Syspro Torque’s IMTS Debut Centers Industrial AI on Approval, Rules, and Cost

Key Takeaways

⇨
Syspro's IMTS demos focused on how an AI agent's actions are approved, sourced, and costed.

⇨
Torque draws its rules from a manufacturer's own policy documents and cites them in each decision.

⇨
ERP teams can test the model on one rules-based workflow before widening agent autonomy.

Manufacturers evaluating industrial AI face a governance decision as much as a technology purchase. In a recap of its IMTS 2026 presence in Chicago, Syspro says the question booth visitors raised most often about its new Torque platform was who approves what an AI agent does.

Syspro launched Torque on Sept. 2 as a platform that detects operational problems, recommends a next step, and takes approved action inside the systems manufacturers already run. IMTS, held Sept. 14 to 19, was its first public showing. The platform is in a controlled availability program, with general availability to follow.

Automation World and Smart Industry placed the launch against labor shortages, tariffs, and supply chain uncertainty. Smart Industry also cited its own reporting that 70% to 85% of generative AI deployments miss their intended ROI. Against that backdrop, Syspro’s demos concentrated on how an agent’s decisions are checked, explained, and paid for.

Agents Check Orders Against the Manufacturer’s Own Rules

Syspro says the demo visitors most often asked to see again was its Sales Order Compliance Agent. The agent checks each order line against pricing policy, credit limits, shipping restrictions, and inventory availability, then returns a pass or fail. Each violation cites the rule that triggered it and attaches the source document. An operator then decides whether to release or escalate the order.

The rules come from the manufacturer’s own policies and terms. A separate Rules Auditor Agent answers plain-English policy questions with the relevant clause, any exceptions, the source document, and the date that document was last updated. Syspro notes that in many mid-sized firms these rules sit in spreadsheets or with a few experienced staff. Bringing them into the agent, in Syspro’s framing, gives each decision a documented basis that managers and auditors can review.

Creed Grimm, Syspro’s vice president of pre-sales, gave Smart Industry a traceability example: finding where a bad batch of material went and presenting next steps, a task that normally requires manual reports across several systems.

Approval and Cost Controls Set the Pace of Autonomy

In Torque, a person approves each action before it runs, according to Syspro. Approved actions then execute through the ERP’s standard integration layer, and each step is logged. Syspro calls this its glass house approach. Chris Lloyd, Syspro’s chief solutions and technology officer, told ERP Today that customers want early use cases to work with a human in the loop before they hand them off.

Manufacturers decide which decisions Torque only recommends and which it can act on. Grimm told Automation World that some customers want procurement insight served up while their buyers keep making the purchases.

Cost visibility follows the same pattern. Torque estimates each workflow’s cost before it runs and tracks what each agent delivers. Dan Abramson, Syspro’s senior vice president of Americas, told Automation World that showing who made a decision and what a workflow costs in tokens and credits is central to how Syspro differentiates the platform.

Torque runs inside the Syspro 8 Web UI and deploys API-first on other ERP systems and earlier Syspro versions. It connects to MES, SCADA, and warehouse systems through Model Context Protocol connectors, according to Syspro.

What This Means for ERP Insiders

Rule provenance is becoming an AI evaluation criterion. An agent that cites the policy clause and source document behind each decision gives managers and auditors something concrete to check. ERP teams can ask any vendor to demonstrate that trail using their own policies.

Start with frequent, rules-based workflows. Order compliance, credit checks, and supplier follow-ups have clear rules and measurable before-and-after results. They let teams build confidence in agent decisions before extending autonomy.

Document the rules before automating them. If policies live in spreadsheets or with a few experienced employees, an agent is only as consistent as that source. Consolidating them first gives any AI deployment a firmer foundation.