The transformation of Deluxe from a century-old check-printing giant into a modern fintech powerhouse is a masterclass in strategic evolution. Led by the vision of redefining business identity, the company has successfully pivoted from focusing on the physical manufacturing of paper checks to mastering the digital flow of alternative payments. This journey involves not just a technological overhaul but a deep cultural shift, where data is treated as the primary asset and AI serves as a force multiplier across every department. By integrating sophisticated platforms and fostering a culture of continuous upskilling, the organization is navigating the complexities of the digital age while maintaining the trust it has built over 110 years.
Transitioning a legacy manufacturing brand into a digital-first payments platform requires a fundamental shift in business identity. How did you redefine the role of technology from a support function to the core product, and what specific cultural hurdles did you encounter during this strategic pivot?
The shift began in 2018 when we realized that while we were known for printing checks, our true business was facilitating alternative payment mechanisms. For decades, the company identified as a manufacturer, viewing technology merely as a back-office support tool, but we had to flip that script to see technology as the actual product and service we deliver. The cultural hurdle was significant because we had to move a 110-year-old organization toward a digital-first mindset, yet we found strength in our roots. Our employees possess a deep-seated commitment to customer trust and quality, which allowed us to drive complex transformations more easily than I expected, as the team was willing to endure the “hard work” of change to do right by our clients.
Moving toward a domain-driven data model often involves months of manual mapping by subject matter experts. Since large language models can now automate these complex data structures, how did you implement this transition, and what metrics demonstrate the efficiency gains compared to traditional manual methods?
We moved away from siloed data by building three core platforms that allow us to have a 360-degree view of the customer across all business domains. Historically, this would have required our subject matter experts to spend months painstakingly mapping disparate data elements into a new unified structure. By leveraging generative AI and large language models, we automated this mapping process, effectively bypassing the traditional timeline that would have taken several months of manual labor. This breakthrough allowed our experts to focus on governance and strategy rather than the rote mechanics of data organization, representing a massive leap in operational velocity.
Traditional finance processes, such as month-end closing, frequently take weeks to finalize due to data silos. When applying AI to automate these back-office operations, what step-by-step changes were necessary to streamline the workflow, and how has this impacted the overall speed of business decision-making?
Our approach follows a three-pillar strategy: AI for tech, AI for business operations, and AI for customers. In the finance department, we targeted the “month-end close,” which ironically used to take us a couple of weeks into the next month to actually finalize. By organizing our data into clear domains and applying AI to automate the reconciliation and data flow, we’ve moved toward a model where “month-end” actually happens at the end of the month. This acceleration means our leadership team isn’t looking at “old news” to make decisions; we now have a much faster pulse on the company’s financial health, enabling more agile strategic moves.
Managing sensitive financial data while using public large language models introduces significant risks regarding hallucinations and privacy. How do you structure an internal “trust layer” to prevent data leakage, and what role does a cross-functional executive committee play in approving high-risk AI use cases?
Protection of data is paramount because financial institutions have trusted us for over a century, so we built a proprietary “Deluxe Trust Layer” that acts as a secure wrapper around public LLMs. This architecture uses Retrieval-Augmented Generation (RAG) to ensure our sensitive data stays within our walls while still benefiting from the intelligence of public models. To manage risk, we established an AI Executive Committee—including our CFO, HR head, Legal head, and myself—to vet use cases, starting with low-risk internal tools before moving to customer-facing ones. We also maintain a Client Advisory Board with our top customers to gain their consensus on how their data is being handled before any rollout.
A “3×3” upskilling framework targets everyone from entry-level staff to executives. What specific technical and soft skills are prioritized at each level to ensure the workforce remains relevant, and how do you prevent hierarchy from becoming a barrier to collaborative learning during this shift?
Our 3×3 framework ensures that no one, from the newest hire to the CEO, is left behind, categorizing skills from beginner to advanced across all tiers of the company. For entry-level employees, we are prioritizing soft skills—specifically the ability to ask the right questions and develop deep domain expertise—because these are the human elements AI cannot replicate. To break down hierarchical barriers, we run joint upskilling programs where junior staff and mid-career professionals learn side-by-side, fostering a collaborative environment where the focus is on solving problems rather than following a chain of command. This approach helps mid-career folks manage the pressure of upskilling while juggling their existing full-time responsibilities.
Automated receivables platforms now use generative AI to extract invoice data and handle cash application deductions. In what ways does this technology improve the end-customer experience, and how do you leverage internal chatbots to provide customized support across highly specialized business units?
We’ve integrated generative AI into our R360 receivables platform to handle complex tasks like extracting invoice data and managing cash application deductions, which previously took human analysts hours or even weeks to resolve. This creates a much smoother, faster experience for our customers who want their payments processed without friction. Internally, we developed “Deluxe AI,” a centralized chatbot platform that uses specialized business rules to appear as a custom assistant for each different business unit. This means a customer service rep in our check-printing division gets a completely different, tailored intelligence experience than one in our promotional products division, all powered by the same core infrastructure.
Beyond internal cost optimization, how can a long-standing company use emerging technology to leapfrog competitors rather than just making incremental improvements? Please share an anecdote or a specific scenario where a rapid technological prototype replaced a traditional, months-long consulting process.
True “leapfrogging” occurs when you use technology to completely reimagine how you solve problems that have existed for 50 years, rather than just tweaking an existing process. I recently put this to the test by using AI to generate a comprehensive research paper on the impact of the Genius Act and stablecoins—a task that would typically cost $500,000 and take three months with a major consulting firm. I completed the research in less than two hours using AI, and when I presented it to our CEO and executive team, they couldn’t find a single hole in the logic or the data. This demonstrates that we can now develop sophisticated strategic points of view in minutes, allowing us to pivot and innovate at a speed that traditional competitors simply cannot match.
What is your forecast for the future of AI-driven digital payments?
I believe we are entering an era where payments will become entirely invisible and hyper-personalized, driven by the convergence of cloud, big data, and generative AI. In the next three to five years, the winners won’t just be companies that use AI for cost-cutting, but those who use it to retrain their people to think like innovators rather than process-followers. We will see a shift where the “entry-level” workforce focuses almost exclusively on high-level soft skills and domain strategy from day one, while AI handles the heavy lifting of data extraction and fraud detection. For Deluxe, this means our 110-year legacy of trust will be paired with near-instantaneous digital execution, making the transition from paper to platform complete.
