Hey all,

We have been playing with OpenAI as much as I imagine many of you have been, and built an internal test file to process PDF documents and analyze them with OpenAI's chat engine. We dropped a quick video up on YouTube showing what we're doing and thought I'd share it here in case anybody's interested in taking a look.

We're pretty excited about what's possible with this kind of technology, and I'm glad to field any questions anyone has.

 

Hey Court! I have been experimenting a bit with the OpenAI completions API.   My company performs about 30,000 product installations per year for a big box store.   We aggregate a huge amount of information as we schedule and perform these jobs.   This has been a great project using the API via Insert from URL.

So, I've built a process (a helpful assistant) which, as we wrap up product installation jobs, will help us improve our level of customer satisfaction. The feature aggregates workorders, customer feedback from various sources, including notes, SMS conversations, email correspondence, and survey data.   The result is a summary report of all customer and partner interactions.  

It takes that deep dive of information and presents it to our expediters, schedulers, installers, and management as areas where we can improve our service.

It is a great way to quickly review all the feedback and identify areas of improvement. 

My next project will include langchain and building a local vector repository that can be used to query not only the working data, but be able to bang it all against vendor specifications/guidelines and our business partners’ specific specifications (their “guardrails”). 

Fun will be had!