7 Hidden ChatGPT Features You Probably Are Not Using

Most people use ChatGPT the same way: open a new chat, type a question, read the answer, close the tab. That works. But there's a layer of features underneath the basic chat interface that most users have never touched — settings, modes, and tools that change what ChatGPT can do for you on a daily basis.

I found several of these by accident, others by actually reading the documentation (which, in fairness, most people don't). This is a tour of the seven features I think matter most for anyone using ChatGPT seriously, with specific examples of how to use each one.

Feature 1: Custom Instructions — Your Personal AI Briefing

Custom Instructions are the feature with the highest impact-to-setup-time ratio on this list. Once configured, they inject your context and preferences into every conversation automatically. No more typing the same background paragraph.

You'll find it at: Settings > Personalization > Custom Instructions

There are two text fields:

"What would you like ChatGPT to know about you?" This is where you put context: who you are, what you do, what you're working on, your expertise level in different areas, your industry.

Example:

"I'm a senior product designer at a B2B SaaS company. My primary work involves UX research, interaction design, and design system management. I work primarily in Figma. My stakeholders are non-technical product managers and engineers. I have a strong background in accessibility. I'm an expert in design but a beginner in SQL."

"How would you like ChatGPT to respond?" This is where you control output: tone, format, length, things to avoid.

Example:

"Be direct. No hedging or excessive qualifications unless genuinely important. Prefer bullet points over paragraphs for instructional content. Skip the pleasantries at the start of responses. If I ask for code, use comments to explain non-obvious parts. Never use the phrase 'Great question!'"

Once set, these instructions shape every chat. The quality of responses improves noticeably because ChatGPT isn't starting from zero about who you are.

For more on how Custom Instructions interact with the memory feature, the ChatGPT memory settings guide covers the two systems and how they work together.

Feature 2: Memory — The Persistent Context Layer

Memory is different from Custom Instructions. Custom Instructions are static — you write them, they stay the same until you change them. Memory is dynamic — ChatGPT builds it up over time from your conversations, saving facts it thinks are relevant.

Find it at: Settings > Personalization > Memory

When memory is on, ChatGPT might save things like: "User is working on a podcast about supply chain" or "Prefers bullet points for technical content" or "User is learning TypeScript."

The key to using memory well is being intentional:

• Tell it explicitly what to remember: "Remember that my client's company is in the healthcare SaaS space"
• Audit what's saved regularly: Settings > Personalization > Manage Memory
• Delete anything outdated or irrelevant

Memory doesn't work across custom GPTs unless those GPTs explicitly enable it. In the main ChatGPT interface, it's one of the most useful background features once you start relying on it.

Feature 3: Temporary Chat — The Incognito Mode You Didn't Know Existed

When you start a regular ChatGPT conversation, it gets saved to your history and can influence memory. Sometimes you want the opposite — a conversation that vanishes afterward with no record.

Temporary Chat is ChatGPT's version of browser incognito mode. Access it by clicking your username at the top of the sidebar — you'll see the "Temporary Chat" option.

In Temporary Chat:

• Nothing is saved to your history
• Nothing influences memory
• The conversation disappears when you close it

Useful scenarios:

• Helping someone else use your account without mixing their session into your profile
• Working with sensitive personal topics you don't want in your history
• Exploring ideas you're not ready to commit to your regular workflow
• Getting a fresh perspective without prior memory context affecting the answer

I use it whenever I'm testing prompts for work that involves client information, since it means nothing gets stored inadvertently.

Feature 4: Voice Mode — More Than a Novelty

Most users who try voice mode use it once, say "huh, that's interesting," and go back to typing. That undersells what voice mode actually does well.

The standard voice mode (available to Plus users on mobile and desktop) lets you have a spoken conversation with ChatGPT. The Advanced Voice Mode (available on Pro plans) is meaningfully different — it processes your actual voice tone and cadence, not just the transcription, which makes interruptions and natural conversation flow work properly.

Where voice mode is genuinely useful:

Thinking out loud: Some ideas develop better when spoken than typed. Starting a conversation in voice mode and just narrating your thinking — "I'm trying to figure out whether to expand our product line or double down on our core offering" — and having ChatGPT respond conversationally can unlock different thinking than typing a structured question.

Hands-free research: On mobile, voice is often faster than typing. Asking research questions, getting quick definitions, or exploring topics while commuting or exercising is a real workflow.

Practicing presentations or arguments: Explain your idea to ChatGPT out loud, ask it to play devil's advocate, and practice responding to pushback. The conversational format makes this feel more like rehearsal than writing exercises.

Language learning: Practicing conversation in another language with real-time responses is something voice mode handles well — better than most dedicated language apps for speaking practice.

Feature 5: GPT Builder — Create Your Own Specialized Assistant

GPT Builder (available to Plus and above) lets you create custom versions of ChatGPT with specific personas, knowledge, and capabilities. These are called GPTs, and you can build one in about fifteen minutes without any coding.

