Gen AI case studies: Productivity and business gains across industries
Written by Mats Johard
4 April 2024
Beyond the buzz, some promising gen AI projects are already seeing daylight. Together with our clients, we are developing and exploring POCs and programs targeting general productivity as well as business-specific and domain-specific areas. To get you an idea of what’s cooking, we’ve gathered some examples here:
- Semantic search at a global biopharmaceutical. We built an intelligent search to parse medical research, learn across the enterprise content landscape and understand product design drawings. The gen AI-powered search tool, where both Azure OpenAI and Da Vinci were deployed, was launched in late 2023. It allows users to search the vast client repository alongside the latest medical publications to extract smart answers. It reduces decision-making time from days to minutes and allows a 360-degree analysis of reports.
- Unlocking enterprise agility for UK insurance company. The customer was dealing with challenges across the software delivery lifecycle including ineffective unit testing coverage in development, prolonged test data preparation in QA, and knowledge silos across teams. Deploying gen AI solutions (Azure OpenAI and GitHub Copilot) across the delivery lifecycle increased efficiency, velocity and agility by automating manual bottlenecks while integrating seamlessly to boost productivity. The efficiencies resulted in 18 percent overall QA cost savings and a 90 percent reduction in data framework development time (from 2–3 weeks to just 2 days).
- Shrinking the distance between BI and decision-makers for a major airline. This client is one of the world’s largest airlines, meaning that they have broad, powerful web performance data to draw on. However, the process of accessing specific, nuanced analytics data from the enterprise data management system creates a workflow bottleneck, requiring query-writing manpower. In late 2023, we built a test environment based on the real-world data structures of the company’s live ecosystem. With our Gen AI SQL generator, business users identify the data they seek, and AI automatically generates the SQL code required to extract insights from the data warehouse. Early predictions suggest an 80 percent efficiency gain.
- Virtual clinician for a Middle Eastern nation’s Department of Health. In connection with the huge technology event GITEX GLOBAL in Dubai, we were asked to build a virtual clinician, powered by AI and complete with an interactive, self-animating avatar.
"...turn conversations between the chat interface and patients into clinical notes for comparison and diagnosis by 30 AI systems, each trained narrowly on a single field of medicine."
To orchestrate that vision, we utilized LLMs (Large Language Models) to turn conversations between the chat interface and patients into clinical notes for comparison and diagnosis by 30 AI systems, each trained narrowly on a single field of medicine. The highest confidence diagnoses were returned to the user, alongside recommended next steps. Once fully functional, more than 100 doctors tested and honed the system before launching at the event.
Why engage Cognizant for gen AI projects then? Our full-scale consulting, advisory and delivery capabilities are backed by industry expertise. We have a platform-based approach, where Flowsource and Neuro AI knit together disruptive AI capabilities with pre-existing tooling to combine the best of both worlds. If you want to learn more, please visit our Generative AI section and Rewire for AI.
