Product Manager (AI Agents)
Anywhere ¡ 88%Data & AIContractMid
About the role
Agentplace is a platform for building AI agents that serve customers: they answer questions, book appointments, handle customer service, and deliver services online. Based in San Francisco, featured in VentureBeat, USA Today, and Fast Company.
Weâre looking for a Product Manager to help shape Agentplace: decide what we build next, turn user problems into clear specs, and make sure what we ship actually improves activation and retention. This is a hands-on contract role. Youâll work directly with the founder, designers, and agent builders. Fractional and freelance PMs welcome: keep your other clients and work with us on a project basis.
What youâll do: talk to users and understand how they build, publish, and use agents; find where users get stuck (sign-up, first agent, publishing, upgrading); turn findings into prioritized problems and short specs; work with design and engineering to ship improvements and measure results; follow competitors and the AI agent market.
Who you are: youâve owned product decisions, you use AI tools every day, and youâd rather ship a small improvement this week than a big plan next quarter. You might be a fractional or freelance PM working with early-stage startups, a PM from an AI, no-code, or developer tools company, or a product-minded designer or founder. Weâre looking for 2+ years of product management, comfort with product analytics (Amplitude, Mixpanel, PostHog, or SQL), clear and short writing, and the ability to prototype yourself with tools like Cursor, Lovable, or v0.
Details: remote, contract, paid per project or milestone (hourly is also possible).
How to apply: complete a short test task (2-3 hours). 1) Sign up at agentplace.io, build a simple AI agent for any business, and publish it as public. Use the Luna model, itâs free. 2) Write a one-page product review: the 3 biggest problems you ran into as a new user, how youâd prioritize them and why, and a short spec for the top one with the one metric that shows it worked. Send the link to your agent and your one-pager to .
Applications are reviewed by email only. Applications without the completed test task will not be considered. We review applications quickly and will reach out to successful candidates with an offer.
Originally posted on Himalayas