Company decides to implement ChatGPT for client work until one employee discovers dozens of unchecked AI mistakes: ‘[Now] staff members have to complete mandatory retraining’

Author image for Cata Holmes
Cata Holmes • Writer
Oct 4, 2026 1:45 PM EDT

What happens when AI confidence meets human oversight?

AI is one of those things that has come into our lives and is not going away. It has come to stay and is going to affect and change the way we live. We're still discovering how it works and what we can do to get the best out of it, without damaging our work. So it's basically trial and error until we find the sweet spot of it. 

In the story we have below, a company decided to implement ChatGPT into their normal workload to make it faster, and what they thought it could be more accurate. They used to send clients emails with very delicate financial information, and they really thought they were doing the right thing. However, one employee checked an information that was sent wrongly to a client and caused a big issue.

They discovered that more than 20 emails had been sent without any fact-checking process, revealing and explaining procedures and information of the company that were incorrect. As a consequence, management decided to stop the use of AI and decided to have staff members complete a mandatory retraining, as well as a professional conduct course.

On the other hand, the employee was asked to write some accounting references for his coworkers because he's the only one who has the accounting degree and is more in touch with the subject. 

Needless to say, IA is not a tool that we can truly trust. It might accelerate some procedures and be helpful for some things, but mainly, it is not really ready to work on its own, so people have to be very careful when deciding to use it. This company now has to apologize to its clients and take back what they said… imagine what that is going to look like.

Advertisement
  • LG อ. 0000 GOOD-CHEAP FAST GOOD-CHEAP-TAT TAST-0000-CLAY CHEAPFAST-0000
    via Unsplash+ (via ZBRA Marketing)

    An employee works at his computer in an office.

    Stock photo unrelated.

  • A few weeks ago, my manager asked us during a team meeting what each of us thought about using Al with clients. I said it could be useful, but that I would never trust it without checking the answer first. We work at a
    via Successful-Stand6699
  • consulting firm, and I've seen ChatGPT confidently give incorrect guidance on tax and accounting matters before, which can affect the small businesses | usually work with.
    via Successful-Stand6699
  • Some members of the senior staff use it regularly, but I didn't expect my comment to bother anyone. One of my coworkers, who mainly works in branding, laughed and said I was being overly cautious.
    via Successful-Stand6699
  • She told us about a time when a client asked her an accounting question and she answered it with ChatGPT's help. I was familiar with the issue she was talking
    via Successful-Stand6699
  • about, and without thinking, I told her that the answer she had given the client was wrong.
    via Successful-Stand6699
  • It turned into an awkward discussion. Another coworker said the idea that I might know better than the software was ridiculous. In the end, my supervisor pulled up the actual text of the
    via Successful-Stand6699
  • regulation, which confirmed that my explanation was correct. Everyone made a few jokes and moved on, so I thought that was the end of it.
    via Successful-Stand6699
  • Yesterday, my supervisor told me that the owner of the company had reviewed more client files and discovered more than twenty cases in which Al-generated information had been sent to
    via Successful-Stand6699
  • A person holds a note labeled “A.I.” in front of computer screens
    via Unsplash+ (via Hitesh Choudhary)

    A person holds a note labeled "A.I." in front of computer screens

    Models pictured are not representative.

  • clients without being properly reviewed. Now several senior staff members have to complete mandatory retraining, as well as a professional conduct course because of the comments they made during the meeting. I was
    via Successful-Stand6699
  • also asked to take on some additional duties and create a simple accounting reference guide for employees, since I have an accounting degree.
    via Successful-Stand6699
  • My supervisor and the people at my level think the response is fair, but I still feel as though I unintentionally caused all of this to happen. I wasn't trying to embarrass anyone or get them sent to training. Should I
    via Successful-Stand6699
  • apologize to the coworkers involved, or would that just make the situation even more awkward?
    via Successful-Stand6699
  • Coworker discuss reports and data during a business meeting.
    via Unsplash+ (via Getty Images)

    Coworker discuss reports and data during a business meeting.

    Recreated scenario included to enhance reader understanding.  

  • Legitimate_Crow_7521 nope - should be praising you
    via Successful-Stand6699
  • Weird_Dependent3732 not your fault. move on
    via Successful-Stand6699
  • neilastin The concept of double checking where Al is involved is a no-brainer IMHO
    via Successful-Stand6699
  • zinsser What little I have messed with Al sometimes produced obviously incorrect answers to relatively simple math questions. I could not imagine using it for actual accounting work and then billing someone for it.
    via Successful-Stand6699
  • iam2bz2p This is NOT the fault of ChatGPT. This is the FAULT of the person using ChatGPT and writing PROMPTS.
    via Successful-Stand6699
  • Various_Education622 Read this exact text word for word a few weeks ago.
    via Successful-Stand6699
  • dbatknight So you're the smarty pants around that place LOL great work now start building a library of reference material for them to check against Al
    via Successful-Stand6699
  • Jyqft That's on your coworker for speaking up and bringing attention to their idiocy. If they kept their mouth shut, maybe you wouldn't have pointed it out there and then.
    via Successful-Stand6699

Want More? Follow Us and Add Us as a Preferred Source on Google.

Tags

employee mistakes chatpgt ai staff members company
Scroll Down For The Next Article