Senior Product Manager (AI)

BukuWarung serves millions of merchants across Indonesia. Every merchant interaction —

sales, onboarding, support, collections, servicing — has a lot of human interactions. This role

owns the mandate to change that: use AI, including our internal agentic platform, to make our

customer-facing functions measurably cheaper to run and measurably delighting for the

merchant, at the same time.

This is not a platform-engineering role and not a research role. It's a business-outcomes role

that happens to use AI as its primary lever. You will spend more time identifying where an

agentic workflow changes a function's cost or quality curve, writing the spec, and partnering with

engineering to ship it, than you will spend configuring the tooling yourself — though you should

be comfortable doing the latter when it's the fastest path to a working prototype.

What you'll do

• Own the roadmap for AI-driven initiatives across all merchant facing functions, prioritized by a combination of cost/efficiency and speed/quality of the customer-facing experience.

• Identify which workflows in these functions are good candidates for agentic automation vs. augmentation vs. no automation at all, and make that call explicitly rather than defaulting to "automate everything."

• Write specs precise enough for engineering to build against, and partner with them

  through build, eval, and rollout, including where the workflow runs on our internal AI platform.

• Stand up the measurement for each initiative before it ships: what cost line moves, what quality/SLA metric moves, and by how much, not just that "AI was used.

• Manage and grow 1–2 PMs, each likely embedded closer to a specific function, while

  you hold the cross-functional prioritization view.

• Work directly with function leads as the primary stakeholders, you are translating their operational pain into a shippable product, not building in isolation.

• Know when an agent should not make the call, flag failure modes, escalation paths, and human-in-the-loop requirements before they become incidents, not after.

What we're looking for

• 6–8 years of PM experience, with hands-on AI/agentic product experience as a must-have — you've shipped something where an agent or LLM-driven workflow made a real decision or took a real action, not just called an API and rendered a response.

• Demonstrated ability to connect a product initiative to a P&L or ops line — you think in cost-per-ticket, cost-per-transaction, resolution time, conversion, not just feature velocity.

• Comfort managing PMs and holding a cross-functional roadmap, even if this is your first time doing so formally.

• Enough technical fluency to partner credibly with engineering on system design trade-offs for agentic workflows — data contracts, failure/fallback behavior, evaluation, guardrails — without needing to write the code yourself.

• A bias toward measurable outcomes over shipped features: you'd rather ship one workflow that provably cuts cost-per-ticket 20% than five that "used AI."
  

What will make you a strong fit

• Prior experience in a company where you personally had to build the case for automating a function you didn't own the P&L for — i.e., you've had to sell ops/sales/support leadership on a change, not just build it.

• Experience in emerging markets, fintech, or high-volume/low-ARPU businesses, where

  unit economics discipline matters more than it does at scale-flush companies.

• A specific, defensible point of view on where AI agents should not be trusted with a decision, this is a role that will need to say no to automation as often as it says yes.