• 0 Posts
  • 13 Comments
Joined 1 year ago
cake
Cake day: August 8th, 2023

help-circle

  • Nice try. But let’s play with the thought that there’s no way we can let a rookie listen in on customer calls and gradually work their way into the role until they have enough experience… What about hiring technicians/professionals that has been working with the products/services for 10 years?

    That would be a way of getting competent customer support people, right?

    And just to clarify my comment that you replied on: The problem today is that most often there’s no career path for the customer support rookies and the pay is so lousy that most people just work customer support until they get something better.

    That’s definitively the correct way to avoid getting experienced people in the customer support.


  • Unpopular opinion and rant: Us, the consumers, brought this on ourselves. Not intentionally but it was a slippery slope.

    No one I know did ever ask the sales representative “does your customer support answer within 5 minutes and will I always reach a representative with att least 10 years of experience, that has the authority to make real decisions?”. No, but we were all very interested about the pricing of the service/product.

    Then these “Please press 1 for…” happened… and no one of us really cared about the change because the service providers offered a much lower price than the ones with customer support representatives with 20 years of experience. Since all of us went for the cheapest provider, the other ones had to cut cost to be able to offer their service on a competitive price level. So then there were no one offering competent support with representatives that knew their shit. And it slowly continued to go downhill…

    So here we are with shitty services, which we pay for, where we all are treated as cattle.

    If people at least started to ask for better customer support there would someone, who wants to climb the corporate ladder, creating a PowerPoint presentation with a real VIP Service Level. Of course it will cost more money, because real people cost money, but we would att least get what we want.

    But no. Consumers will still go for the lowest price.


  • How good is good do you say?

    We got a pretty good results with CER at 4% and WER at 15%!

    This was on a limited dataset used to test and train which most likely means that if you introduced an even larger dataset with greater variations in handwriting style for testing the numbers might be even worse.

    Very simplified: A risk of a character wrong every 20th character and a word wrong every 7th word. The SER was around 20%.

    There’s an reason why no one has released a good model for western letters yet and why companies pay up to 1€ for capturing data from 10 handwritten pages.

    It will come but OCR isn’t as sexy as developing text2image solutions.






  • To train an AI to recognize handwriting you need a huge dataset of handwriting examples. That is millions of samples of handwritten text + information about what the written text says in every example).

    This is why the best engines only exists as a service in the cloud. The OCR engines you can install lovely that are acceptable, but far from perfect, are commercial. Parascript FormXtra is one of the better commercial ones.

    The only OCR Engine that’s free and really good is Tesseract OCR but it doesn’t handle handwritten text.



  • No degree at all but been working IT since the 90’s.

    It’s fun that when I started in IT everything went from centralized (mainframe and terminals) to decentralized (PC). Then came Citrix and everything went towards centralized. Smartphones and apps came so we went decentralized again. Then cloud came and we essentially went centralized again.

    It’s all about trends, the pendulum swings back and forth…