See who uses ChatGPT, Codex and ChatGPT agents, how deeply, for which tasks, and what it costs per team. Worklytics turns OpenAI's Enterprise APIs into team-level adoption, AI fluency and cost metrics, without storing a single prompt.
.png)
.png)
.png)
The problem
OpenAI's own analytics show activity inside ChatGPT. They can't tell you which teams get value, whether people are getting better at using it, how ChatGPT, Codex and ChatGPT Work change the way work gets done, or what it all costs per team.
Usage and cost are not tied to your HR structure, so you cannot see which departments, managers or locations justify their seats.
A one-line question and a long working session look the same in seat activation data. They need different fixes.
If you also run Copilot, Claude or Gemini, there is no single view of AI use across tools.
A vendor dashboard cannot show how much of the working day involves AI, or how it changes focus time, meetings and output.
See who adopts AI and why, what it changes in the working week, and what it costs by team, across ChatGPT, Codex and ChatGPT Work.
What Worklytics adds
OpenAI reports on its own product. Worklytics connects ChatGPT, Codex and ChatGPT Work to your org structure, your other AI tools and the work itself.
Active users, days used and depth of use for ChatGPT, Codex and ChatGPT Work by department, manager, location and tenure, from your HR data.
Compare each team with the rest of your organization and with peer companies (benchmarks in beta).
Fluency goes beyond logins: how often and how intensively people use AI, and AI penetration, the share of available working time that involves AI.
See which teams have made AI part of daily work and which are still experimenting, and track movement after training.
Every prompt is classified into the kind of work it supports, such as drafting, summarization, research, analysis or code, without storing the prompt text.
See what each function actually uses ChatGPT for, and where agents are taking on whole tasks.
Find what predicts AI use in your organization: whether a person's manager uses AI, tenure, role and team norms.
Aim enablement where it will move adoption most, starting with managers.
Compare heavy and light AI users, before and after rollout, on focus time, meeting load, collaboration, engineering output and customer outreach.
Shows whether AI is changing the work itself, not just adding another tool.
Usage-based cost for ChatGPT and Codex by team and per active user, next to licensed seats and estimated time saved.
See where spend goes and what it returns before you renew or consolidate tools.
Worklytics vs. native analytics
OpenAI's admin analytics cover activity inside its own products well. These are the gaps for teams that run more than one AI tool or need to connect AI use to the work.
| Capability | Native Workspace | Worklytics |
|---|---|---|
| Seat activation & weekly active users | ✓ | ✓ |
| Sessions & messages/session, 14-week trend | Partial | ✓ |
| Usage-based cost by department and per active user | ✗ | ✓ |
| Task classification by team and function | ✗ | ✓ |
| Cross-tool view (Copilot, Claude, Gemini, Cursor) | ChatGPT only | All AI tools unified |
| Department segmentation without SCIM | Requires SCIM | ✓ |
| Impact on focus time, meetings, collaboration and output | ✗ | ✓ |
| AI penetration of the working day | Not available | By team, trended |
| User-level pseudonymization (no raw data) | Basic aggregation | SHA-256, minimum group sizes |
How it works
Worklytics reads ChatGPT, Codex and ChatGPT Work activity through OpenAI's Enterprise Compliance API, and usage-based cost through OpenAI's Analytics API. No agent, no SDK and no code changes, and nothing is written back to your workspace.
Personal identifiers are hashed with SHA-256 before analysis, and results are shown only for groups above a minimum size. Prompt and response text is never stored. With the Worklytics proxy, this happens before data leaves your environment.
Connect your HRIS, calendar or GitHub to see AI use by team and next to meetings, focus time and engineering work. ChatGPT works on its own too.
Adoption, fluency, task, impact and cost views come pre-built. Most teams see metrics within a week of setup, with trends building over the following 30 days.
Privacy & Compliance
Worklytics measures AI use without reading employee content. It connects through admin-level APIs only, is audited annually against the SOC 2 standard, and is privacy-first by design.
The data is already there. Worklytics connects through read-only APIs and gives IT, HR and leadership pre-built views, each with a clear action to take.
