Principal Product Manager | Prism Workflow Engine

Every product built on Prism eventually needs to do real, multi-step work against the ERP, not just retrieve information. The Workflow Engine is the shared foundation that makes that a customer’s quote to cash work via AI systems reliable, reversible, auditable, and reusable. You own the product definition of that foundation: how a unit of work is represented and tracked from start to finish, how information is handed off cleanly between different parts of the system, how much autonomy an AI agent gets based on how reversible its action is, and the approval, policy, and audit systems that make the engine safe to point at real production data.

Epicor is a technology leader in ERP software for manufacturing, distribution, and retail, with deep roots in both enterprise software and startup culture. Our product organization is in the middle of a deliberate move to a platform-first, AI-enabled delivery model: building shared capabilities that many teams can build on, broadening how roles work together, and removing the silos and handoffs that slow good teams down.

The biggest change versus a traditional PM job is how the work gets done. We expect this leader to operate fluently in the model below, and to help the rest of the org adopt it.

• Built as a platform, not a feature. The Workflow Engine is built for the other product teams, consulting teams, and customers that will build on it. Ease of adoption, reuse of shared building blocks, and a clear path for a new team to get their first workflow running are treated as first-class goals, not afterthoughts. Success is measured by how much operational work in the ERP is done via the workflow engine vs. manually.

• A small, senior team, not a feature factory. Delivery happens in a lean, senior, cross-functional team working closely together against a system with built-in human oversight. You are inside that loop as the product decision-maker, not reviewing status from the outside.

• Roles blur on a team like this. The lines between product, engineering, design, and architecture blur on a team this size. You work directly in specs, cross-team sequencing decisions, and pilot results, not a backlog handed over a wall. You keep clear ownership of product scope and decisions while picking up adjacent work that a small team requires.

• Responsibility is the differentiator. How reversible an action is (fully reversible, reversible within a limited window, or effectively permanent) determines how much autonomy an AI agent gets to take it. Policy, approvals, and a clear audit trail are product features you own, not compliance add-ons, because they're what let a customer trust the system with real business decisions.

• Share openly. Open questions, pilot results, and test outcomes are shared across the org so people can see clearly what the engine can and can't yet promise.

• Empathy for two different customers. Real understanding of the end customer whose order, supplier requalification, or month-end close runs through the engine, and of the engineering teams deciding whether to build on it or build around it. You validate both before committing to a build.

• A platform mindset. You think about ease of adoption, reusable building blocks, and how many teams choose to build on your system the way a good platform product manager thinks about retention, not the way a feature team thinks about a single roadmap.

• Working Backwards: Demonstrates mastery of customer empathy, working backwards from customer pain points to validate the customer problems, ideal customer profiles, and value propositions.

• Founder Mode' Mindset: Demonstrates an entrepreneurial spirit by taking full ownership, embracing creative problem-solving, and relentlessly challenging the status quo to drive innovation and continuous improvement.

• Prototyping Pro: Quickly generates prototypes to get early feedback from internal stakeholders and customers

• Strategic Direction & Prioritization: Lead the orchestration of strategic direction and set priorities for our AI applications and functionalities. Define the vision and roadmap that drive innovation in AI-driven products.

• Own the product vision and build sequence. Decide what ships first as you take the Workflow Engine from its initial build to market, and keep the roadmap tied to real demand from the teams that want to build on it, not to technical completeness for its own sake.

• Define how work is represented and handed off. Own the core concepts that make the system work: how a unit of work is tracked from start to finish, and how information is passed cleanly between different parts of the system so nothing gets lost along the way. These are the two things every team building on the engine depends on.

• Own the full system, not one piece of it. Be the single product owner across all of the engine's functional areas: authoring (the tools people use to build and edit a workflow), execution (running workflows and connecting them to the ERP), quality (automated testing, simulation, and health monitoring), and governance (policy, approvals, and audit trails). Make the sequencing and trade-off calls that span all of these.

• Own risk classification and approvals as a product, not a policy document. Own the system that classifies how reversible a given action is, and the framework that decides who needs to approve what, based on things like reporting structure, dollar amount, and risk level, and make sure that logic is built into the system rather than handled manually.

• Resolve open questions with the right partners. Drive resolution of open questions the technical design has already surfaced, in partnership with the right team each time: which AI agents should be allowed to take which actions (with the identity and security teams), requiring automated testing before anything reaches production, how the system should behave when a multi-step process needs to roll back partway through (with the team building Epicor's agent development platform), how far a manual “stop” should reach in an emergency, and what spending limits should exist per customer (with sales operations).

