Generating efficiencies and staff engagement with AI
Milliman technologists automate manual processes for French insurer

Business model optimization wasn't CEGC's only concern. Staff well-being was also top of mind.
CEGC is a leading French insurer of home loan guarantees and other financial instruments, and accurate claims compensation depends on thorough document review. But with many processes performed manually, Mansour Sow, CEGC risk manager, knew that some steps could be automated to enhance efficiency.
In particular, claims handlers regularly had to stop what they were doing to flip through thick paper files to find an elusive signature or ID card. This low-skilled work was tedious and disrupted higher-value projects. It could cause staff frustration. And the slow pace kept partnering banks waiting for answers, potentially jeopardizing transactions and relationships.
If CEGC continued to operate like this, Sow worried the firm would lose its competitive edge. But actuaries and statistical engineers are contributing to the firm’s digital transformation. As head of a team charged with integrating artificial intelligence (AI) into CEGC’s decision making and risk management, Sow knew the opportunities the technology provided.
“We wanted to simplify these processes and automate some of the time-consuming tasks so staff could focus on more complex, high-value work,” he recalls. “We knew that using AI to make these processes more efficient would yield significant value.”
But CEGC needed help. For nearly 13 years, Sow and his team had collaborated with Milliman on projects in several areas, from risk scoring to statistical modeling. Once again, they turned to Milliman, this time with a different kind of assignment: develop process automation that would help speed document review, allow employees to focus on more high-value work, and foster innovation at CEGC.

Actuarial acumen meets tech talent at Milliman
Founded in 1947, Milliman has a deep history in traditional actuarial work—but as the field has evolved, the firm has embraced innovations. Today, actuaries partner with data scientists, technologists, and other professionals across Milliman to develop tools that help clients around the world manage the risks in their industries.
In Paris, a dedicated research and development (R&D) lab has been focused on generative AI.
Alexandre Boumezoued, a Milliman principal and the R&D lab director, leads a team comprising experts in large language models (LLMs), mathematical engineering, software architecture, and other specialties. Many have PhDs in their field.
“Everyone has a strong technical background and a rigorous scientific approach to problem-solving,” says Boumezoued, himself an engineer with a doctorate in applied math.
In the past year alone, the R&D lab has developed vision models for interpreting charts and formulas within a PDF, LLMs that incorporate actuarial reasoning, and computational AI agents that can generate code for performing calculations and responding accurately to user requests. The team also has used AI to help insurers process actuarial spreadsheets and streamline economic scenario production.
Across Milliman, actuaries and other professionals are incorporating AI into their work, from advising life insurers on the benefits of LLMs to building AI solutions to manage property and casualty insurance claims. They’re also keeping abreast of the evolving global response to AI among insurance regulators.
When CEGC approached Milliman with its latest request, the team was ready with tools that could be adapted to develop a custom solution.
Collaborating with Milliman on these technologies will enable the automation of business processes for these various tasks. It will provide a significant time saving and will make the work much easier and more enjoyable for employees.”
Mansour Sow, Risk Manager, CEGC
Automation improves CEGC operations
For the CEGC project, the R&D lab worked with Milliman Senior Consultant Loup Ortiz, who leveraged his longstanding relationship with the client to map the insurer’s existing workflows, breaking down each activity into individual steps and taking the time to understand each process from start to finish. Because of the volume of work, CEGC chose to focus on changes that would have the most impact. The Milliman team then identified which tasks were performed manually and reviewed them one by one to determine potential automated solutions, from document reading to data extraction.
“AI is not a magical tool that can solve every problem,” explains Adel Cherchali, a Milliman senior consultant who manages the R&D lab’s AI researchers. “It is most effective when tailored to a specific task.”
“What was innovative—and what made this project successful—was how we matched each task to the most appropriate AI technique,” Ortiz adds.
This was no small feat. “One manual job involved reviewing 20 different documents, which meant training 20 models,” Sow says.
The ability to leverage insights from the Milliman R&D lab—such as customizable optical character recognition tools—facilitated the project. For example, to spare a claims manager from flipping through those thick paper files, Ortiz worked with Cherchali to train a model that can scan documents, locate a needed signature no matter where it might appear on a form, and extract the necessary information—accurately and quickly.
When Sow brought these solutions to CEGC’s senior managers, they were enthusiastic about implementation. Automation, they saw, would immediately add value and set CEGC apart from the competition.
“Collaborating with Milliman on these technologies will enable the automation of business processes for these various tasks,” Sow says. “It will provide a significant time saving and will make the work much easier and more enjoyable for employees.”
The ROI on AI for a large insurer
Thanks to Milliman, CEGC claims managers today spend five to six minutes less per document review, a significant gain in the high-volume business. While employees are spared the low-value work, customers get faster answers. CEGC leaders have already seen the financial benefit of this project. And the return on investment (ROI) will continue to grow, Ortiz says, because the tools can be adapted to fit other CEGC processes.
“We can apply the model to review other types of documents without completely redeveloping it,” he explains. “We fine-tuned it on one specific case, but we can easily scale it to others. This flexibility is very important and improves the project’s ROI.”
While Sow enjoyed working again with Milliman on another fruitful project—and he is pleased with the balance sheet impact—there are other benefits to CEGC from implementing AI.
“Beyond the financial and time savings,” he says, “it will empower our actuaries to tackle more complex challenges, enhancing their job satisfaction and well-being.”
This article was written by Milliman Senior Writer Adin Bookbinder.
Explore more on this topic
See how Milliman’s AI solutions cut through the hype to solve real-world problems.

Solutions for a world at risk
Technology is powering breakthroughs in healthcare, insurance, and life sciences—but high costs, insufficient access, and inequity persist. Factors like market volatility and longevity risk are threatening the retirement security of millions of people worldwide. And as governments, businesses, and communities face the complex, cascading effects of climate change, the need for bold solutions of unprecedented scale has never been more urgent.
We're here to help
Ask the tough questions. We're ready for them.