PDP Chat Bot
Product Statment
The Provider team is facing the challenge of reducing onboarding time to the Provider Data Common (PDC) platform, which has resulted in daily calls between the architect and Horus teams. This frequent communication is time-consuming and inefficient.
In simpler terms, the problem is that it takes too long for new team members to get up to speed on the PDC platform, leading to daily calls between different teams. This constant communication hinders productivity and efficiency.
Link to Canvas Dashboard: PDP BOT
Solution
The proposed solution consists of two main approaches:
- LLM (Language Model) + Vector DB + Knowledge graph approach for storing PDC documentation: This approach suggests using LLM as a tool to extract relevant information from the documentation related to the PDC platform. The extracted information will be stored in a Vector DB using a knowledge graph structure. This will make it easier and faster for team members to access the necessary information without relying on daily calls.
- LLM as an agent to extract information + SKG (Structured Knowledge Graph) approach: In this approach, LLM will be utilized to parse through documents and extract specific details like email addresses, names, etc., based on prompts or queries given to it. This automated extraction of information will further reduce the need for frequent calls and allow team members to quickly find and retrieve the required data.
Key Metrics
- Number of office hour questions resolved without the need for meetings: This metric measures the reduction in the number of queries or issues that can be resolved during office hours without the need for additional meetings or calls. It indicates how effectively the solution enables self-service access to information, reducing reliance on manual communication.
- GD (General Discussion) automation: This metric tracks the level of automation achieved in GD discussions related to the PDC platform. It measures the extent to which LLM and other automated processes can assist in extracting relevant information and providing valuable insights during discussions, reducing manual efforts and improving efficiency.
These key metrics focus on quantifying the reduction in communication overhead and improvement in self-service access to information, both of which are crucial for addressing the problem statement and evaluating the success of the proposed solution.
Target Customers
- The target customers for the proposed solution are the members of the Provider team. This includes individuals involved in onboarding new team members to the Provider Data Common (PDC) platform and those who frequently communicate with the architect and Horus teams. The solution aims to streamline their onboarding process, reduce communication barriers, and improve overall productivity within the Provider team.