Plenty of businesses have adopted AI without ever really knowing what it costs them. Someone activates a subscription, a few employees start using it, and then the bill climbs with nobody quite sure why or who’s driving the usage. On July 2, 2026, Anthropic, the company behind the Claude assistant, released an update that tackles this blind spot head on.

The change isn’t about a faster or smarter model. It’s about something far more practical for a manager: seeing where the money goes, setting limits, and getting a warning before the bill spirals. For a Quebec SME just starting to weave AI into its daily operations, this is the kind of tool that matters as much as raw performance.

1. What Anthropic announced on July 2

The update targets the Claude Enterprise plan, the one organizations use to give their teams access to AI. Anthropic added three major capabilities: a detailed analytics dashboard, spend alerts, and the ability to choose which model starts by default. Together they answer a recurring complaint from managers who received one lump bill with no way to break it down.

In practice, an administrator can now see usage and cost broken down by group and by person. The dashboard also shows what was actually produced: documents created, files edited, connectors and skills used, all sitting next to their cost. According to cost-analysis firm Finout, the accompanying API even returns a breakdown by named individual, complete with their usage and dollar spend. It’s a shift from a vague estimate to a sharp picture.

2. Seeing where the AI money actually goes

The heart of this update is visibility. Before, a bill spike looked like a black box. Now a manager can answer specific questions: which teams use AI the most, on what kind of tasks, and at what cost. The dashboard filters by your company’s actual structure, so the numbers follow your existing teams rather than some abstract breakdown.

Anthropic also added a section specifically for developers using Claude Code, its programming tool. It shows the number of active users, the most frequent commands, and an estimate of the productivity gain. A natural-language analytics feature even lets you ask a question about usage and get back an exportable chart, handy for presenting the numbers to leadership without building a report by hand.

• Usage and cost by team and by person, filterable by your existing groups.
• A breakdown of what was produced (documents, files, connectors) alongside the cost.
• An API that connects to spend-tracking tools like Datadog or CloudZero.

This focus on visibility echoes something we’ve preached in IT management for years: you only really control what you measure. It’s the same principle behind our IT management services, where tracking a fleet’s real consumption avoids nasty surprises at month end.

3. Alerts before you hit the wall

The second addition is the one that speaks most directly to an SME manager: spend alerts. As an organization approaches its budget cap, the administrator gets a warning at 75% and again at 90% of the limit. That leaves time to react, adjust the budget, or raise the limit before someone gets cut off mid-task.

On the employee side, an in-app notification appears at 75% and 95%, with the option to request a limit increase directly, without leaving the tool. The manager still has final say, but the request arrives cleanly instead of through a panicked email. For larger organizations, an admin interface lets you automate how these requests get handled and flag usage that’s climbing fast.

This kind of guardrail changes your relationship with AI. Instead of discovering the budget after the fact, you steer it in real time. It’s exactly the kind of control we put in place when helping a business get a handle on new digital tools. If you’re wondering how to manage AI use at your company, let’s talk through our contact page.

4. Picking the right model at the right price

The third piece is more subtle but very cost-effective. Anthropic now gives administrators control over which models are available. You can set which Claude model starts by default for a new conversation, both in chat and in work tools. You can also decide which models are accessible to which roles.

Why does this matter? Because not all models cost the same. Running every routine task on the most powerful model is like taking a delivery truck to pick up milk. By routing ordinary work to a lighter model and reserving the top-tier one for genuinely complex tasks, a business can cut its bill without losing quality where it counts. It’s a setting you configure once and it pays off every month.

• Adjustable default model so routine work doesn’t default to the priciest option.
• Model access controlled by role or across the whole organization.
• Individual-level analytics turned on by default starting July 11, 2026, worth knowing from a privacy standpoint.

5. What your SME can take from this

You don’t need an Enterprise plan to learn something from this announcement. The underlying message applies to any digital tool: adopting something without measuring it always ends up costing you, in money and in risk. AI follows the same rule as the cloud or software licensing. If nobody’s watching consumption, it drifts.

For a business in Mauricie or Trois-Rivières just starting to bring AI into its operations, three habits are worth their weight in gold. First, know who’s using what. Second, set a clear limit with a warning before you hit it. Third, save the costliest tools for the cases that actually justify them. These are the same principles we apply to fleet management, and they now apply to AI too.

Frequently asked questions

Are these new Claude features free?

They’re part of Anthropic’s Claude Enterprise plan, built for organizations. The new dashboards, spend alerts, and model controls come at no extra cost for customers on that plan. Individual plans don’t offer the same level of administrative control.

How do I find out which employee is using the most AI?

The admin dashboard shows usage and cost by person and by team, filterable by your existing groups. Note that individual-level analytics will turn on by default on July 11, 2026, which is worth checking against your privacy policy before you use it.

Can I actually cap my AI spending?

Yes. An administrator gets an alert at 75% and 90% of the set limit, and employees are warned in-app before being cut off. You can also choose which models are accessible, which keeps routine work from defaulting to the most expensive option.

Managing AI in your business without losing control

AI is becoming a line item like any other, and it deserves the same rigor as a fleet of workstations or a suite of licenses. Claude’s new tools show that even the big providers are taking cost seriously. At OKTO Solutions, we help SMEs in Trois-Rivières and Mauricie adopt these technologies with clear guardrails in place. Find out how we can help on our IT services page, or reach out directly through our contact page to get your AI usage under control.