Hey all,

I have an idea like many of you to start leveraging generative AI like GPT as a foundation but build my own models on top of it.

Essentially I want to find a really annoying business process in an enterprise and then introduce AI to that process to completely streamline it or eliminate the manual component of it.

Example, I’m a sales person and I would like to submit all the orders I got from my customers today. Instead of emailing an admin or me having to go into an ERP and key them in, I just type up my orders in a chat gpt like interface and it will automatically route those orders into my ERP and enter them for me and then I just go about my day.

Are enterprises receptive to this kind of software? Will they trust a small company to do it?

  • AHardCockToSuck@alien.topBEnglish
    1·
    1 year ago

    You would need to make damn sure the results are always correct and predictable. A missing order could be a lot of money lost

    • Belmeez@alien.topOPBEnglish
      1·
      1 year ago

      My example was a simple one but consider the alternative.

      Entering an order in an ERP requires filling out a form. Anywhere between 10-30 fields, many of which are indirectly redundant (which customer, now which address, etc)

      Many times it’s also hard to know what the right information is. I know what the product is called in the field but does it have the same name internally? (Even worse if it’s just a product ID)

      So entering an order in many cases is a 20-40 minute affair, and that’s just a single order.

      If you use a generative AI on top of this process, you can just free form text like you and me and millions of others already know how to do and get those orders out by just describing the order. You can even describe 20 orders at the same time and get them all in.

      Oh and for new people joining an org, the ERP method is even slower. The learning curve is steep