From credit ratings to AI in the cloud: Experian’s quiet reinvention – News

It began as a note stuck to the wall during a 2018 brainstorming session at Experian plc’s consumer business unit. The two-word idea — “score boost” — was one of a hundred that came up that day, but it was the one that helped save the company’s $2 billion consumer services business.

Experian’s consumer division was struggling at the time, squeezed by fintech upstarts such as Credit Karma LLC and posting multiple unprofitable quarters. In early 2019, it launched Experian Boost, a free service that enables consumers to add payments to their credit records that don’t typically appear on credit reports. Consumers who make their bill payments on time can improve their credit scores and do so by an average of 13 points. More than 17 million people now use the Boost mobile app.

Today, Experian’s consumer business is not only profitable but also central to a larger transformation of the 125-year-old company, shifting from a credit rating bureau to an analytics and software provider. The firm now generates 35% of its global revenue from software and platforms. The shift has been fueled by a large-scale migration to the Amazon Web Services Inc. cloud and a new push into artificial intelligence.

Cloud-first, AI everywhere

Experian’s core business of reporting on the creditworthiness of individuals and businesses still matters, but the company is increasingly positioning itself as a provider of infrastructure for decision-making in financial services, from fraud detection to real-time risk assessment. It’s using a combination of cloud services and AI to automate complex data migrations, minimize downtime, enhance data accuracy and deliver more scalable and secure services.

That tracks with broader industry shifts. Like most financial data firms, Experian is racing to stay ahead in a cloud-native world where data flows continuously and customer expectations are higher than ever.

Alex Lintner led the team that developed Boost. He has since been rewarded with a broader role as chief executive officer of Experian technology, software solutions and innovation, overseeing North American operations and shaping global strategy. In an interview with News, he explained how Boost is part of a larger narrative about how Experian is rethinking its role from scorekeeper to an enabler of smarter financial decisions.

The company’s transformation has been powered by a 10-year agreement with AWS that gave it “enhanced performance, scalability, reliability, lower operating costs and better security,” Lintner said.

The cloud also enabled Experian it to consolidate data from around the company into a data lake that is expected to surpass 100 petabytes in size within 18 months. It built its own tools to manage the migration, including software for loading and reloading data into AWS S3 buckets with real-time refreshes. “Our clients tell us we have the freshest data,” Lintner said. “Not every bureau can say that.”

The heart of Experian’s AI operations is Ascend, a platform that began as a sandbox for data scientists to visualize data and test models. That led to “AscendOps,” which automates the handoff between the data science and information technology departments, translating models built in Python or R into production-ready code for financial institutions.

“That process used to take three weeks,” Lintner said. “Now it takes two or three days, and still includes human quality control.”

Fraud reduction

Experian has developed AI-driven models for fraud detection that it says dramatically outperform traditional rules-based systems. Detection rates have improved by 37% in loan scenarios and 45% for credit card applications. One lender avoided more than $250,000 in losses during a single fraud attack cycle, Lintner said. Meanwhile, manual review volumes were cut in half, and customer experience improved from a reduction in false positives.

The company is also investing heavily in AI agents, which can operate semi-autonomously and take actions. Digital agents monitor the health and accuracy of models, prompt data scientists when anomalies are detected, and suggest new data sources to enhance model performance.

Like many companies experimenting with agents, Experian is testing them broadly and cherry-picking the most effective ones.. “We’ve built a couple of hundred agents,” Lintner said, “but the top 20 get most of the usage.”

One popular agent focuses on “model drift,” alerting teams when a model’s performance starts to degrade and recommending changes that can be made via a drag-and-drop interface. “They used to have to rebuild the model manually,” Lintner said. “Now they just approve or decline the prompt.”

AI systems monitor for suspicious behavior patterns, like servers firing off tens of thousands of loan applications. The company uses behavioral biometrics, such as typing speed or the frequency of typographical errors to determine if a person is real or a bot. Near-real-time updates catch fraudulent activity earlier than ever. “You need to react as fast as the bad guys deploy a new strategy,” Lintner said.

AI is also helping combat “Frankenstein IDs,” a tactic in which fraudsters assemble fake identities using real but unrelated data usually pulled from different sources. Lintner said its systems can now detect such fake documents far more quickly than before.

“The goal is to minimize the blast radius, whether that means catching fraud before it happens or limiting the number of affected accounts,” Lintner said.

LLMs for compliance

Beyond fraud detection, Experian has built large language models tailored to regulatory compliance. One early application is automating the creation of reports required under the Supervisory Guidance on Model Risk Management, a set of principles known as SR 11-7 required by the Federal Reserve to validate credit risk models.

Financial institutions used to spend months producing the 200-page-plus SR 11-7 reports, often consuming entire data science teams during the final months of the year. Now, an LLM pulls from the model metadata and other sources to auto-generate the required documentation, check for completeness and output a PDF for regulators. “It used to be 30 people doing nothing but documentation,” Lintner said. “Now, we can auto-generate it with an LLM.”

Technology alone doesn’t drive transformation. Lintner acknowledged that getting Experian’s 23,000 employees up to speed on AI tools is an ongoing process. “We have 11,000 using it at a proficient level,” he said. “It depends on what kind of work they do.”

The company uses what it calls “love metrics” to track adoption — measuring time spent in tools, repeat usage and organic growth. “We don’t advertise these tools,” Lintner said. “If usage grows, we assume they find value.”

From helping consumers raise their credit scores to equipping banks with real-time insights, the company’s AI shift is ultimately about broadening access and trust. As Lintner put it, the goal isn’t just faster systems, but smarter, fairer financial decisions.

Photo: Experian