<h5><b>Insurance case study</b></h5>

data-xy-axis-lg:null; data-xy-axis-md:50% 0%; data-xy-axis-sm:47% 0%
<h3 class="m-0 pb-1"><b>At a glance</b></h3>
<p><span class="text-accent1-light"><b>Industry<br> </b></span>Insurance</p> <p><span class="text-accent1-light"><b>Location</b><br> </span><span style="line-height: 26.0px;">United States</span></p> <p style="margin-bottom: 0; line-height: 26.0px;"><span class="text-accent1-light"><b>Challenge</b></span></p> <p>Implementing an agent-based solution for financial analysis and report generation to enhance the Investor Relations team's capacity for strategic initiatives.</p>
<p><b><span class="text-accent1-light">Success Highlights</span></b><br> </p> <ul> <li>95 to 98% positive feedback on model-generated report parameters, reflecting strong alignment with the client’s expectations.</li> <li>About 40% rate between recommended questions and those asked during the insurer’s earnings call.</li> <li>24% reduction in report drafting time, improving operational efficiency.</li> </ul>
<h3><span class="text-accent2-dark"><b>Our approach</b></span></h3> <p>Cognizant collaborated with the client’s Investor Relations team to evaluate the feasibility of implementing an agent-based solution for financial analysis and report generation. This approach aimed to minimize manual summarization tasks, enhance competitive analysis through data insights, improve preparation for earnings calls, and elevate stakeholder communication readiness for the client’s executive leadership team.</p>
<h5><span class="text-accent2-light"><b>Gen AI-powered autonomous agent—a step in the right direction</b></span><br> </h5> <p>We developed a generative AI-powered autonomous agent based on the Reason and Act (ReAct) framework to automate financial analysis and report creation. This agent was deployed in an AWS environment using Agent Calls orchestrated through Lambda. Unlike conventional AI and RPA tools, the agent operated systematically, unsupervised and with a self-improving approach to handle end-to-end task execution. It utilized a reflection module for self-correction and to enhance output quality. The solution leveraged standard operating procedures (SOPs) provided by client SMEs to ensure the output met the client’s expectations.</p>

<p class="pt-1">Key components of our solution included:</p> <ul> <li><b>Hyperautomation:</b> Automated complex financial analysis and report generation, significantly reducing the need for human intervention.</li> <li><b>Superlative insights:</b> Enhanced analysts' abilities with generative AI-powered reasoning and judgment, leading to improved financial insights.</li> <li><b>Bias removal:</b> Demonstrated consistency in countering inherent human biases, promoting more objective financial reporting.</li> <li><b>Decision support:</b> Showed proven potential to generate AI-based questions for earnings call preparation, providing new insights and boosting strategic readiness.</li> <li><b>Document processing:</b> Managed a wide range of documents and unstructured data, simplifying the analysis of diverse financial materials.</li> </ul>
<h3><span style="font-weight: normal;"> <b><span class="text-primary">Business outcomes</span></b></span></h3> <p>The solution was developed using publicly available data. All reports generated were evaluated by the client’s SMEs to assess the suitability of the proposed approach to their business needs. This resulted in better focus, improved reporting, timely insights and better preparation for earnings calls. Specific benefits achieved include:</p> <ul> <li>95 to 98% positive feedback on model-generated report parameters, reflecting strong alignment with the client’s expectations.</li> <li>About 40% rate between recommended questions and those asked during the insurer’s earnings call.</li> <li>24% reduction in report drafting time, improving operational efficiency.</li> </ul>

<h3><span class="text-primary">Related case studies</span></h3>