Reimagining data matching in the AI era: How modernization helped Deluxe triple processing speed and cut costs by 99% - Impetus
Blog post

Reimagining data matching in the AI era: How modernization helped Deluxe triple processing speed and cut costs by 99%

Discover how Deluxe enhanced scalability, improved performance, and paved the way for sustainable growth in partnership with Impetus and AWS
Impetus/Resources/Reimagining data matching in the AI era: How modernization helped Deluxe triple processing speed and cut costs by 99%

In financial services, data matching is a foundational capability that keeps transactions accurate, compliant, and trustworthy. Whether it’s ensuring payments reach the right recipients or enabling precise customer targeting, the accuracy and speed of data matching greatly influences business outcomes. Yet, as data volumes grow exponentially and customer expectations increase, traditional matching systems often struggle to keep pace.

Deluxe, a payments and data leader, wanted to rebuild its core matching platform which handles a wide spectrum of workloads and powers critical customer-facing services. Deluxe uses this platform to compare and reconcile data from diverse sources—often based on individual or business names and addresses. With growing volumes of matching workloads and concurrency needs, their homegrown legacy matching platform struggled across multiple dimensions:

  • Inconsistent matching accuracy, leading to false positives/negatives
  • Limited concurrency, with failures beyond ~30 parallel workloads
  • Rising infrastructure costs from always-on systems
  • Operational inefficiencies, including job failures and SLA risks

As workloads increased, these issues compounded, impacting not just system performance but customer experience and profitability.

From bottlenecks to breakthroughs: Building a cloud-native, event-driven matching platform on AWS

To break past these barriers, Deluxe tried leveraging OpenSearch for vector embeddings and search, but the results were inconsistent, with high-scoring matches occasionally getting missed. The company then partnered with Impetus to rebuild its matching platform and transition to a modern, AWS-native architecture purpose-built for scale, speed, and efficiency. Deluxe’s ongoing partnership with Impetus for data and AI development ensured a deep understanding of their systems, enabling seamless collaboration and effective delivery.

Impetus redesigned the existing matching platform to a flexible, modular, distributed architecture on AWS, leveraging their in-depth experience of modernizing large-scale data platforms.

Key innovations included:

  • Transitioning from a monolithic system to independently scalable jobs—each optimized for its workload
  • Replacing always-on infrastructure with on-demand compute that automatically adjusts resources
  • Swapping vector embeddings with string similarity algorithms to meet deterministic matching needs
  • Converting Python UDFs to Spark-native processing to improve execution speed 7–10x
  • Eliminating redundant data processing steps to reduce overhead costs

A powerful combination of AWS services was leveraged to ensure intelligent orchestration:

  • AWS Lambda and Amazon DynamoDB to manage configuration settings and concurrency limits, queuing and validating job requests 
  • AWS Glue for filtering incomplete records, performing hash-based matching, and partitioning data
  • Amazon SageMaker to process partitioned datasets using FAISS-based in-memory vector search for matching precision
  • AWS Glue jobs to apply custom matching rules and write the results to Amazon S3 for downstream use

“When we attempted transforming our legacy matching service the first time, we didn’t get the scale or performance we were looking for. We needed a trusted partner who could bring in the vision needed to build a new, modern matching solution. This is where Impetus came in and helped us extensively.” — Satish Balasubramanian, Divisional CTO, Data & Enterprise Architecture, Deluxe

Enabling high performance and limitless scalability at 1% of the cost

With the new cloud-based matching platform, Deluxe shifted from slow, failure-prone batch processing to reliable, high-speed execution powered by on-demand compute. The transformation redefined what’s possible for data-intensive operations in financial services, proving that scale, speed, accuracy, and cost-efficiency can coexist. Matching for massive datasets (100 million+ records) which once took weeks to process—was now completed in approximately two hours. What’s more, monthly vector search costs dropped from about $75,000 to nearly $700—bringing the cost per run down to roughly $1 per million records. Modernization also helped reduce operational overheads, eliminating the need for manual infrastructure management. Additionally, Deluxe was able to process more workloads simultaneously, handle larger and more complex datasets, and adapt quickly to evolving business needs.

Key business benefits:

  • 99% reduction in monthly vector search costs
  • 3X faster processing for match jobs
  • $1 cost per match run with million records
  • Hundreds of parallel jobs supported in production

“With Impetus and AWS, we’re no longer constrained by processing time or concurrency limits. That’s been a game-changer for scaling our operations,” added Satish Balasubramanian, Divisional CTO, Data & Enterprise Architecture, Deluxe.

Laying modern data foundations for an AI-driven future

As data volumes surge and AI adoption increases across industries, the ability to scale core operations like data matching is becoming a defining advantage for financial services organizations. The winners will be those who can process faster, scale smarter, and optimize costs, without compromising accuracy or reliability.

Deluxe’s transformation journey reflects a broader shift toward cloud-native, event-driven data architectures. By rethinking how core processes like data matching are designed, Deluxe didn’t just solve for performance; it built a foundation for continuous scale, agility, AI-driven innovation, and sustained revenue growth.

Ready to modernize your data foundation and ensure AI-readiness?

With deep expertise in data engineering, agentic AI, and cloud modernization, Impetus helps power truly intelligent enterprises. Learn how you can transform your legacy data platforms into high-performance, cost-efficient engines—which fuel AI innovation and support your unique analytics use cases.

Read the full AWS case study to dive deeper into Deluxe’s modernization journey with Impetus and AWS.

Authors

Vigyan Agrawal – Director, Strategic Accounts, Enterprise Solutions

Hussain Saify – Senior Technical Architect, Enterprise Solutions

Learn more about how our work can support your enterprise

Explore more resources
Review our services
Get in touch