Head of Product, Enterprise Platform

At Productboard, we're building the first agentic product management system. We're rethinking how product management is done in a world full of AI agents, and how we help our customers build what matters for their audience. AI is changing what an interface even is: static screens are giving way to conversations, generated artifacts, rich interactive UI, and open-ended workflows, often on the same screen.

1. The core data model and data services. The objects every feature and agent works with: products and features, goals, customers and segments, feedback, and how they relate. You decide what an object is, how objects connect, what an agent may write, how data is stored, queried, retained and exported, and what a customer can control about what an agent sees.
2. Views, query and reporting engine. One query layer over the data model: filtering, grouping, aggregation, computed fields and saved views, served the same way to people, agents and the public API. Roadmaps, grids and dashboards are built on it by other teams.
3. Scoring and planning primitives. Configurable scoring models, formulas and weights.
4. The document editor. Where specs, notes and customer feedback live, for people and agents alike.
5. Permissions, authorization, sharing and administration. Roles and access control for people and for agents acting on their behalf, sharing with people outside the workspace, and the admin controls a buyer checks before signing.
6. Security and compliance. The product side of security questionnaires, residency, retention, audit logs, and the record of what every agent did, and for whom.
7. Search. Finding anything in a workspace, whether a person or an agent is asking.
8. Notifications and alerting. Routing, batching, subscriptions and delivery across in-app, email and the channels teams already live in, so agent generated activity reaches the right people without becoming noise. The notification experience sits with your AI Product peer.
9. Explainable and provable agent work. Customers will not hand an agent their roadmap on faith. The evidence is yours: where an answer came from, which sources it used, why it acted, what changed as a result, and how a team measures whether the agent is getting better. Evals and quality signals ship as product surface. The eval suite itself sits with your AI Product peer.
10. The enterprise ecosystem. Two-way integrations with the systems an enterprise already runs: CRM such as Salesforce and HubSpot, engineering tools such as Jira and Linear, feedback sources, data warehouses, and the public APIs partners build on. Field mappings, sync rules, auth scopes, rate limits and each customer's own schema are complexity you absorb, so an agent reads a customer record or writes a ticket and nothing more. The protocol layer that lets external agents call Productboard sits here; the specific tools exposed through it sit with your AI Product peer.

1. The data model is written down once: what every entity means, how entities relate, what an agent may write. AI Product builds on that document.
2. The enterprise controls run as platform services: who may see what, where data lives and for how long, what a plan allows, what every agent did. An agent acting for a person sees and changes exactly what that person could.
3. Every part of the platform has a product manager, and the seats still open when you arrived are filled by people you hired. Every part also has a decision on its future, stay, merge or retire, and the roadmap reflects it.
4. The top three capabilities AI Product asked for have shipped as platform capabilities, each used by at least two surfaces.
5. The CRM and engineering tool integrations have an owner and a roadmap.

1. You have built a platform other products stand on, and hidden its complexity from them. You have owned a data model, a permission system, a compliance surface or an API that other teams built against without having to understand it, and you know the cost of getting the abstraction wrong.
2. You have a computer science or engineering background. You own the data model, the permission system and the API, and those conversations go faster when you can read an entity diagram, a query plan and an API spec yourself and argue with an engineer about them.
3. You build with agents every day. You use Claude or its equivalent for real work, connected to your own tools and data. You know what an agent needs from a platform because you have felt what happens when it is missing.
4. You are fluent in headless. The platform will be used through agents, open agent protocols and other tools. You can decide which capabilities need a screen and which need an API.