The client sought to overhaul their approach to summarizing medical inquiries, a process that was traditionally cost and labor-intensive. They partnered with HCLTech to implement a GenAI-based solution that revolutionized their operational efficiency by producing accurate, human-like summaries of scientific data rapidly and at a reduced overall cost.
The Challenge
Growing need to expedite and refine the summarization process while enhancing the accuracy and ensuring cost-effectiveness
Health Care Professionals (HCPs) often have crucial medical inquiries regarding scientific data on drugs that pose significant challenges for the medical team. Addressing these queries involves extensive literature searches and content summarization. Presently, this process heavily relies on external consultants who invest substantial time and effort. Reliance on human resources leads to inefficiencies and delays in providing timely responses to HCP's queries. Using GenAI, our client wanted to automate the content summarization process and streamline the entire process to provide prompt and accurate answers to HCP inquiries, improving overall responsiveness and client satisfaction.
The Objective
The overarching goals were distinct and metrics-driven
- Drastically reduce manual effort in literature summary tasks
- Achieve a higher level of accuracy and reliability in summaries
- Significantly cut down the cost of operations related to summarizing medical literature
The Solution
Innovating with GenAI lab and Azure OpenAI services to automate PubMed article summarization
HCLTech crafted a state-of-the-art solution leveraging the client’s GenAI Lab along with Azure OpenAI services to create a system that could interpret instructions in natural language and automate content summarization. The technology focused on extracting context-specific information from voluminous PubMed articles. After meticulous testing of numerous GenAI models like GPT 3.5, Bio-GPT and MedPalm2, the system selected the most effective based on a ROGUE Score — a measure used to evaluate and compare the AI-Gen summaries to those crafted by subject matter experts.
The Impact
The results of implementing this GenAI-based summarization tool were outstanding and measurable
- Time Efficiency: The manual effort was decreased from 8 hours to a mere 10 minutes, achieving a striking 95% time savings
- Accuracy: The model-generated summaries achieved a 70-80% similarity to human-produced reference summaries, thereby assuring a high degree of accuracy
- Customer Experience: HCPs experienced significant improvements in the timely, efficient and accurate reception of responses to their queries, thus enhancing overall satisfaction
