I strongly doubt any AI agent would make that typo. And I'm really not sure why you trust the Verizon AI chatbot here.

The easier option I'm afraid would have almost been to just not pay your bill.

No, I'm not even kidding here, Verizon has apparently in the last year randomly just been turning on multi-month past due accounts and going "lol we'll just let you come back if you pay your next month's bill like nothing ever happened" and zeroing out thousands+ dollar past due balances in the name of "customer loyalty".

[1] https://old.reddit.com/r/verizon/comments/1tqh4or/400_overdu...

I once set up a 1 month hotspot with Verizon when I moved into a new place that didn’t have internet yet.

After the month, I returned the hotspot to the store and they told me I was all good (in writing too!).

Then I started getting monthly bills that I couldn’t cancel online or over the phone. I repeatedly went into the store and they told me it was a billing mistake and they’d fix it.

Eventually they sent me to collections, my credit score dropped, and a debt collector started coming after me.

It took 2 years to finally get them to stop and to remove it from my credit history.

Verizon is a bad company that doesn’t care about its customers.

ah, verizon and billing mistakes, name a more iconic duo.

immediately reminded me of the "verizon doesn't know dollars from cents" phone call circa 2006.

https://verizonmath.blogspot.com/2006/12/verizon-doesnt-know...

https://verizonmath.blogspot.com/2007/08/original-recording-...

In general, the first thing you should do when you encounter a bad AI agent is to ask for a human agent, threatening to cancel otherwise. For a truly dumb agent, threatening to cancel could work even if you don't have an active account. In fairness, before you do escalate to a human, do give the AI agent a fair chance, as not all are bad.

Unnecessary ragey terms in the title -- how about Verizon’s AI Billing Agent Should Be a Case Study on How Not to Deploy an LLM in Business