Senior Product Manager — Workflow Builder Runtime & Extensibility; Agentic, AI-First Enterprise

About Adobe

Adobe’s mission is to change the world through digital experiences. We empower creators and global brands with the tools they need to design and deliver exceptional digital content across every screen. Adobe is committed to an inclusive workplace where everyone is respected, supported, and empowered to do the best work of their career.

We are building Workflow Builder, an AI‑first workflow runtime that orchestrates production workflows across services, tools, and systems. As a Senior Product Manager, you will own the runtime and extensibility foundation end‑to‑end, defining the execution model, developer surface area, governance guardrails, and roadmap that unlocks real production adoption at enterprise scale.

• Building platforms that enable other builders—creative technologists, solution teams, partners, customers, and developers—to move faster.
• Designing primitives and contracts that scale across many workflows and use cases.
• Crafting a developer experience teams choose because it’s clear, composable, and reliable.
• Owning hard problems in multi‑tenant runtime execution: isolation, performance, observability, governance, and lifecycle management.
• Turning power‑user needs into a system that is approachable and production‑ready, especially when workflows include AI/agentic steps.

• Own the product strategy and roadmap for the Workflow Builder runtime: orchestration, execution semantics, state, scaling, reliability, and operability.
• Own extensibility end‑to‑end: define how customers build, test, deploy, version, govern, and operate custom actions, including packaging & distribution, versioning, compatibility, migration, authentication/authorization, data access boundaries, secrets handling, tenant isolation, policy enforcement, approvals, governance controls, runtime limits, performance, cost controls, and observability.
• Define platform contracts: workflow definitions, triggers, connectors, action schemas, retries, idempotency, error handling, and SLO‑backed reliability.
• Define and ship agentic runtime primitives: structured state/context, tool/action invocation, planning + execution loops, bounded autonomy, deterministic fallbacks, and human‑in‑the‑loop checkpoints.
• Build the trust layer for AI‑first execution: permissions, policy enforcement, audit trails, evaluation hooks, and enterprise controls over agentic steps.
• Partner deeply with engineering, security, and platform teams to ensure the runtime is enterprise‑grade: compliance, auditing, tenancy, and secure‑by‑default execution.
• Drive discovery with enterprise adopters and internal solution teams to validate requirements and support real production workloads.
• Lead cross‑functional execution, aligning stakeholders, clarifying trade‑offs, and shipping iteratively with measurable adoption.

• Safe, multi‑tenant execution of customer‑built actions (isolation, quotas, data boundaries).
• Deterministic, reproducible workflow behavior across versions.
• Governance that scales (policy, approvals, auditability) without killing builder velocity.
• First‑class observability across workflow and action boundaries.
• Reliable execution under production load (timeouts, retries, back pressure, failure modes).
• Making agentic behavior enterprise‑trustworthy: bounded autonomy, predictable outcomes, and auditable decisions.
• Operating AI‑first workflows at scale: latency/cost controls, fallback strategies, and incident‑ready visibility across model and action boundaries.

• A production‑ready extensibility framework where customers can ship custom actions with clear lifecycle management and governance.
• Defined execution semantics and reliability targets that unlock customer confidence (SLOs, incident playbooks, operational maturity).
• A developer experience that materially reduces time‑to‑integrate and time‑to‑debug (SDK/API, testing/sandboxing, great docs).
• Early enterprise reference workloads running end‑to‑end with measurable adoption and reusability.
• A clear, shippable foundation for AI‑first workflows: policy + evaluation hooks + human‑in‑the‑loop patterns…