Verizon customers are increasingly struggling to reach human support staff as the carrier embeds artificial intelligence deeper into its operations. Reaching a live agent has become notably harder, prompting some users to try unconventional methods to break through the automated systems.
Absurd prompts to bypass the bots
One customer on Reddit described spending seven hours attempting to submit a case to a customer champion, only to be routed repeatedly between departments before the connection dropped every time. Another recounted calling in, being handed to an AI agent, then transferred to a second, more sophisticated AI agent, with neither able to resolve the issue. That customer eventually switched carriers.
A separate user, posting as PepperCat1019, claims a nonsensical prompt does the trick: telling the bot they have a purple cat. According to their account, that detail is enough to break the AI’s script and secure an immediate transfer to a live agent. The method is not proven, and it may stop working once the systems adapt.
AI overload at Verizon
Chief executive Dan Schulman was brought in last year to stem customer losses and steer the company back on track. Although Verizon used AI before his arrival, Schulman has leaned heavily into the technology as part of a broader efficiency drive aimed at cutting billions in costs.
More than 13,000 jobs have already been cut, and several employees suspect they have been training AI systems designed to replace them. Schulman expects AI to replace 20% to 30% of the American workforce over the next two to five years. Verizon is being rebuilt around AI, with customer service bearing much of the impact.
Robotic responses and hidden automation
Verizon has developed specialised small language models for support, drawing on Google’s Gemini and Anthropic’s Mythos models. The company claims a 90% accuracy rate, yet many customers report that the AI chatbots fall short of expectations. Verizon has said queries requiring human input would be handled by human representatives, but numerous online accounts indicate users are frequently left dealing with bots instead.
Some customers report that the automation goes further, with AI at times presenting itself as human. One user described a support agent that appeared to be AI despite background call-centre noise designed to sound authentic. After being disconnected and calling back, they reached what sounded like the identical voice, a female voice with a southern accent that read as scripted rather than conversational.
While humour and repeated requests for a live agent occasionally work, customers report inconsistent results across their attempts to reach human support. Verizon continues to route support queries through its layered AI systems built on the Gemini and Mythos models.
Source
Image: phonearena.com
