Use ChatGPT Work to build your data agent
Summary
ChatGPT Work enables non-technical teams to build data agents by connecting AI to existing tools—Databricks, Slack, Google Drive—using a plugin system that mirrors how you'd onboard a new analyst with context and access to company data sources.
Key Takeaways
- Treat agent setup like employee onboarding: provide the same tools, data sources, and context you'd give a new analyst to ensure the agent can perform the job effectively.
- Use the plugins tab to connect ChatGPT Work to multiple data sources simultaneously—Databricks tables, Slack channels, Google Drive docs—creating a unified data context the agent can query across.
- Have your data team review and refine the auto-generated 'data context skill' before packaging it as a reusable plugin, ensuring accuracy and enabling company-wide distribution.
- Enterprise and business plan users may require admin approval to enable certain plugins, creating a governance layer for data access and agent capabilities.
- The data agent plugin automates cross-source analysis by assembling context from disparate tools, reducing manual data gathering and enabling faster insights for non-technical users.
Related topics
Transcript Excerpt
Um, the way I typically like to think about this is what would I give a new analyst? Like what kind of context would I give a new analyst on the team? Give that to the agent. And so for this analysis, the first thing I'm going to do is make sure that Chat GPT work is using and connected to the same tools that I would use myself to do this job. And so to do that, we can go to the plugins tab, um, which is up here on the top left. Um, the other thing I want to call out here is that I'm also I also have the data uh agent installed here, which Molly just spoke about. Um, and to install any of these, all you have to do is click this plus button and then log into your account if that's required. Um, and if you're on a business or an enterprise plan, you might find that you have to also ask your …