, and carry the relevant interaction history forward so the human handoff remains continuous and fully auditable.
Real-time observability and auditability: Expanding AI agency across patient workflows is impossible without continuous, traceable oversight at every touchpoint. Ensure observability is built directly into the execution layer to dynamically monitor sentiment and record traceable audit logs for every interaction.
The future belongs to the builders
Nearly two decades ago, Twilio started with a simple idea: democratize communications for all. Since then, Twilio has become a trusted foundation, running the global infrastructure beneath billions of daily conversations.
We believe deeply that the future is for the builders, and the world’s most critical challenges will be solved by the leaders bold enough to build the infrastructure their communities deserve.
The next generation of healthcare leaders won’t treat AI as another application to layer onto legacy systems. They’ll be the ones willing to do the foundational work of breaking down data silos, designing interoperable systems, and integrating disjointed workflows.
That vision is why Twilio is proud to serve as the Cloud Communications sponsor for Pharmacy For All. We are committed to providing the resilient, secure foundation that healthcare innovators need to re-imagine patient access for the next generation of healthcare engagement.
Whether you’re modernizing core engagement channels or deploying autonomous, multi-step AI workflows, Twilio manages the foundational complexity, so you can focus on delivering exceptional care.
We can’t wait to see what you build.
Learn more about Twilio for Healthcare, today.