Integrating Cognitive Search & OpenAI
Customer
Medical Provider
Challenges Faced by Medical Practitioners in Acquiring Recent Scientific References on Disease Relationships Amidst Time Constraints
Challenge
- Medical practitioners face constraints in time and capacity when searching for relevant content
- Challenges arise when seeking current scientific references to comprehend the correlation between two diseases
- Hallucination risk: Generative models may inaccurately identify or connect irrelevant information
Solution
- Use of Azure-powered tool and combining Cognitive Search & OpenAI
- Implement fine-tuning model or prompt engineering crucial to minimising hallucination risk
- Test: Experiment with limited data source for nephrology articles
- Putting proper guardrails resulted in no response to out-of-context queries, which was the proper and expected behaviour
Business Impact
- Quickly sifts through vast scientific articles
- Doctors can now find references that potentially contradict the mainstream findings to provide the practitioner with a wider view
- Rapid extraction of insights from unstructured data aids informed decision-making
- A competitive advantage achieved in the respective industry