Forum Discussion

acamar22's avatar
acamar22
Helper I
1 year ago

Using advanced AI to analyze free-form text

Hi guys!

 

I'm looking for some direction using AI in my imported data sets. I have mass data sets with complex free-form text.

 

I've done some research, and it looks like the built-in AI is not advanced enough.

 

Ideally, I'd like a chatGPT type AI where I can type a prompt and the AI searches for specific scenarios/keywords.

 

Using things like Text.Contains() is not ideal as there are so many variations of keywords I'm looking for.

 

example) "Analyze columns that indicate a part failure: if provided, record part number and failure mode."

 

Is something like this possible? If so, what programs do I need? I'm willing to learn any skills needed.

 

Thanks all!

2 Replies

    • Callum_McMahon's avatar
      Callum_McMahon
      New Member

      Hi! Tabular data is one area where built-in AI solutions are lagging behind more dedicated products.

       

      There’s a few main players out there, each targeting different demographics. e.g.
      clay.com - mainly for smaller one-off lists, say up to 100 rows, costing a few hundred dollars a month

      everyrow.io - tackling much larger datasets, in the 5000+ row domain, at costs designed for scale

      ChatGPT directly - supports uploading your CSV, but uses a single model to slowly work through your dataset, not ideal for systematically processing data in a repeatable way

      Manus - great if your analysis questions involve clicking through web forms and interacting with complex sites to uncover the answers you need. Very pricy per individual answer

       

      As far as usability goes, learning to effectively specify your analysis criteria is a pretty transferable skill across all such tools.