Summary
Everything up to this point is to understand use cases and where to apply resources to evaluate and decide what to do with a ChatGPT solution. This chapter is an agile method that applies to ChatGPT interactions, recommender experiences, backend solutions, or any software development project. We needed a method to bridge the gap between defining use cases and deciding what use cases to tackle. Creating a repeatable method applies to any development project.
Do a few things with the learnings in this chapter:
- Estimate the value of existing stories, use cases, bugs, or features
- Work with development to explain the story and get their cost estimate
- Work out issues in scoring stories and prioritize the backlog
- Integrate WSJF into the product life cycle
Once a solution is up and running, we can apply more sophisticated approaches to evaluate LLM performance, but the same scoring concepts can be used there. We are almost done with our pre-development...
