What did you want to find out at the start of your project?
How AI can be utilized to convert documents that have been run through Optical Character Recognition and create summaries of the documents for use in a collection.
What did you actually do?
AI was used to on pre-scanned documents and, using custom queries, summaries, keywords, the documents were prepared for a more accessible online collection
What was the most difficult thing you found?
The sheer amount of variance in the source documents, and the speed at which AI models improved or changed; it was difficult to determine what may be the most stable going forward
What surprised you?
The amount of data that was able to be compiled, and how fast it was (relatively). It would have taken months to do what has been done with the sample set
What is next for you and/or your organization?
Continue to run the documents through the AI and see about stabilizing the process
What 3 quick takeaways would you give to help future participants/your current fellow participants?
Keep lots of notes, keep everything concise.