A digital divide is emerging in ambulatory care between patients — who increasingly rely on chatbots and virtual assistants to help with everyday tasks — and most of the clinics they visit, which often lack this level of technology at their digital front door.
An April 8, 2025, MGMA Stat poll revealed that only about one in five (19%) medical group practices use some version of chatbot or virtual assistant for patient communication, while 81% do not. The poll had 375 applicable responses.
Multiple market research firms estimate the global healthcare chatbot market has already surpassed $1 billion in 2025 — forecasted to reach or exceed $10 billion over the next decade.
Did you know?
- Anyone can use the Ask AI virtual assistant on MGMA.com to help them navigate resources on the MGMA website to have an AI-powered virtual assistant experience and get contextual responses to your questions.
- MGMA members can read about how Garnet Health uses conversational AI to address claim denials through pre-registration automation and digital follow-up in the latest issue of MGMA Connection magazine.
Today’s capabilities
Modern chatbots and virtual assistants perform a wide array of patient-facing tasks, helping them access information and services after hours or during peak times. Among the top capabilities today:
Appointment reminders and no-show reduction
Chatbots are used to send automated appointment reminders via text or portal message; the AI assistant enables appointment confirmation, rescheduling, or cancellation. Adoption of SMS text-based messaging platforms for patient communication — AI-powered or otherwise — was frequently cited by practice leaders in our August 2024 poll that found half of practice leaders reporting no change in their no-show rates last year.
Appointment scheduling and registration
One of the most impactful capabilities is freeing up your staff from phone calls by shifting appointment scheduling to a 24/7 chat interface that allows patients to find the appropriate providers and available time slots. This approach, used at Weill Cornell Medicine in recent years, led to a 47% increase in appointments books digitally via an AI chatbot.
Patient Q&A and self-service information
Perhaps one of the most simple and straightforward uses of virtual assistants and chatbots as a “digital receptionist” to answer common patient questions about clinic hours, directions, parking, and other routine inquiries. Multiple vendors in the conversational AI space have gone beyond automated conversations within secure portals and embedded versions of them directly onto providers’ websites.
Symptom triage and clinical guidance
One of the more intriguing types of chatbots are those that go beyond logistics and offer some level of clinical triage, asking patients about symptoms and providing guidance on managing at home, coming in for an appointment, or seeking emergency care. In the past year, EHR giant Epic recently tested a bot for MyChart for post-surgical patient communication and recovery tracking.
Medication refills and billing assistance
Health IT vendors have specifically focused chatbot development on cutting down on the phone time and tasks associated with handling patients’ medication refill queries via integration with the EHR and/or pharmacy system(s). Some chatbots also send medication reminders or post-visit care instructions to help improve adherence and patient understanding.
Multilingual support
Many solutions now ensure their chatbots can converse in Spanish and other common languages to ensure non-English speakers can use self-service tools, helping to reduce disparities in access to information.
- Did you know? MGMA Translate, powered by Boostlingo, is a technology-driven interpretation service designed to help healthcare organizations reduce costs, streamline scheduling and improve patient care. It offers on-demand interpreter access, seamless EHR integration, and data-driven insights to enhance efficiency. MGMA Translate allows medical groups to manage language services more effectively while saving time and resources.
Keys to integration
A critical factor for success in chatbot deployment is how well the tool integrates with existing EHR and practice management (PM) systems. From a technology perspective, integration can happen through application programming interfaces (APIs) provided by EHR vendors (e.g., Epic’s APIs, Cerner’s Ignite APIs, FHIR endpoints, etc.) or through HL7 interfaces. The depth of integration can vary:
- A “shallow” integration might just pull demographic info or upcoming appointments (read-only).
- A “deep” integration means the chatbot can create or modify records in the EHR/PM (write-back) — for example, creating or modifying records, scheduling appointments directly on the practice’s calendar, checking insurance eligibility, or documenting outcomes of patient interactions.
Full automation of a process ought to yield a better return on investment (ROI). Consider appointment scheduling: A standalone chatbot that isn’t integrated may only collect appointment requests and forward them via email to staff, which can introduce delay or risk of human error. In contrast, a well-integrated chatbot can check real-time availability, book appointments, and write the details directly into the EHR — all while the patient is still in the chat. This close-looped process creates a better experience for staff and patients.
Most major EHR companies have embraced interoperability with chatbot vendors. Epic and Cerner (now Oracle Health) have marketplaces where third-party digital front door tools can plug directly into their systems. Athenahealth, NextGen, and Allscripts (Veradigm) have similarly expanded their integration capabilities, leading to a growing number of products successfully connecting with systems.
When evaluating AI assistants for their practice, decision-makers should prioritize solutions that partner well with their EHR/PM platforms. Many case studies of successful chatbot deployments attribute their success to this kind of seamless integration in the workflow (often describing the solution as “embedded” in the EHR workflow).
Metrics to measure
To evaluate the impact of a chatbot or virtual assistant, practices should track the mix of operational, financial, and patient access metrics. Key performance indicators (KPIs) may include:
- No-show rate: Compare the no-show rate before and after chatbot implementation, and measure how quickly cancelled slots are backfilled.
