Royal London Group

Solution

Cognizant and Royal London Group implemented a governed generative AI review assistant to shorten the compliance cycle for regulated marketing content. The solution uses Microsoft Azure’s OpenAI LLM inside a modular assistant split into three personas—Creator, Refiner and Approver—so content generation, improvement and approval follow explicit stages instead of a single opaque model call. That persona-based architecture mirrors Royal London Group’s existing content workflow and makes outputs easier to audit. Each asset moves through the assistant under human-in-the-loop controls, with subject matter experts performing the final review before anything is approved. Cognizant also refined the assistant through stakeholder feedback, quality assurance checks and scaled evaluation, while embedding generative AI skills inside Royal London teams so the process can be reused economically in a highly regulated environment.

Data flow

Regulated marketing assets enter the modular assistant, where the Creator, Refiner and Approver personas apply staged Azure OpenAI processing rather than a single model pass. Outputs are then checked by human subject matter experts in the final SME review, and the assistant is continuously refined through stakeholder feedback, QA checks and scaled evaluation.