Connect
You used to know what your monthly bill for developer tools would look like. A seat count, a fixed rate, and you were done. But as AI becomes a standard part of your team’s workflow, that predictability is gone.
We’re moving from fixed subscriptions to a token economy. A single developer experimenting with a high-reasoning model can spike your costs overnight.
This isn’t a small shift. Globally, LLMs and AI agents are already processing an estimated 1.5 quadrillion tokens every month.
Managing access is still the foundation of your security. But at this scale, you also need to see how these resources are being consumed. Protecting your budget now requires direct oversight.
JumpCloud AI & SaaS Management gives you that visibility. You can govern AI environments without relying only on traditional identity controls. This article walks through the challenges of unmanaged AI spending and uses JumpCloud’s Cursor connector to show how you can bring these tools under control.
Why AI Tools Break the Old Budget Model
AI tools are different from traditional SaaS apps. They’re often free or cheap to start. But enterprise costs tie directly to usage. Specifically, to how many tokens get consumed.
Cursor is a good example. It’s an AI-powered IDE that development teams are adopting fast.
Your team might create accounts and start using the tool lightly. Then usage ramps up. Costs accumulate in ways that license-based SaaS models never produced.
These shifts create three challenges that traditional IT management doesn’t handle well:
- Many AI tools don’t work through browser workflows, so extension-based discovery won’t catch them.
- Organizations often lack clear insight into usage and cost drivers across AI tools.
- Without direct integrations, IT can’t validate who has access or spot unauthorized use.
Without direct visibility into tools like Cursor, you’re left with open questions. Which teams use AI tools, and how much? What is your total spend on AI model usage? Are tokens being used redundantly across tools? Do your current policies cover secure AI adoption?
There’s a way to answer all of them.
Getting Real Visibility with the Cursor Connector
JumpCloud AI & SaaS Management closes the visibility gap around AI tools by adding native connectors for the AI space. Instead of inferring usage from a browser extension, these connectors pull data directly from provider APIs.
That means accurate usage data, cost clarity, and deeper visibility across your AI tool stack.
To show how this works, let’s look at Cursor — one of the most common AI tools in the modern development stack.
The Cursor connector pulls user lists, MCP server connections, and AI model usage from the Cursor API every six hours. It includes token counts and costs. It also captures new users and app associations through regular polling to keep your data current.
When you connect Cursor to JumpCloud, you can see:
- An up-to-date inventory of all team members with Cursor accounts, with shadow accounts clearly labeled.
- Which external MCP servers your team is connected to and how they relate to other apps you’ve discovered.
- Detailed breakdowns of which AI models are in use, with token counts and estimated costs per user.
- A view of total AI token usage and estimated costs over time, right in your JumpCloud dashboard.
Turning Data Into Better Decisions
Visibility is the foundation. The real value is what you do with it. With the Cursor connector, you can move from reactive blocking to strategic enablement.
Lock Down Developer Accounts
When shadow accounts surface, you can move those users into your managed environment fast. That keeps development work inside your security perimeter and prevents source code from floating in unmanaged instances.
Track Costs by Team and Model
By watching consumption at the model level, you stop guessing and start attributing costs accurately. Usage tables show which models drive your highest spend. That gives you what you need to forecast budgets or negotiate enterprise contracts based on actual trends rather than seat-count estimates.
Reduce Integration Risk
Monitoring MCP server connections helps you verify that AI integrations stay secure. These connections can be entry points for attackers or vectors for data loss. Seeing them in real time means agentic activity stays within your sanctioned environments and doesn’t bypass your data policies.
Simplify Compliance
Meeting SOC 2 or GDPR requirements usually involves manual, point-in-time audits. The Cursor connector supports compliance efforts by logging user activity and app associations automatically. That gives you a reliable audit trail and proves to auditors that you have centralized oversight of your AI stack.
Setting Up the Cursor Connector
If you need visibility into Cursor usage and token spending, the connector is a practical starting point. You’ll need admin access to both Cursor and JumpCloud.
- Generate an admin API key in Cursor’s Team Settings with the “Admin” scope.
- Add the connector in your JumpCloud Admin Portal under Access > AI & SaaS Management > Settings.
- Once connected, data collection begins. You can track usage, costs, and adoption patterns almost right away.
See the Help Center for setup details.
Make Secure AI the Default
The core shift in IT is enabling AI in a way that is secure and aligned with your goals.
When IT has visibility into AI usage, the focus shifts from blanket blocking to strategic enablement. IT can identify which tools drive the most value and which pose the highest risks. Teams can then standardize on the right tools while keeping data protected.
Shadow AI doesn’t go away through restriction alone. It’s mitigated when IT provides better visibility, governance, and support than employees can find on their own.
Ready to stop guessing about your AI costs? Head to the JumpCloud Admin Portal or sign up for a free account to set up AI & SaaS Management.
