Vaibhav Vohra will become Epicor’s CEO on October 1, 2026, inheriting a business that has expanded its cloud ERP strategy and is embedding AI into its operational workflows. For manufacturers and distributors, the real question is what those capabilities improve and how much authority customers should give them.

Vohra succeeds Steve Murphy, who will step down on September 30 and remain on the board. Epicor says it surpassed $1 billion in annual recurring revenue under Murphy’s leadership. The company’s succession announcement describes a planned transition following Vohra’s appointment as president in 2024.

I recently sat down with Vohra on KramerTALK to discuss that transition and where he believes ERP is headed. His vision builds on industry knowledge, operational data and AI-assisted action. The test is whether those elements help businesses respond faster while preserving control over the decisions that follow.

ERP already supports planning, business rules and automated workflows. AI can extend that role by helping people interpret changing conditions and determine what to do next. But trusted data and existing access permissions are not enough on their own to determine whether an agent should be allowed to act. Disclosure: KramerERP provides paid research, advisory and consulting services to technology companies, including ERP and data vendors listed in this article.

A Product Leader Steps Into The CEO Role

Vohra joined Epicor as chief product officer in 2021, became chief product and technology officer in 2023 and was named president in 2024. He began his career designing satellite hardware and software, where he also had responsibility for ERP, before moving into roles at SAP and robotics companies including Gecko Robotics. That background gives him a perspective spanning enterprise software, AI and automation in physical environments.

Vohra describes Epicor’s evolution in three phases. Epicor V1 was shaped by acquisitions, on-premises software and industry specialization. V2 was the cloud chapter under Murphy. V3 is the next phase, combining industry knowledge, customer processes and community insight with AI, faster deployment and agentic workflows. He also sees Epicor’s sub-industry customer communities as a source of shared learning that can help shape roadmap priorities and reusable workflows.

What stood out was how Vohra described Epicor’s foundation. “I use the word trust because the legacy is really the trust of dealing with customer data,” he said.

That history has value. Transactions, security models and business rules preserve important knowledge about how a company operates. They can also carry outdated assumptions, inconsistent records or permissions that need review. Making that foundation reliable for AI means confirming the information is current, the rules still apply and the proposed action is appropriate.

From Vision To Available Capabilities

Epicor has already announced capabilities that support parts of this direction. In May, it introduced Prism Agent Foundry and Lux , its agentic design system, as part of a broader stack for building and governing agents inside ERP workflows. In an August 2026 Prism announcement , the company described conversational access to live ERP data and documents in Kinetic and Prophet 21, along with preapproved actions subject to human oversight. Epicor also says Prism respects existing governance and role-based security.

Those capabilities support parts of the V3 direction, but they do not mean every customer can achieve that vision today. Buyers still need to confirm what is available in their product, deployment and region, what configuration is required and what results have been demonstrated in comparable operations.

As I have written about ERP’s expanding execution role , the opportunity reaches beyond any single application. Manufacturing and distribution processes span ERP, supply chain applications, warehouse systems, machines, commerce and partner platforms. Useful AI has to work with that broader operating environment.

AI Still Needs Someone Accountable

The harder question starts when AI moves from recommending an action to taking one. An agent might identify a supplier delay and suggest expediting an order, moving inventory or changing a production schedule. Who determines whether it has authority to act, which information it should trust and who owns the outcome?

“The decision and the outcome should rest in the hands of people,” Vohra said. He believes agents should operate within the responsibilities and access established in ERP rather than bypass controls because the interface has changed.

Vohra shared a story from his robotics experience to illustrate the risk. He described a robot that mistook a visual display for an empty shelf, causing an automated replenishment system to order unnecessary inventory.

The lesson is what happens when a bad signal moves directly into execution. Before an agent acts, the system needs to validate the signal against inventory and business rules, enforce limits on the action and escalate exceptions. Existing access rights are a starting point, but they do not answer every question about what an agent should be allowed to do.

That distinction sits at the center of the decision-rights issue I have explored. For buyers, three questions make the discussion concrete.

  • What can the agent execute, and what financial or operational limits apply?
  • When must a person approve or intervene, and can the business trace and correct the action?
  • What evidence will show that the workflow improved without increasing errors, risk or rework?

Time Saved Is A Starting Point

Vohra offered his view on measuring AI. “Did we save someone time? That’s how we know the future is successful,” he said.

That could mean reducing time spent finding information, preparing a procurement decision or handling routine supplier interactions. Those gains matter, particularly where teams are stretched and experienced employees hold operational knowledge others still need.

Time saved should be paired with quality. A faster purchasing decision has limited value if it creates excess inventory or more work correcting errors. The better measure compares the time required, exception rate and a relevant business outcome before and after the change.

The workforce implications matter too. As I have argued about the enterprise operating model , technology may drive transformation, but people drive technology. Removing repetitive work creates value when employees have the knowledge, authority and support to manage exceptions and make better decisions.

Resilience Requires Visibility And Context

When I asked what manufacturers and distributors should notice three years from now if this next phase works, Vohra started with resilience. The value is in recognizing disruptions earlier and evaluating options faster, whether that means changing a supplier, reallocating inventory or adjusting production. Vohra also pointed to the need for visibility beyond direct suppliers, where risk can sit deeper in the network.

Modernization speed matters as well. Vohra favors getting a core environment running sooner and building from there rather than optimizing everything before go-live. Getting the core live faster does not mean the transformation is finished. Scope, integrations, data preparation and organizational readiness still determine what can move safely.

The Test For Epicor’s Next Chapter

Toward the end of our conversation, Vohra made what I thought was his most interesting observation. “AI has been looking for a vehicle and I think that vehicle is ERP,” he said.

There is logic behind that perspective. ERP connects transactions to business processes, and industry context helps explain which actions make sense. A fabricated metals company and an HVAC distributor may use similar categories of data while operating under very different constraints.

I would add one qualification. ERP cannot become another closed layer. Customers will continue to use multiple applications, data platforms and AI technologies. ERP does not need to own everything. It needs to become part of the decision architecture that connects operational context to action across those environments.

Murphy helped move Epicor through its cloud chapter. The next test is turning that foundation into measurable execution. For customers, the evidence should show up in shorter response times to disruptions, fewer avoidable exceptions and better inventory performance. Those gains still need to be weighed against implementation effort and ongoing oversight.

A system of action should ultimately deliver better operational decisions with clear accountability for what happens next.