In 1915, Deluxe Corporation was one of the most innovative companies in the nascent payments space when it invented the pocket checkbook and a proprietary process for printing checks. One hundred and ten years later, the company honors its legacy by continuing to manufacture checks but has evolved into a trusted payments and data powerhouse.
That shift required not just new technology, but a new mindset and a careful balance between innovation and institutional memory. UiPath, the automation provider Deluxe is working with as it transitions to the AI era, also is cognizant of that balance. It has traveled a similar path as an RPA pioneer in the early days of process automation—with an understanding the technology is still foundational—to the current agentic age.
Automation Today sat down with Satishkumar Balasubramanian, Vice President and Head of Architecture and Shared Services at Deluxe, during the UiPath FUSION event to talk about how he has used automation to carve a path for the established check printing and manufacturing company to evolve into the payments and data brand that it is today.
“We were traditionally known as a check printing company,” says Balasubramanian. “Even now, close to 60 percent of our revenue comes from check printing. But over the last seven or eight years, we’ve been actively pivoting ourselves into a proper B2B payments company.”
He explains that the pivot required both business and technology transformation.
“We started looking at how to change our traditional technology landscape into the digital platform company that we want to be,” he explains. “That way, we could support the payments and data side of the business.”
From Manual to Machine Intelligence
When Deluxe began its automation journey in early 2022, the initial focus was on eliminating the most repetitive manual tasks through robotic process automation (RPA). Teams relied on UiPath to integrate simple automations across multiple legacy systems.
“We started with simple RPAs—three-step processes that people were doing manually across screens,” he remembers. “We just brought in a bot between them to automate those steps. That’s how our automation started.”
As Deluxe gained confidence with automation, RPA quickly evolved into more intelligent forms of automation. The company’s internal platform—dubbed deluxe.ai—launched in late 2023 as a hub for combining automation with artificial intelligence.
“The idea with deluxe.ai was to build a proprietary platform within Deluxe to use AI and automation to benefit employees, business operations, and customers,” Balasubramanian says. “It became our AI and automation engine.”
Early initiatives tackled document-intensive operations such as invoice and check processing. But the company’s needs evolved quickly and RPA was not enough. Rather than relying on static OCR (Optical Character Recognition), Deluxe adopted machine-learning-based extraction to identify and interpret diverse document types.
“We had to move on to the next stages, using AI to extract information and make cognitive decisions before automating,” he explains.
Building Trust and Adoption
For Balasubramanian, the technology required to meet their needs was not the challenge. Changing habits in a century-old organization proved harder.
“The technology transformation is always easy—maybe I say that because I’m a technologist,” he notes. “The people transformation is the toughest part.”
To ensure AI was applied responsibly and effectively, Deluxe created a structured framework around three domains: AI for Tech (employee productivity), AI for Business (operational efficiency), and AI for Customer (embedded AI within products and services).
“We were very clear we are not going to use AI just for the sake of it,” he says. “We identified needs and then categorized them so we could focus on the right problems.”
On the employee side, the rollout of Microsoft Copilot served as an accessible entry point. Staff used it to generate meeting transcripts, summarize emails, and draft documents. Over time, teams explored GitHub Copilot for code generation and AI-assisted test design.
“People were a little taken aback at first, because we’re a legacy company and new technology takes time,” he says. “But then everyone started playing around with ChatGPT and coming back asking, ‘What are we doing in AI?’”
That curiosity gave rise to an internal evangelization program: a SharePoint hub listing approved tools and use-cases, newsletters highlighting early wins, and live roadshows—both in-person and virtual—where employees shared their experiments, creating curiosity among others to go and try things themselves.
From Automation to Agentic Workflows
The next phase of modernization at Deluxe centers on agentic automation—AI-driven systems capable of autonomous decision-making. One showcase project, the “pricing audit agent,” automates cross-checks between contract terms and invoices to detect discrepancies that previously required human spot-checks.
Developed using UiPath and AWS Bedrock, it went from concept to production in less than six weeks. That speed underscored the company’s preference for combining off-the-shelf components with internal engineering talent.
“Our mantra is not to build something that’s already available,” he points out. “We use UiPath, we use SaaS providers, but everything sits under the deluxe.ai umbrella. We have our governance, our trust layer, and our audit trails established.”
To maintain control, Deluxe now hosts UiPath in its own AWS environment rather than the vendor’s cloud.
“Every LLM call goes through our trust layer so I know the data is secure, and I’m logging everything for the compliance team,” he explains. “UiPath runs as a natural extension to all the other things we’re doing in AI.”
Re-Engineering for a Century Ahead
Even after winning industry awards for its intelligent automation efforts, Balasubramanian insists the company is “just getting started.” The long-term vision is to embed AI capabilities throughout the organization so that every employee can use them in context, whether writing code, managing accounts, or interacting with clients.
“There is no shortage of processes that we want to automate, re-imagine, or transform,” he states. “From a vision standpoint, we want AI used by each and every employee in whichever way it helps their business or day-to-day work.”
The ultimate goal is not merely to replace manual steps but to redesign how work flows altogether.
“When an agent can make autonomous decisions, I don’t need steps one, two and three to go in sequence,” he says. “The agent can combine everything and make a decision by itself. So, we are automatically re-engineering the process.”
For other long-established enterprises exploring similar transitions, Balasubramanian’s advice is pragmatic: Focus first on the business problems worth solving.
“Every executive should look at their company’s biggest opportunities and see which technology helps address them,” he advises. “Don’t ask, ‘How can AI help me?’ Ask, ‘Can AI help me solve this specific problem?’ Then it becomes an easier fit.”
