Frequently asked questions

Tokenize is an AI usage management platform for the whole organization. It shows how your teams use AI, where they could work more efficiently, and gives you controls to manage access and spending. It brings observation, optimization, and governance into one place.

Companies often buy AI seats across several teams without a clear view of how they are used. It becomes difficult to tell which teams use them, what the spending produced, or how costs may change next month. Tokenize answers those questions with usage data.

All of them. Tokenize is an organization-wide platform, not a tool for one department. Usage rolls up by person, team, and department, so you get one picture of AI across the company and can still drill into any part of it.

Tokenize is usually owned by whoever is accountable for AI across the business. That may be an AI or platform lead, IT, finance, or operations. The owner is typically responsible for explaining what the company spends on AI and what that spending produces.

Vendor dashboards show you that vendor, and only the teams on it. Tokenize shows every AI tool side by side, across every team, with one definition of cost. It also connects spend to the work it produced, which vendor billing pages do not do.

Claude and Claude Code, OpenAI Codex, and Cursor today. We add tools as customers adopt them, so ask us about anything your teams use that is not on that list.

No. Tokenize sits alongside the tools people already have. Nobody has to switch tools, change accounts, or learn a new way of working.

Tokenize shows what your teams asked AI to do, which models they used, what it cost, and what the work produced. You can view a single session, a person, a team, or the whole company from the same dashboard.

Yes. That is the default view. Usage rolls up by person, team, and department, so you can see where the work and the money go and who to talk to about it.

Yes. Tokenize measures usage consistently across tools and teams, which makes the comparison reliable. Vendor dashboards are harder to compare because each one reports usage differently.

Yes. Individuals see their own activity and their own suggestions. In our experience that is where most of the improvement comes from, because people adjust once they can see what their habits cost.

No. It is not a productivity scoring tool and we do not build leaderboards of who is best. It measures how AI is being used and where the money is going.

Tokenize identifies repeated work, oversized prompts, and tasks where a less expensive model would work just as well. It gives people specific suggestions based on their usage instead of simply telling them to use less.

It connects spend to output, so you can see what a piece of work cost and what it produced. You still make the judgment call. We give you the evidence to make it with.

The usage data shows which approaches are working well inside your company. Experienced users receive suggestions based on their own habits, while newer users can learn from practices that have already worked for their colleagues.

You can set budgets by team, review spending forecasts, control which models people can use, and redact sensitive data before it leaves a laptop.

You get alerted, with the forecast that led there. Budgets are there to make spend visible early, not to cut people off mid-task.

Yes. You can set model access rules for the organization. That is the usual way teams keep expensive frontier models for the work that needs them.

No. The controls run in the background, and Tokenize emphasizes visibility over approval steps. People do not have to ask permission each time they use an AI tool.

Records of AI assistant work: which tool and model ran, when, how many tokens, what it cost, and what kind of task it was. Whether prompt text is included is your organization's choice, and it is a setting you control.

No. We do not scan, copy, or store your documents or your repositories.

That depends on your organization's prompt content setting. When prompt content is off, Tokenize records usage details without recording the text. When it is on, admins can review prompt content to investigate usage patterns. Sensitive values such as passwords and API keys are redacted before anything is sent.

Secrets such as passwords and API keys are removed before data leaves the device, so Tokenize never stores the original values.

Yes. An organization-level setting prevents prompt text from being sent. Cost, token, and usage data remain available when prompt content is off.

Yes. We have a documented deletion process that covers the usage records and the reports built from them. Ask us and we will walk you through the scope and the timeline.

To Tokenize's own hosted pipeline, over an encrypted connection, authenticated with a key issued to your organization.

Yes. Single sign-on, plus directory sync so that people, teams, and groups stay in step with your identity provider as they join, move, and leave.

Access follows the roles you set. Admins see the organization-wide view, individuals see their own. Directory sync keeps that accurate without manual cleanup.

Our latest security and compliance documentation, including our current SOC 2 status and available reports, lives in the Tokenize Trust Center. You can review the public materials there or request access to restricted documents.

You will need API keys from each AI provider you want to connect. A small background agent also runs on each person's laptop, on macOS and Windows. Most companies deploy it through their device management tooling, so people do not have to install it by hand. During a trial, people can install it themselves instead.

You can create an organization and begin receiving data the same day. The timing of a full rollout depends on your device management process.

There is a small menu bar icon on macOS and a tray icon on Windows so people can see it is on and check its status. It is deliberately quiet otherwise.

Yes. The agent keeps a local record and sends it once the machine is back on a network. Nothing is lost because someone worked on a plane.

Usage data appears immediately. Patterns such as repeated work and model mismatches usually become clear within the first couple of weeks.

Yes, and we recommend it. Start with a few teams, look at what comes back, then decide how wide to go.

Explain what Tokenize does and whether prompt content is enabled. People should understand that Tokenize is intended to measure how the company uses AI, not to monitor individual productivity.