Find it at: Explore GPTs > Create in the sidebar, or at chatgpt.com/gpts/editor.

The Builder interface is conversational — you tell it what you want your GPT to do, and it configures itself. You can:

• Give it a specific persona and expertise area
• Upload documents it should reference (PDFs, text files, spreadsheets)
• Set capabilities: web browsing, image generation, code interpreter
• Restrict what it talks about
• Add custom actions (API connections for advanced users)

Practical GPT ideas:

• A GPT trained on your company's documentation that answers employee questions
• A writing assistant that matches your brand voice and style guide
• A code review assistant with your team's specific conventions loaded
• A customer support agent trained on your product FAQs
• A research assistant for your specific industry

You can keep GPTs private, share them with a link, or publish them to the GPT Store. For work purposes, private GPTs are the most immediately practical.

The ChatGPT plugins guide covers how to extend GPTs further with external integrations.

Feature 6: Data Controls and Privacy Settings

Most users have never opened the data controls panel, which means they're running on default settings they may not agree with.

Find it at: Settings > Data Controls

Key settings here:

Improve the model for everyone: This controls whether your conversations are used to train future ChatGPT models. It's on by default. If you share confidential work information in chats, you may want to turn this off. Go to Settings > Data Controls and toggle "Improve the model for everyone" off.

Shared links: ChatGPT lets you create shareable links to conversations. This setting controls whether shared conversations can be used for training. Worth reviewing if you share conversations externally.

Export data: You can download all your ChatGPT conversation history as a JSON file. Useful if you've ever wanted to search through past conversations or back up your prompt experiments.

Delete all conversations: A nuclear option if you want to start fresh. This removes your history but does not remove memories (that's a separate deletion).

Understanding these settings matters because they affect what happens to the information you share with ChatGPT. Taking ten minutes to configure them intentionally is time well spent.

Feature 7: Conversation Branching — Editing Your Way to Better Answers

This one surprises most users: you can edit any previous message in a conversation and create a branch. The original conversation continues to exist, and the new branch is a parallel conversation from the edit point forward.

How to use it: Hover over any of your previous messages and click the pencil/edit icon. Change the message and submit. A new response generates, and you can navigate between versions using the arrow controls that appear.

Why this is useful:

Refining prompts without starting over: If ChatGPT gives you a mediocre response, instead of asking "can you try again?", you can go back and edit your original question with better context. The new version often performs significantly better because you're giving it a cleaner starting point.

A/B testing prompts: Write a prompt, see the response, branch it with a different version of the prompt, compare results. This is how you learn what prompting approaches work for specific types of requests.

Saving the good version: If you had a conversation going well and then took it in a direction that didn't work, you can branch back from before things went wrong instead of re-doing the whole session.

Building prompts iteratively: Start with a rough question, see what's missing from the response, edit the original question to include that context, see the improvement. Repeat until you've found the prompt structure that works.

This feature is underused because it's not obvious — you have to hover over messages to see the edit option. Once you know it's there, it changes how you work through complex conversations.

For a complete breakdown of prompting techniques that work well alongside these features, the prompt engineering guide and ChatGPT prompt bible cover structured approaches that compound when combined with Custom Instructions and GPTs.

Putting It All Together

These seven features aren't independent — they work better together. Custom Instructions set your baseline context. Memory builds on that context over time. GPTs let you create specialized assistants for recurring work. Temporary Chat handles sensitive sessions. Voice mode handles situations where typing is inefficient. Data controls let you manage privacy intentionally. Conversation branching lets you refine prompts without friction.

A power user setup might look like: Custom Instructions configured with your professional context, memory enabled and audited monthly, one or two custom GPTs built for your highest-frequency tasks, data controls reviewed and set deliberately.

Most users set up none of this. The people who spend thirty minutes on initial configuration get a noticeably different experience from everyone else — not because they have a better plan, but because they've taken the tool off its default settings.

Conclusion

ChatGPT's default experience is deliberately simple — open a chat, ask a question. That simplicity is a feature, not a bug, for new users. But it means that the configuration layer, which significantly changes what the tool can do, stays invisible to most people indefinitely.

The seven features covered here — Custom Instructions, Memory, Temporary Chat, Voice Mode, GPT Builder, Data Controls, and Conversation Branching — are all worth exploring in order. Start with Custom Instructions because it has the highest immediate payoff. Spend ten minutes filling in both fields tonight. Then try a conversation the next morning and notice whether the baseline quality of responses has shifted.

The ChatGPT vs Claude comparison is worth reading if you're curious how these kinds of customization features compare across different AI platforms.

Further Reading

• ChatGPT API Tutorial: Build Your First AI-Powered App in 1 Hour
• ChatGPT for Lawyers: How Legal Professionals Use AI Ethically
• ChatGPT for Language Learning: Become Fluent Faster