LLM constructs as viable stores of institutional knowledge

Description

This project would explore setting up a framework of topically focused LLMs structured to capture the prompts and extended dialogue of history across DEP experts. These could be deployed to assist the experts while being themselves built and refined by expert use. Special effort could be made to collect grey materials and to gather white papers and essential resources that could be specific to topics and to NJ places and scenarios. These tools would inherently be GPT-like tools usable by others and especially able to blunt the impact of retirements and vastly enhance succession outcomes.

Problem Statement

DEP operates on knowledge and capabilities that are significantly possible only through individuals and their awareness of, fluency with and access to highly specialized information that only their experience and unique perspective can synthesize into intelligent outputs (communication, answers, recommendations and decisions). These people readily retire without effective capture or transfer of their knowledge or capability.

Project Justification

Each departure from DEP incurs some setback in knowledge and capability that may not be replaceable or is costly and delayed in restoring.

Estimated Transactions

None

Components

Large Language Model (LLM)

Target Rollout Date

None

Target Rollout Date Reason

None

Attachments

Activity