Part 1 of our 2026 hackathons covered our company-wide event on February 25th, where people across every function at Jane spent a day building with AI. Part 2 was an opportunity for our Product, Engineering and Design teams to go deeper, experimenting with what Jane AI has been building.

Jane AI is a startup team within Jane, building our AI infrastructure and new experiences across audio, mobile, and more.

Where Jane is in 2026

Over 65,000 clinics run their practices on Jane. Millions of patients book through us every month. We're profitable, growing fast, and every problem we solve lands directly in the hands of real practitioners and patients.

We're at an inflection point. We've been putting AI in the hands of every team, moving faster, and making some of the most significant technical bets in Jane's history, all in service of the practitioners and patients who rely on us every day.

Among the biggest of those bets is Jane AI, a dedicated team initiated by our Co-CEO Trevor, whose mission is not to replace Jane but to transform it, in the shared pursuit of helping the helpers. That means:

• Building a new AI-first product layer that's the best in our industry
• Creating the systems that let Jane extend that product with speed and confidence
• Driving the culture change that sets our product up to thrive in the age of AI

Our 2026 mantra says it best: Customer Obsession, Innovation, and Momentum.

The platform underneath

The Jane AI SDK is a Rails engine inside the Jane monolith. It provides agent session persistence, message storage, agent loop execution, streaming, and a REST API that product teams build against.

The SDK is built on Riffer, an open-source Ruby framework the Jane AI team built for agents. Riffer handles the LLM provider adapter layer, the agent model, tool invocation, and streaming. It is designed so the same agent code works across providers without rewriting. Our Rails engine sits on top of that with Jane-specific context, guardrails, and session infrastructure wired in.

When product teams experimented with the SDK, they were building agents and tools on infrastructure the Jane AI team had already built and shipped.

Privacy, Governance and Safety

Building AI in healthcare means the bar for trust is higher than in most industries. Similar to part 1, before the hackathon, Jane's internal AI working group, spanning Security, Privacy, Legal, People, and Product, reviewed every tool in play, ran education sessions across the team, and set clear boundaries on data handling and escalation. We use RunLayer for MCP governance so integrations get security reviewed before they go live.

A fraction of what we built…

Voice for any Riffer agent

One engineer extended Riffer itself, adding a unified voice layer so any Riffer agent can become a real-time voice experience with minimal additional setup. Building voice today means stitching together separate speech-to-text, text-to-speech, and streaming providers with their own failure modes. The goal was one clean adapter model, the same way Riffer already works across LLM providers.

Get a new clinic live in minutes, not days

The Onboarding team coordinated three agents working together: a web extractor that pulls data from a new clinic's existing website to pre-fill their Jane setup, a setup agent that closes remaining gaps through conversation, and a learning agent that generates a personalized path based on what the clinic actually needs. New clinics today face a setup process that can sometimes feel daunting. All three agents ran together inside a single hackathon day.

Intake forms in seconds

The Clinical Forms team built a Smart Form Builder: an agent that collects context about a practice's specialty, jurisdiction, and patient population, then generates a ready-to-use intake form inside Jane. What currently takes practitioners hours of manual configuration becomes a conversation.

Tackling Complex insurance

The Insurance teams went somewhere technically hard. Eli 1.0 surfaces a patient's insurance eligibility automatically before an appointment, so practitioners stop walking in blind. A second project took on a bigger open question: outside of eClaims, there is no standardized claims path in Canada. Practitioners are manually logging into separate insurer portals to re-enter information that already lives in Jane. The hackathon goal was to find out whether an AI agent could navigate one of those external portals reliably and where the real hard problems are. They came back with clear answers and a defined target for May.

Patient context before the appointment starts

The Charting team built Facesheet, a dynamic pre-appointment dashboard surfacing patient vitals, history, and relevant context at a glance. A practitioner can also chat with an agent to surface additional information or modify the view based on their discipline and the appointment at hand. Today that picture requires digging manually through historical chart data.

Business intelligence for clinic owners

Jane Pulse gives clinic owners a plain-language view of how their business is doing: revenue patterns, scheduling gaps, patient retention signals. Built as a standalone agent-powered web experience on the SDK, and scoped so it can evolve further over time.

The hackathon was one week of a much bigger build

The projects above show what product teams can do when the infrastructure is already there. The Jane AI team is simultaneously shipping products and infrastructure that will define how practitioners interact with Jane going forward.

We're building new ways for practitioners to access Jane; including a mobile experience designed to support a clinician's day-to-day workflow by surfacing the right information at the right time.

We're also exploring new surfaces where AI capabilities come together for users, including natural language interfaces and voice. The goal is for AI to show up wherever it's useful across Jane, not just in one feature. Several projects from this hackathon are now in active development.

If this sounds like your kind of problem

We are building AI that impacts real lives: the helpers running businesses and providing care, the patients getting care, the family members, the community. Our work has ripple effects, and if we can keep improving how this ecosystem operates, we are going to make everybody's life better. The problems are harder here because the stakes are higher. That is also what makes getting it right worth doing.

We are profitable, growing, and moving with urgency because the window to define what AI in practice management looks like is open right now.

If you are excited about helping the helpers, building AI products that practitioners and patients actually rely on, and working on a platform that is growing in real time, we want to hear from you. We are looking for builders who have shipped something from scratch, a product, a company, a side project, that real people used.

Take a look at our open positions at jane.app/careers.

If Jane AI specifically is where you want to build, you can learn more about the team and what we are working on here.