Anthropic's Smart Reports lets managers summarize how their team uses Claude

What Anthropic has launched

Anthropic, the company behind the AI assistant Claude, has announced a feature called Smart Reports. It is aimed at managers. According to Anthropic's own support page, as futurism.com reported, the tool produces summaries of how a team uses Claude.

The stated goal is to let managers describe the work that gets done with the assistant during the day. In plain terms, it turns a team's scattered individual chats into a report that someone in charge can read.

Futurism's headline frames this as a way for your boss to spy on your chats. That is the outlet's own framing, and it is sharper than Anthropic's. The company presents the tool as reporting, not surveillance.

What a manager can actually see

Today's coverage does not spell out exactly what managers can see inside a given conversation, so it is worth being careful here. What the source does say is that the tool goes beyond summaries of sessions.

Per Anthropic's support page, managers can sort their team's AI activity by overall cost and by the output the chatbot produced. That gives a manager a view of who is using the assistant heavily and what that use is costing the company.

Futurism reads this as direct oversight of how efficient employees seem when they use AI tools. That is an interpretation, but it is a reasonable one. Once usage and cost sit side by side, it is easy to compare people, whether or not that is the intent.

The anonymity promise and its critics

Anthropic says the system anonymizes sessions. Anonymizing means stripping out details that point to a specific person, so a report describes activity without naming who did what.

Futurism is skeptical. It says it remains to be seen how secure this is in practice. It also points to an earlier Anthropic effort with "anonymized" data, which it calls a privacy disaster because the information was easily de-anonymized using common large language models. De-anonymizing means working backward to figure out who the data came from. Large language models are the type of AI that powers chatbots like Claude.

That earlier episode is Futurism's account, and today's report gives no more detail on it. Still, it shows why a promise of anonymity deserves a closer look. Whether a summary can be traced back to a person depends on how it is built, and Anthropic's support page claim alone does not settle that.

Why Anthropic wants this feature

Futurism ties Smart Reports to Anthropic's push to win large business contracts as it looks ahead to a possible IPO, the first time a company's shares are sold to the public. Corporate customers account for a reported 80 percent of Anthropic's total business, with individual consumer subscriptions making up the rest.

If that split is right, businesses matter far more than individual subscribers. Companies that buy AI tools for many employees tend to ask the same questions: who is using it, how much is it costing, and is it worth the money? A reporting tool answers those questions directly.

Futurism adds that how useful Smart Reports turns out to be has yet to be determined. The price is not stated in today's reporting either.

What to do if you use Claude at work

The practical lesson is simple. If your Claude account comes from your employer, assume your company may be able to see summaries of how you use it. Keep personal matters out of those chats, such as health, money or job hunting.

It is also reasonable to ask your employer what is switched on. Today's report does not say how Smart Reports is turned on or who can see its output, so those details are worth asking about directly.

• Use a personal account for personal questions.
• Use the work account for work tasks you would be comfortable discussing with your manager.
• Ask your IT team or manager what reporting is enabled.