3 Min Read

ChatGPT Now Connects to Your Business Tools

Featured Image

Wondering how to get started with AI? Take our on-demand Piloting AI for Marketers Series.

Learn More

OpenAI just made some big updates to ChatGPT...especially if you're a business user.

The update getting all the attention? Connectors within ChatGPT. 

These let ChatGPT plug directly into popular business apps like Google Drive, Dropbox, SharePoint, Box, and HubSpot. With a few clicks, teams can pull live data from these tools straight into a chat and use it to search, summarize, and generate content.

Want to find last week’s roadmap? Summarize Q2 goals? Analyze project pull requests? Just ask ChatGPT

On Episode 152 of The Artificial Intelligence Show, I unpacked OpenAI’s latest announcements with Marketing AI Institute founder and CEO Paul Roetzer. And let’s just say:

The excitement is palpable. And so is the caution.

ChatGPT Wants to Be Your New Operating System

This is more than a nifty integration. It’s a clear signal that OpenAI envisions ChatGPT as an all-encompassing work platform. One where your data, tools, and documents live inside the chat window, not scattered across disconnected apps.

"They don't want you to leave ChatGPT," says Roetzer. "They want you to just connect to everything you have access to and talk to it right there."

They seem to be getting their wish. According to OpenAI COO Brad Lightcap, the company now boasts 3 million paying business users, up from 2 million just a few months ago. That includes ChatGPT Enterprise, Team, and Edu customers across industries like finance, healthcare, and retail.

Security, Governance, and a Growing Complexity Crisis

But as enticing as the new connector feature sounds, Roetzer was quick to pump the brakes. While connectors promise a huge leap in productivity, they also introduce serious governance concerns.

"My immediate response [to our team] was: Do not connect this to anything," he says.

Why? Because once connected, ChatGPT can access, retrieve, and potentially remember sensitive data from those apps. And unless you're on the Enterprise plan, your admin controls might be limited.

In fact, Roetzer discovered that even on the Team plan, there were few safeguards in place to prevent users from connecting confidential sources like Google Drive or HubSpot without oversight.

"If I have HR data or confidential information that only a select few people in the organization have access to, how do I know that's not going to end up in chat?" he asks.

"How do I know someone can't just literally say 'Send me the salary information for all the employees'?"

And that’s just the start. For agencies or companies handling third-party data, especially client information, these connectors could violate existing privacy agreements or contractual obligations.

Add to that a lack of clarity around what data ChatGPT "remembers" and whether it's used to train models (spoiler: it might be, if you're on the Free, Plus, or Pro plans with model training enabled), and the risks become all too real.

Roetzer’s advice:

Slow down. Assign clear ownership over connector testing. Update your AI policy. And most importantly, don’t treat this like just another shiny new feature.

"I think you have to have someone own this from a governance perspective," he says.

Also making waves is a newly announced record mode, which lets users transcribe meetings and auto-generate follow-ups through OpenAI's Canvas tool. Currently rolling out on macOS for Teams users, this feature could become indispensable for operations and project teams alike.

But again, the same cautions apply. What data is being recorded? Where is it stored? Who has access? As with connectors, the ease of use might mask a host of downstream data risks.

The Bigger Picture: Speed, Policies, and Staying Ahead

OpenAI's aggressive rollout strategy underscores a broader truth: This technology is evolving fast. Features like connectors and record mode don't just change workflows—they redefine the expectations for knowledge work.

"It'll probably become ubiquitous throughout the enterprise," says Roetzer. "You're going to connect your AI models to these outside sources. It's going to enrich all these use cases. But pump the brakes a little bit."

For organizations without a formal AI council or written AI policy, these launches should be a wake-up call. It's no longer optional to have guardrails. It's essential.

The Bottom Line

ChatGPT's latest features promise faster insights, more seamless workflows, and a radically reimagined way to get work done.

But they also demand a new level of vigilance.

So yes, go ahead and explore. Just make sure someone on your team is asking the tough questions before you connect the dots—or the drives.

Related Posts

OpenAI's o3 Just Beat Humans at Reasoning: Here's What That Means for Your Career

Mike Kaput | January 7, 2025

OpenAI just announced a brand-new model that may have just crossed a major threshold in AI capabilities—and it has everyone talking.

Inside OpenAI's Blueprint to Reshape America's AI Future

Mike Kaput | January 21, 2025

OpenAI just dropped a sweeping “Economic Blueprint” that sketches out how the United States can—and should—take the lead on AI infrastructure and policy.

OpenAI Cancels Its For-Profit Plans

Mike Kaput | May 13, 2025

OpenAI just pulled off a dramatic pivot—and it could have profound implications for the future of AGI, AI funding, and even global social systems.