Agentic Resource Planning (ARP)
Agentic Resource Planning (ARP) is an operating model where AI agents, working inside a company's existing systems, draft resource decisions end to end (production plans, purchasing, inventory moves) while named humans approve what matters. Where ERP records what happened, ARP drafts what should happen next.
What is Agentic Resource Planning (ARP)?
ERP systems became the system of record: they store orders, inventory, and transactions, and people plan around them. Agentic Resource Planning is the step after that. AI agents, sometimes called AI employees, read the record, carry the operating rules, and draft the decisions themselves: tomorrow's production plan, the purchase that covers it, the inventory move it requires. A named human approves before anything executes. The system of record stays exactly where it is; the planning work moves to the agent.
How is ARP different from ERP and APS?
An ERP tells you what you have and what happened; a person still turns that into a plan. An APS can optimize a plan when the data is clean and every rule is encoded as structured master data. ARP differs on where the rules come from and who does the work: the agent learns undocumented rules from people and messy systems, drafts finished decisions, shows its reasoning in plain language, and leaves sign-off with a named human. It is a division of labor, not a new database.
How Semia relates to ARP
Semia builds toward this model in food and beverage manufacturing: an AI employee that reads the plant's existing systems, drafts tomorrow's production plan, and asks a named human to approve it. Production planning is the first ARP workflow; purchasing, inventory, and quality follow the same pattern. The original essay is at Agentic Resource Planning: the next letter after ERP.