Situation
A global research firm produces frequent reports on global macros trends for clients. Their users faced challenges navigating and leveraging their extensive research database. The sheer volume of data made it difficult for both internal teams and clients to efficiently find and utilize relevant insights.
The firm sought to integrate AI to create a chatbot capable of more efficient querying, aiming to:
- Enhance client engagement with their research.
- Potentially open new revenue streams by licensing content to other clients.
This initiative was part of a broader digital transformation to modernize their platform and improve content delivery.
Task
The client engaged us to design and build a working prototype chatbot that could:
- Answer client questions about their research.
- Resurface valuable but hard-to-find content.
We were tasked with leading the entire project end-to-end, including prototyping, feature scoping, design, integration, testing, and deployment.
Action
We structured the project into four key phases: Discovery, Implementation, Refinement, and Deployment.
- Discovery Phase:
- Benchmarking: Collaborated with the client to define key questions the chatbot needed to answer and what “correct” responses would look like.
- Data Readiness: Reviewed and categorized their research documents, aligned on necessary metadata and tags, and designed a storage framework.
- Prototyping: Uploaded sample documents to tools like Google Notebook LLM to explore potential chatbot outputs.
- Cost Modeling: Scoped out budget requirements for implementation and scaling.
- Implementation Phase:
- Built a scalable data pipeline using their existing cloud storage provider for documents and spreadsheets containing metadata.
- Developed a daily job to automatically update the chatbot’s knowledge base by processing new or updated documents.
- Leveraged a tech stack including Next.JS, Vercel AI SDK, Langchain, React, Qdrant, and WordPress integration
- Refinement Phase:
- Iterated on chatbot responses based on real-world use cases to ensure accuracy and relevance.
- Incorporated feedback loops for continuous improvement.
- Deployment Phase:
- Delivered the chatbot as an integrated feature of the client’s existing platform, aligning it with their ongoing digital transformation efforts.
- Provided post-launch support to ensure smooth adoption.
Results
- Increased Client Engagement: The chatbot enabled users to query and discover valuable insights within the client’s research database, improving client satisfaction and interaction time.
- Resurfaced Hidden Value: Helped the client rediscover and leverage underutilized content, leading to new upsell opportunities.
- Exploring New Revenue Streams: The Client is now positioned to explore new revenue streams leveraging this new platform.
Catalyst for Transformation: The project served as a cornerstone for the client’s broader digital transformation, setting the stage for future innovations in how they deliver content to clients.
Key Takeaway For Executives
This project showcases how leveraging AI can turn existing assets into competitive advantages. By integrating a customized chatbot into their platform, the client enhanced user engagement, uncovered new revenue opportunities, and laid the groundwork for continued digital innovation.