If you've ever had to re-explain your business to ChatGPT at the start of a new conversation, that problem may be about to disappear.
OpenAI is rolling out a significant upgrade to ChatGPT's memory system. Starting this week, ChatGPT can now reference all of your past conversations, not just information you explicitly asked it to save. The update went live for Pro subscribers ($200/month) first, with Plus users ($20/month) getting access "soon" and Team, Enterprise, and Edu accounts following within a few weeks, according to OpenAI.
This is a meaningful shift in how the tool works.
Two Kinds of Memory, One Smarter Chatbot
Before this update, ChatGPT had one memory mode: you could ask it to remember specific things, and it would. Useful, but limited. You had to know what to save.
The new system adds a second mode called "reference chat history." In this mode, ChatGPT automatically scans your past conversations for relevant context, even from exchanges you never flagged for memory. If you mentioned three months ago that you run a 12-person catering company in Austin, it will know that the next time you ask about staffing strategies. You don't have to tell it again.
OpenAI CEO Sam Altman described the goal on X: building "AI systems that get to know you over your life."
For small business owners who use ChatGPT as a regular work tool, the practical upside is significant. Every conversation you've had about your business, your customers, your pricing, your competitors, your writing style, your preferences, becomes part of the AI's working knowledge of you. Over time, the tool gets more useful, not less.
What You Can Do Right Now
If you're on a ChatGPT Pro or Plus plan, check your settings this week to confirm the upgrade has rolled out to your account. Once active, you'll see both "Saved Memories" and "Reference Chat History" options under personalization settings.
A few things worth doing once the feature is live:
Review what's being retained. ChatGPT will show you what information it has saved. It's worth scanning this list, especially if you've discussed anything sensitive about your business, clients, or finances in past chats.
Test it with context. Start a new conversation and ask ChatGPT to recall something specific from a past session. See what it remembers and how accurately it applies that context to a new request. This will tell you how much your previous conversations have actually been useful training data for future interactions.
Consider what you share going forward. The memory feature is opt-in and can be disabled entirely. If you use ChatGPT for client work or sensitive planning, decide how much context you want it retaining across sessions.
The Fine Print
The memory upgrade is not available in the EU, UK, Switzerland, Norway, Iceland, or Liechtenstein. OpenAI has not committed to a timeline for those regions, likely due to data retention regulations that complicate long-term memory storage.
If you want to opt out entirely, you can turn off memory in settings or use ChatGPT's Temporary Chat mode, which leaves no trace across sessions.
This is not the first AI assistant to make this move. Google rolled out a similar long-term memory upgrade for Gemini in February 2026. OpenAI is now catching up, and the competition between these two platforms for the role of "AI that knows your business" is accelerating.
The Bottom Line
The more you've invested in ChatGPT over the past year, the more this upgrade pays you back. Every conversation you've had is now potential context. The tool you've been training, without realizing it, is now able to use what it learned.
For small business owners, that's worth paying attention to.
Sources: OpenAI blog, The Verge, Sam Altman on X