Puppeteer AI is a clear example of the "Vertical Agent" trend in the AI ecosystem. They are moving past the chat window by building agents that possess agency within a specific system of record—in this case, the EHR. Their relevance to the agent stack is primarily in their integration layer and their domain-specific guardrails, demonstrating how agents can be deployed in highly regulated environments like medicine.
For builders, Puppeteer AI serves as a blueprint for how to handle the "last mile" of agent deployment: connecting a reasoning engine to a proprietary database while maintaining safety. They are champions of the idea that agents shouldn't just talk to users, but should perform transactional tasks that traditionally required human oversight, such as clinical triaging and automated rescheduling.
Most conversational AI companies attempt to be everything to everyone. Puppeteer AI takes the opposite approach, focusing exclusively on the healthcare sector. Launched in late 2023, the company builds voice and text agents designed to handle the administrative and clinical tasks that typically require a human staff member: scheduling, intake, and triage. The core of their technical proposition is not just the ability to speak or text in a human-like manner, but the depth of their integration into the healthcare data stack.
They have prioritized building connectors for major Electronic Health Record (EHR) systems, including Epic, Cerner, and Athenahealth. This allows their agents to act as more than just a front-end interface. When a patient calls to reschedule, the agent isn't just taking a message; it is checking the provider's live calendar through an API, identifying an open slot, and updating the system in real-time. This "closed-loop" automation is what differentiates an actual agent from a basic chatbot.
Working in healthcare introduces a set of constraints that would kill a general-purpose AI startup. Puppeteer AI addresses these through a combination of regulatory compliance and safety guardrails. Beyond being HIPAA-compliant in the US and GDPR-compliant in the EU, the company has built specific detectors into their platform to handle high-risk patient interactions. Their agents are programmed to recognize signs of self-harm, substance abuse, or medical emergencies, triggering immediate escalations to human care teams or emergency services.
This safety layer is paired with a transparent dashboard that allows providers to trace the decision-making process of the AI. For an industry that is notoriously risk-averse, this observability is a requirement for adoption. The company claims a 30-day go-live timeline, a remarkably fast cycle for healthcare IT, which often measures implementation in months or years. Their case studies suggest tangible outcomes, such as a 74% appointment confirmation rate for clients like Swing Care and a 36% increase in waitlist utilization.
Puppeteer AI operates on a SaaS-plus-usage model. Their "Business" tier starts at $1,350 per month, plus $0.20 per minute of voice and $0.05 per text message. This pricing structure suggests they are targeting mid-to-large clinics and digital health startups that have sufficient volume to justify the base fee but need a variable cost structure to scale. Interestingly, they have also begun experimenting with outcome-based pricing for specific use cases, a move that aligns their incentives directly with the clinic's revenue.
The company secured pre-seed funding in June 2024 led by Haatch Fund. While they are small, their list of integrated partners—which includes modern platforms like Healthie and Medplum alongside legacy giants—suggests they are successfully navigating the fragmented healthcare interoperability sector. As the cost of voice synthesis and LLM inference continues to drop, Puppeteer AI is betting that the real value lies in the vertical-specific logic and the secure data plumbing that connects an LLM to a patient record.
HIPAA-compliant AI voice and text agents for clinical intake, scheduling, and triage.
Puppeteer AI is hiring