• Run the AI operating system. Own the agent workflows that drive our product pipeline — from raw customer request through prioritisation, spec, engineering handoff and release. You will write the prompts, skills and scheduled jobs (cron) that do the heavy lifting, review their output with a sceptical eye (does this account for what we've already shipped?), decide what gets automated and what stays a human call, and improve the system every week
• Set the quality bar on UI design. Partner closely with design, critique flows with a genuine eye for interaction and visual detail, and refuse to ship experiences that are merely functional. You will sketch, wireframe or prototype to make ideas concrete, and you will be one of the people who says "this is not good enough yet" — and can say exactly why
• Get into the technical detail. Read the API, query the data, understand the data model, and reason about trade-offs with engineers rather than relaying messages between them and the business. Write specs that engineers do not have to reverse engineer
• Prioritise with revenue in mind. Shape the roadmap by weighing net-new revenue, churned and at-risk deals, support ticket volume, and logo and staff volume — split by region (AU, US, CA, UK, global) — alongside strategic fit and competitive position. You bring ranked options to the decision with sources, not vibes
• Market what you ship, in plain language. Write the blurb, the announcement and the help- article draft that turn a technical capability into a benefit a busy care provider immediately understands. You own adoption, not just delivery — shipped-but-not-adopted is your failure, not marketing's
• Own and groom the backlog. Keep a healthy, well-prioritised backlog in Linear with clear user stories, acceptance criteria and requirements, so the team always knows what is next and why
• Drive delivery. Run refinement, planning and reviews, keep work unblocked, and make scoping calls in the moment rather than escalating them
• Stay close to customers. Gather and synthesise feedback through support conversations, user testing and direct customer contact, so the team stays grounded in the real pain points of care providers and support workers
• Measure and learn. Track adoption, engagement, retention and rollout state (feature flags across regions), and use what you learn to decide what to build, improve or retire
• 
  

Requirements

AI-native operation (essential)

• You don't just use AI — you build with it. You can write a production-grade prompt, author a reusable agent skill, set up a scheduled job, and reason about what should be automated versus what must stay human
• You have a deep working knowledge of where product is headed with AI — agents, agentic software, AI-native workflows, and how the product craft changes when machines do the lifting. You have a point of view on this, and you will be asked to defend it
• You are healthily sceptical of AI output. You verify against ground truth, and you've caught a confident model being wrong before
• You can point to agent workflows or automations you designed and ran — at work or because you couldn't stop yourself
• 
  

Strong UI design opinions (essential)

• You notice when spacing, hierarchy, copy or interaction is wrong, and you can articulate why — not just "it feels off"
• You can sketch, wireframe or prototype an idea well enough to make it concrete. You can point to shipped work whose interface quality you were personally accountable for, and you're proud of it
• You give and receive design critique directly and well
• 
  

Technical Depth (essential)

• A genuinely technical background. This might be a computer science or engineering degree, time spent as a developer, QA engineer, data analyst, solutions engineer or technical BA, or a self-taught track record you can demonstrate
• Comfortable answering your own data questions without waiting on someone else
• Able to read and reason about APIs, integrations and system architecture at a level that earns credibility with engineers
• 
  

Autonomy And Range (essential)

• Proven ability to manage your own time across competing priorities without a delivery manager sequencing your work
• Experience running multiple concurrent projects or product areas, with evidence you kept all of them moving
• Bias to action. You unblock yourself, chase the answer, and close the loop
• 
  

Communication (essential)

• Outstanding written English. Our specs, tickets, release notes and customer communications are written, asynchronous and public inside the company, so precision and clarity are core to the job
• Able to hold your own verbally with customers, engineers and executives, and to translate confidently in both directions
• 
  

Note: all candidates complete a written exercise and a verbal discussion, assessed against the same standard. Expect to design a small agent workflow, critique an interface, and present your point of view on where AI is taking the product craft.

Experience

• Tenure is not what we are optimising for. Roughly 4 years owning product outcomes is a guide, not a gate — a demonstrated builder beats a long CV. We care about evidence: agent workflows you've built, interfaces you held the bar on, adoption you drove
• 4-6 years solid product management experience, ideally SaaS, excellent product judgment and discovery skills, strong track record shipping products customers use. On top of that, you are hands on with modern AI tools and have built your own agentic workflows. Show us what you've built. Curious, fast-moving, comfortable experimenting
• Exposure to B2B SaaS is valuable. Deep experience in Agile ceremony is not
• Bonus: care, health, disability or workforce management domain experience; hands-on experience with Figma, Linear, GitHub, LaunchDarkly or HubSpot. Tools: Linear, GitHub, HubSpot, LaunchDarkly, Notion, AI agents (Hermes, Claude Code), AI-native spec and analysis workflows
•