Introduction
Industry-wide digital challenges
Many industries, including healthcare and nutrition, struggle to make their websites easy to navigate and ensure accessible content. This hinders their ability to serve their clients effectively.
Industries need better ways to support customers, provide easy access to information, and offer personalized experiences. In healthcare, for example, professionals often struggle to find the information they need. Companies can improve their service delivery and user satisfaction by implementing tailored solutions.
Challenge
Optimizing online presence for healthcare professionals
In the competitive healthcare and nutrition industry, having a website that is easy to navigate, offers accessible content, and provides personalized experiences is crucial. Our client, a leading company in this sector, faced several challenges in improving their online presence to improve their service to healthcare professionals (HCPs).
Website navigation
HCPs struggled to find specific content on the website, making it difficult to access critical educational materials, services, and products related to medical nutrition. The existing system did not provide comprehensive answers to their queries and lacked sufficient references and links to scientific evidence and tools available on the website.
User engagement
The website setup did not effectively engage HCPs, as shown by the limited time spent on the site. The lack of engagement was due to the website’s inability to accurately capture and address HCPs’ needs, which limited the brand’s ability to offer relevant and engaging content.
Personalized experience
Users had a generic and less effective experience because of the lack of personalized guidance. The system was inadequate for enabling future engagement, such as follow-up communications. This hindered the brand’s ability to move HCPs along the brand engagement journey.
Solution
Comprehensive digital transformation
To help our client improve their online presence for HCPs, we took a comprehensive approach to enhance website navigation, content accessibility, user engagement, and personalized experiences. Our solution was designed to make it easier for HCPs to find information about products and tools and to support future communications.
What we provided:
An advanced chatbot
We developed AMANDA, an advanced chatbot to help HCPs visit the client’s website. AMANDA comes in two versions: a non-gated version with a limited knowledge base, accessible to any visitor, and a gated version with a more extensive knowledge base, available only to logged-in users.
Our approach started with engaging stakeholders to understand their requirements, finalizing user flows, and designing the solution architecture. The solution included several advanced components to ensure a seamless user experience:
Stakeholder engagement and requirement gathering:
Engaged with stakeholders to understand their specific needs.
Finalized user flows for various scenarios, overall approach, and solution architecture.
Advanced Q&A module:
Built on Azure, using services like Azure OpenAI, AI Search, Azure Functions, and private endpoints to ensure scalability and security.
Implemented custom chunking strategies for different document types to generate efficient embeddings, with legal safeguards as required.
Summarization module:
Developed a summarization API to condense chat sessions, enabling future analysis of query types and product searches conducted by HCPs.
Frontend development and personalization:
Designed and developed the front end to match the brand’s aesthetics, integrating a personalized chatbot to enhance user interaction and engagement.
Deployed the chatbot and integrated it with the existing website, allowing it to identify users and offer personalized interactions.
Automated content ingestion:
Set up automated jobs to ingest updates for URLs, podcasts, and PDF files, ensuring the chatbot’s knowledge base remains current and up-to-date.
We chose a Large Language Model (LLM)-based approach over traditional Natural Language Processing methods. LLMs offer superior understanding and generation capabilities. This ensures more accurate and context-aware responses, effectively handling complex queries. The LLM-based solution, built on Azure, provides the necessary scalability, security, and precision to meet the specific needs of HCPs.
Results
Progress and future enhancements
Phase 1 of the project, spanning 14 weeks, demonstrated significant progress and set the foundation for future enhancements. The timeline was strategically divided to ensure comprehensive development and fine-tuning of the solution:
Weeks 1-2: Requirements assessment and project planning
Engaged in detailed discussions to assess requirements and create a robust project plan.
Week 3: Design and architecture preparation
Finalized the design and architecture, ensuring alignment with the client’s goals and technical specifications.
Weeks 4-12: Design, development, and fine-tuning
Focused on designing, developing, and refining the solution.
Conducted multiple show-and-tell sessions to gather feedback from business, legal, and other stakeholders.
Weeks 13-14: User Acceptance Testing (UAT) and release
Allocated the final weeks for thorough testing and successful solution deployment.
Phase 2 discussions are underway to finalize the scope and timeline. This phase will focus on further enhancing AMANDA’s functionalities. It is anticipated that one to two months post-launch will be required to fully understand and realize the deployment’s benefits.