Using AI to connect the dots, precisely

KPMG helped a U.S. payments network combine big data, artificial intelligence and statistical modeling to advance its capability to detect fraudulent transactions.

U.S. payments network
Financial services
Transaction security innovation
  • Client challenge
  • Benefits to client
  • Approach
  • KPMG insights

Client challenge

In the algorithmically-driven world of transaction security, this payments network knows precision is everything: Incorrectly blocked purchases mean lost sales and diminished customer loyalty. The company asked KPMG to help it develop an AI-driven fraud analysis engine as a value-added service for the network partners and card issuers who rely on them for fraud prevention support.

KPMG combined advanced machine learning algorithms and pattern recognition modeling to extract previously uncovered insights from our client’s transaction data. The company has gained the ability to offer more precise “second opinions” about suspect transactions, in real time. Working with KPMG, they’ve strengthened their competitive position as a trusted technology partner in a changing payments landscape.

Benefits to client

Working with KPMG, our client has combined big data, AI, and statistical modeling to add a ready-to-deploy transaction fraud detection capability. Specific benefits include:

  • transaction fraud prevention developed as an added-value service and competitive differentiator to power business goals of increased transaction volume and partnership network expansion
  • reputation enhancement among card issuers and merchants, as an active, innovative participant in “collaborative security” fraud prevention initiatives
  • data science practices advanced, as the company moved beyond statistical sample model testing to tackle the computational challenges of testing with massive pools of transaction data.


KPMG worked with our client to develop a fraud transaction proof of concept, along with an implementation roadmap. KPMG:

  • structured a co-development relationship between KPMG and the company, drawing upon proprietary KPMG AI-driven statistical modeling capabilities and large pools of the network’s transaction data
  • designed an AI-driven payment fraud detection model, using proprietary KPMG machine learning algorithms and statistical modeling techniques such as Extreme Gradient Boosting and Statistical Regression
  • verified the fraud detecting precision of the new model by parallel-processing 365 days’ worth of client-supplied transaction data, stepping beyond traditional statistical sampling models
  • developed and delivered an implementation roadmap to integrate the newly developed fraud detection model into the client’s existing transaction fraud infrastructure.

KPMG insights

Organizations need AI advisors who understand their competitive implications

As AI succeeds in widening the competitive divide between leaders and laggards in all industries, clients appreciate KPMG’s ability to help them integrate the right combination of technology, organizational capital, and data strategy to advance their own AI agendas.

We actively explore ways to bring frontier ideas to life

As part of our ongoing commitment to designing and building practical solutions, we structured an innovative co-development relationship model that leveraged the AI modeling capabilities of KPMG with massive pools of proprietary data from our client.

We put advanced analytics at the service of business objectives

KPMG actively distills the AI experience and insight gained in data analytics engagements around the world into ready-to-deploy tools, technologies, and accelerators so clients can take advantage of our capabilities and move quickly to their objectives on day one.

Being an active, innovative participant in 'collaborative security' fraud prevention initiatives has enhanced this company's reputation among card issuers and merchants.