Assess Usage and Value of ChatGPT Work

3 min video3 key momentsWatch original

⚡

TL;DR

OpenAI's admin console now lets organizations track where ChatGPT usage is actually driving value—from sales research to code commits—so leaders can decide where to invest next.

Key Insights

The admin console breaks ChatGPT usage into predefined use case categories—sales research, coding, planning—so you can see which teams are actually getting value versus just burning credits.

For coding specifically, Codex contributions now show measurable outcomes like merged commits and code review engagement, giving engineering leaders concrete ROI data to justify expansion.

Single conversation — The admin plugin lets you ask follow-up questions across all metrics at once—usage, spend, outcomes, team breakdowns—turning scattered data into a single narrative for leadership reviews.

Want this for every new video OpenAI posts? Brevyd summarizes each upload automatically, the morning it drops.

Deep Dive

The problem OpenAI is solving

As companies roll out ChatGPT org-wide, admins face a hard question: where is this actually working? The new admin console starts simple—showing who's using ChatGPT, which teams are consuming credits and tokens, and where usage is growing. But raw usage numbers don't tell you if the tool is delivering value or just wasting budget. That's why OpenAI added the Insights tab, which classifies user activity into specific work categories—sales research, account planning, coding, and more. This shift from 'how much are we spending' to 'what are we spending on' is the core idea.

Connecting usage to actual outcomes

The breakthrough is linking usage data to business results. In the sales example, the demo shows account research and planning filtered by team, with model, reasoning, speed, and plugin breakdowns. An admin can see if the right tools—like Salesforce plugins—are wired in for account research, and if teams are actually using them or need training. For coding, the console shows Codex contributions directly: commits merged, code review engagement, and velocity over time. This lets engineering leaders move beyond 'are people using it?' to 'is it making them measurably faster?' The data becomes a conversation starter with team leads instead of a spreadsheet.

The admin plugin brings it all together

Rather than jumping between tabs, OpenAI added an admin plugin that treats all these analytics as one dataset you can query conversationally. The demo shows an admin asking ChatGPT to prepare for a monthly rollout review—visualizing Codex and usage trends by team, top use cases, and engineering outcomes—then turning that into a leadership deck with charts and next steps. The plugin reduces friction from 'I need to investigate this' to 'show me a deck.' It's designed for the cadence of real org decision-making: monthly reviews, team syncs, budget allocation cycles.

Takeaways

✓Open your admin console today and audit the top five teams using ChatGPT—check if their plugin stack matches their stated use cases, because that's the fastest way to find training gaps.
✓For engineering leaders, track Codex contributions alongside velocity metrics monthly—merged commits and code review engagement are how you build the ROI case for expanding access beyond your pilot.
✓If you're planning an AI rollout, use the admin plugin to generate one unified dashboard of usage, spend, outcomes, and recommendations rather than assembling it manually—it saves time and keeps leadership aligned on next steps.

Key moments

0:33Use cases breakdown introduced

“user activity is classified into predefined use case categories with a breakdown into specific tasks”

1:20Measurable coding outcomes

“for coding, we can also see measurable outcomes directly in the admin console. So here, I can see Codex contributions to commits, merged code, and code review”

2:36Admin plugin generates leadership deck

“turn this into a short leadership deck with charts, key takeaways, and recommended next steps, and share it in the AI Rollout Slack channel”

You just read one. Brevyd does this for every upload.

Follow OpenAI and every new video comes back as a summary like this, in your morning briefing. No watching required.