• Drive adoption across other product teams. Partner with the other teams building on Prism so they choose to build on the Workflow Engine instead of building their own version. Own the onboarding experience for a team running its first workflow on the system.

• Own the outcome-based pricing model. Own how usage is tracked and measured, and how the engine supports a shift toward pricing based on outcomes and usage rather than per-seat licensing. Build the financial case for this, and be explicit about how it affects existing pricing.

• Lead pilots with real customers. Lead the customer pilots that prove the system works on real workflows, including what happens when something goes wrong, not just the ideal case, in partnership with the professional services team, before any broader rollout.

• Stay current and avoid duplicated effort. Stay current on how other companies are approaching AI agent orchestration and workflow automation, and fold what's useful into the roadmap so the system keeps improving instead of other teams quietly building duplicate versions of the same thing.

• Prototype. Create hands on working prototypes of product concepts in addition to documentation and design.

• Strong product management fundamentals. Proven ability to validate customer needs, prioritize ruthlessly, and build a roadmap for a complex product, ideally one that other teams build on top of rather than a single customer-facing feature.

• Real technical depth in distributed systems. Working understanding of how multi-step, distributed processes are typically designed, including patterns for handling partial failure and rollback, sufficient to make real product trade-offs with an architect and a senior engineer rather than just relaying their recommendations.

• Applied AI knowledge. Strong working knowledge of AI agents, how they retrieve and use information, and how to evaluate whether they're working correctly, enough to make real architecture and trade-off decisions, not just sponsor them.

• Comfort with a fast-moving, hands-on way of working. Comfort defining clear specs and success criteria, and directing a small, cross-functional team without ambiguity about who owns what decision.

• Understanding of governance and trust. Understanding of how risk classification, approvals, audit trails, and emergency controls translate into customer trust, since that trust is the product's core value.

• Analytical and financial fluency. Data-driven decision-making, and the financial fluency to build a usage- and outcome-based pricing case and a rigorous return-on-investment model.

• Leadership without formal authority. Ability to drive alignment across peer leads in architecture, design, engineering, and quality, and across adjacent teams like security, sales operations, and professional services, without having formal authority over any of them, and to explain a clear, credible product vision to each audience.

• Familiarity with modern software delivery. Familiarity with modern, AI-assisted software delivery and cloud/SaaS platforms. Previous ERP experience is a plus.

• Experience: 8+ years in product management, with a track record of shipping platform or infrastructure-style products that other engineering teams build on top of.

• Relevant technical depth: 3+ years of experience with workflow orchestration, distributed systems, or applied AI/agent products, with real technical depth rather than purely commercial exposure.

• Education: Bachelor’s degree in Computer Science, Engineering, Business, or a related field (or equivalent experience). Advanced degrees are a plus.

• Workflow orchestration experience. Direct experience with a workflow orchestration platform (for example, Temporal, Camunda, or AWS Step Functions, or something comparable) is a strong plus.

• Hands-on AI experience. Practical experience with prompt engineering, retrieval-based AI systems, evaluation methods, and agent orchestration, ideally including using AI coding tools to prototype directly.

• Programming ability. Working proficiency in a language such as Python or C#, enough to read the system's code, prototype against it, and work credibly with engineers.

• Data and integration background. Familiarity with data pipeline concepts, knowledge graphs, and how modern AI systems connect to external tools and data.

• Machine learning exposure. Exposure to frameworks such as TensorFlow, PyTorch, or scikit-learn is a plus.

This role is for someone who wants to take a product from its early build to market, not inherit a finished roadmap: owning how work is represented, tracked, and governed as it moves through the system, and making the Workflow Engine the thing every other product team and customer loves to build on instead of rebuilding for itself.

#LI-MB2 #LI-HYBRID

Whatever your career journey, we’ll help you find the right path. Through our training courses, mentorship, and continuous support, you’ll get everything you need to thrive. At Epicor, your success is our success. And that success really matters, because we’re the essential partners for the world’s most essential businesses—the hardworking companies who make, move, and sell the things the world needs.

Range:

The salary range provided reflects the national average for this job title and does not represent compensation specific to Epicor Software Corporation. Actual compensation will vary based on experience, qualifications, and market factors relevant to the position.

Recruiter: