Pamarco accelerates finance operations with autonomous agents and Copilot Cowork

Global manufacturing depends on lean operations

Pamarco supplies precision-engineered anilox rolls, blades, accessories, and technical services to printing and packaging businesses. Its products support corrugated packaging, flexible packaging, coating and laminating, and towel and tissue production. With approximately 225 employees, operations in the United States and Europe, and customers in almost 70 countries, the company combines a global reach with a deliberately lean operating model.

"We're a decentralized business operating across many different platforms," says John Burgess, Pamarco's President, Flexo Division. "But for the last 18 years, Microsoft ERP products are the one common thread." The shared system based on Dynamics 365 Finance and Dynamics 365 Supply Chain Management gives Pamarco and its seven legal entities a consistent source for financial, operational, production, and sales data. It also serves as the foundation for the company's next priority: using AI to reduce repetitive work without adding complexity or weakening control.

When Pamarco's board and chief executive asked the technology team to explore AI, Chief Information Officer Glyn Wilson focused on finance as the first objective. Finance was a logical starting point because it was closely connected to the data, its workflows were well understood, and it included repeatable tasks that could deliver measurable business value. Pamarco chose Microsoft Copilot Studio because the company already relied on Microsoft technology and wanted agents that could work with its established Dynamics 365 environment.

Wilson highlights Microsoft's Model Context Protocol (MCP) integration layer, which lets agents interact with business applications without requiring bespoke integrations. The Dynamics 365 ERP MCP Server gives bounded agents, limited to a defined scope, access to the business data and actions they need, while allowing Pamarco to keep the work inside its existing Microsoft environment.

A longer close limits business insight

Pamarco's finance organization is small and decentralized. Accountants and operational leaders often carry responsibilities beyond traditional finance work, so urgent requests can pull attention away from the monthly close. Burgess explains that as the business grows, so does the complexity of the monthly financial close.

The delay affected more than reporting speed. Finance professionals had less time to compare current results with prior months and years, investigate changes, and proactively influence outcomes. Routine work such as invoice coding, close-task coordination, report assembly, and ad hoc analysis competed for the same limited capacity. 

Bounded agents take on repeatable finance work

Pamarco is building autonomous finance workflows with Microsoft Copilot Studio, the Dynamics 365 ERP MCP Server, and Microsoft Dataverse skills. Rather than giving every user unrestricted access to a general-purpose assistant, the company defines agents and skills for specific jobs. Each scenario needs a clearly defined workload, appropriate permissions, and measurable time savings.

The financial close workflow uses the financial period close workspace in Dynamics 365 as its operating framework. Tasks have owners, due dates, and dependencies. An hourly check identifies tasks that are due and whose dependencies are complete. The appropriate agent performs the work, marks the task as complete, and emails the results. Human owners remain responsible for work that requires judgment or review.  

Invoice coding is another early scenario. The agent reads invoice information, connects it with Dynamics 365 data, and applies the appropriate coding. When Pamarco acquired a business in Germany, it needed its US team to code invoices there. To overcome the language barrier, Wilson developed a set of skills that can code German invoices and send a summary email to the finance team explaining how they were coded. During validation, the process demonstrated a high success rate, with fewer than 5% of invoices requiring manual coding. Not only is the risk of late vendor payments reduced, but the company estimates that a fully adopted process could reduce a one-hour daily task to approximately five minutes.

The majority of Pamarco's non-purchase-order invoices are freight-related and have complex tax code requirements. For example, multiple tax codes may apply to the same invoice. The AI workflow splits the journal into different tax lines and coding lines and "it's doing such a good job for us. It's around a 95% success rate. We are so confident with the AI freight coding that we have turned on auto-posting for all non-purchase-order invoices," says Wilson. 

Cowork turns recurring work into actionable insights

Pamarco also uses Copilot Cowork for tightly controlled scheduled work. Wilson created a virtual user that runs daily, weekly, and monthly routines, drawing on sales and manufacturing data from Dynamics 365 to generate recurring reports. Results flow into Microsoft Power BI, Microsoft Excel, Microsoft Word, email, and presentation outputs. 

For plant managers, scheduled reporting turns production data into an immediately usable management tool. Each Monday at 8:00 a.m., the team receives a report identifying jobs that performed well and those that may need attention, giving leaders the information they need to begin their weekly meeting without waiting for manual analysis. Pamarco also routes daily and weekly business summaries covering key measures such as sales activity, cash movements, and banking information. 

The same scheduled approach now supports board reporting, extending automation from operational updates to executive decision support. A sales director who previously spent about five hours each month preparing sales analysis for the board now receives an automatically generated presentation on the sixth working day. The presentation arrives with the latest figures already populated, along with designated fields for adding business context, leaving the director to review and refine the narrative rather than assemble the report from scratch.

External guidance from Microsoft partner Crowe played a critical role in the early stages of Pamarco’s journey. Crowe helped Pamarco get started without displacing the customer at the center of the work. Crowe supported an early proof of concept and advised on the company's AI direction, while Microsoft specialists helped Pamarco understand Copilot Studio, Dynamics 365 integration, and the MCP approach. Wilson and his team then adapted those capabilities to suit Pamarco's finance processes.

More time for proactive financial decisions

Finance professionals are shifting from coordinating repetitive tasks to reviewing exceptions and acting on information. Scheduled reports arrive without the same manual assembly effort, invoice coding handles routine cases, and autonomous close tasks advance when their dependencies are ready. 

Burgess sees this as an improvement in the overall management of the business. “Our business is better served when our senior financial team has a role in various management decisions, but they were so busy rolling from close to close that it wasn’t possible.” Wilson describes the progress made in the close process, “We’ve been able to automate roughly 40% of financial close tasks with autonomous AI agents with the potential to reach 50% to 60% by year-end." 

The result is not a fully automated finance function. It is a controlled operating model that reserves human attention for decisions that need context and judgment. “We don’t think about it in terms of reducing headcount; it just means we can achieve more with our employees because they can skip repetitive tasks and focus on things that humans are better at. That’s how we’re looking at AI and that’s how it will benefit us long term.” 

Knowledge continuity and adoption

An added benefit is the reduced risk of losing operational knowledge as experienced employees retire or move into different roles. By placing repeatable knowledge into governed agents connected to Dynamics 365, the company can make critical routines less dependent on any one person while keeping experts accountable for the final decision.

Adoption is progressing, and support from executive leadership reinforces the practicality of the AI approach. Wilson regularly meets with users to uncover recurring tasks and translate them into bounded scenarios, helping the tools work for the people who know the processes best.

Extending agents into commercial operations

Pamarco plans to deepen finance automation and then extend the pattern into sales, quoting, order entry, customer response, and vendor communications. The company sees particular value in using historical transactions and pricing information to improve quotations, reduce unexpected outcomes, and respond to customers more consistently. Planned customer-facing scenarios will follow the same requirement for defined access, controlled actions, and measurable value.

Looking for more inspiring agentic AI transformation stories? Explore additional customer successes on the Microsoft Agent Transformation Stories site.