- Appointment conversion and scheduling volume: Track the percentage of inbound inquiries that result in booked appointments, overall appointment volume growth, and the share of appointments made outside normal business hours.
- Call deflection/reduction in call center volume: Reducing incoming phone calls cuts staff labor, shortens wait times, and can improve patient satisfaction. Track change in the volume of inbound calls and average call handling time pre- vs. post-chatbot.
- Patient engagement and satisfaction: High satisfaction often means better loyalty and retention, so be sure to track Net Promoter Score (NPS), survey scores, chatbot usage rates, and completion rates for digital tasks (e.g., check-ins, forms).
- Staff efficiency/workforce metrics: Track changes in the ratio of support staff to providers, as well as average time to respond to patient requests.
- Same-day/urgent appointment fulfillment: Track the number of same-day appointments successfully booked via the chatbot, how many cancellations are automatically refilled, and any changes in access wait times.
- Revenue impact: Watch for changes in revenue-per-provider-day, total billed encounters, the percentage of successful balance collections (if the chatbot handles payments), and new-patient volume driven by AI-based scheduling.
- Escalation/handoff rates: Track the percentage of patient inquiries resolved entirely by the chatbot versus those requiring staff assistance, and reasons for escalation.
Practice leaders should select a subset of these metrics that best align with their strategic objectives. Setting baseline values pre-implementation and measuring changes will help assess ROI and refine chatbot/virtual assistant workflows for ongoing improvement.
Is there a clear ROI for chatbots?
For many organizations, the answer is “yes” — with qualifiers. Early adopters have often seen measurable labor cost savings amid staffing shortages. Every routine task offloaded to a chatbot is one less burden on staff — and over thousands of interactions, those savings add up.
For example, if a virtual assistant can handle after-hours calls instead of an answering service, that could replace a paid service or reduce on-call staff time. If it automates 50 rescheduling calls a week, that might save several hours of staff time, freeing up capacity for other needs without hiring additional FTEs.
On the revenue side, preventing no-shows and adding new bookings directly contributes to income. The Weill Cornell example (47% more digital bookings) likely translated into tangible revenue from visits that might not have been booked otherwise.
However, ROI is not necessarily uniform across all scenarios. Practice size and scale matter: a large health system handling millions of patient contacts may see a dramatic absolute ROI from automation. In contrast, a five-physician practice might see more modest gains. When evaluating vendors, ask if they offer scalable pricing (e.g., per patient or per visit), so smaller groups can still achieve net benefits.
Remember that some benefits are indirect or long term. Patient satisfaction, for example, can eventually drive revenue through improved retention or word-of-mouth referrals. Those are hard to measure quarter to quarter but are part of the strategic ROI.
Achieving ROI also isn’t automatic:
- Training and fine-tuning: AI chatbots need to deliver correct information — inaccurate or confusing responses can create more work and undermine patient trust.
- Data privacy and compliance: Because these tools handle private health data (PHI), practices should ensure vendors sign a business associate agreement (BAA) and adhere to HIPAA guidelines. Any breaches or lapses could cost far more than the savings.
- Ongoing oversight and maintenance: Someone should periodically review chatbot logs to ensure it functions well and automatically updates its knowledge base with changes (e.g., if the clinic’s hours change for a holiday, the bot needs to reflect that). These are relatively minor duties but need to be accounted for. Many vendors provide dashboards to monitor usage, success rates, and handoff frequency. These analytics can help a practice optimize the ROI (for example, if many patients ask a question the bot can’t answer, adding that answer increases automation rate).
Conclusion
AI-based chatbots and virtual assistants have evolved from experimental pilots to practical tools that enhance patient access and communication, increasingly customizable to a practice’s specialties and workflows.
Today’s chatbots can book flu shots, triage a rash, and answer patient questions 24/7, in multiple languages, all while integrating with core systems.
Looking ahead, the trend is toward even deeper integration and "smarter" AI. Future models may access to a patient’s health history (with consent) to deliver personalized advice: “You’re due for a blood pressure check, shall we schedule that?” As generative AI becomes more accurate and better trained in medical knowledge, their reliability will increase, likely expanding the scope of issues they can handle safely.
From an ROI perspective, more evidence is emerging as large health systems share outcomes from implementing digital front door. For smaller medical groups, the question may shift from if chatbots add value to how best to implement and integrate them most effectively. A small pediatric office may not need AI chatbots for every channel, but a deep integration to allow for scheduling vaccines and urgent sick visits could offer real value. An orthopedic practice might not be ready to use AI for appointment booking, but it might be OK automating post-surgery rehab check-ins.
While challenges such as ensuring accuracy and maintaining the human element remain, the trajectory is clear: AI assistants are becoming integral members of healthcare teams. Practice leaders can make a strong case that deploying a chatbot is not just a tech novelty, but a practical step to reduce no-shows, boost same-day scheduling, deflect calls, improve satisfaction, or even increase revenue — with a demonstrable ROI when aligned with the practice’s workflow.
