The Social Security Administration is pushing a new generative AI chatbot into everyday staff work at a difficult moment for the agency. Employees have been told the tool is available for all staff, but the rollout has already raised questions about training, accuracy, and whether workers have time or confidence to use it.
The chatbot is called the Agency Support Companion. According to an agency email reviewed by WIRED, it was “designed to assist employees with everyday tasks and enhance productivity.” Before using it, employees are required to watch a four-minute training video.
A chatbot arrives during agency disruption
The rollout comes amid chaos and upheaval at the Social Security Administration caused by Elon Musk’s so-called Department of Government Efficiency (DOGE), according to the source article. That context matters because the agency is already dealing with reduced headcount at SSA offices, and some employees described being too busy with actual work to engage with the new AI tool.
Work on the Agency Support Companion did not begin with Musk or DOGE. One SSA employee with knowledge of the app’s development told WIRED that work started about a year ago, long before Musk or DOGE arrived at the agency. The app had been in limited testing since February before being made available to all SSA staffers last week.
That timeline suggests the chatbot was not a last-minute addition. Still, the broader workplace environment appears to have shaped how employees received it. Multiple SSA employees, including front office staff, told WIRED they ignored the email announcing the chatbot because they were focused on compensating for reduced headcount.
The training problem is hard to ignore
The most immediate issue is not simply that the training video looked dated. The required four-minute video features an animated, four-fingered woman in a style WIRED described as resembling websites created in the early part of this century. Some employees reportedly mocked the graphics.
More importantly, the video failed to communicate one of the most critical rules for using the chatbot: employees should not use personally identifiable information, or PII, when using the assistant. For a public agency whose work involves sensitive information, that omission is central to the story.
The SSA acknowledged the issue in a fact sheet about the chatbot that was shared in an email to employees last week and reviewed by WIRED. The agency wrote, “Our apologies for the oversight in our training video.” The same fact sheet tells employees to “refrain from uploading PII to the chatbot.”
That correction is significant because AI tools used inside large organizations depend on clear boundaries. A staff chatbot may be meant to help with everyday tasks, but employees still need precise guidance on what information can and cannot be entered. When that guidance is missing from the mandatory training itself, the rollout risks creating confusion at the exact point where clarity is most needed.
Employees report low interest and weak results
Several employees told WIRED the chatbot has not become a major topic inside the agency. One source said, “Honestly, no one has really been talking about it at all,” adding, “I’m not sure most of my coworkers even watched the training video. I played around with the chatbot a bit and several of the responses I received from it were incredibly vague and/or inaccurate.”
Another source said coworkers were making fun of the training video. “You could hear my coworkers making fun of the graphics. Nobody I know is [using it]. It’s so clumsy and bad,” the source said. That person also said the chatbot gave inaccurate information.
Those accounts point to two separate adoption barriers. First, employees may not be paying attention to the rollout because they are already stretched. Second, the employees who did test the chatbot did not describe a strong first impression. A productivity tool has to earn trust quickly, especially when it is introduced into a busy operational environment.
The source article does not provide a full technical evaluation of the Agency Support Companion. It does, however, show that early employee reactions included concerns about vague responses, inaccurate information, and the usefulness of the required training material.
Why this rollout matters
The Social Security Administration’s AI rollout illustrates a broader challenge for organizations adopting generative AI: deployment is not just about making a tool available. Staff need training that covers practical risks, especially around personally identifiable information. They also need a reason to believe the tool will save time rather than add another task to a crowded workday.
In this case, the agency’s own materials described the chatbot as a productivity aid for everyday tasks. But employee comments reviewed by WIRED show that availability alone did not generate enthusiasm. Some ignored the announcement, some briefly tested the tool and were unimpressed, and others focused on the quality of the training video.
The strongest lesson from the rollout is straightforward. For an internal AI assistant to be useful, the basics have to be credible: accurate responses, clear data-handling rules, and training that does not bury the most important instruction outside the video employees are required to watch.
The Agency Support Companion may have been in development for about a year and in limited testing since February, but its broader introduction is now being judged by staff under real workplace pressure. That is where AI adoption succeeds or fails: not in the announcement email, but in whether employees trust the tool enough to use it while doing the work already in front of them.