In a fresh development, this doesn't affect our editorial independence. When you purchase through links in our articles, we may earn a small commission.
The report highlights that of course, taking ChatGPT up on its offer would inevitably lead to another “Want me to…” pitch, making them feel like transparent ploys to extend ChatGPT conversations. Once upon a time, ChatGPT seemed to end every response with “Want me to…” follow-up questions, suggesting there was just one more thing it could do to make its reply complete.
According to the latest update, it dialed down the “Want me to…” questions while dialing up “If you like…” teases instead (e.g., “If you like, I could draft a list of possible topics for your fresh podcast”). Initially, I thought the change was good, figuring that silently passing on an “If you like…” suggestion was easier than ignoring a direct “Want me to…” question. But all the “If you like…” pitching got tiresome, too. Eventually, ChatGPT changed its tune.
The report highlights that claude likes to follow up with interactive questions and step-by-step modules, as well as “If you do this, I’ll do that”-style suggestions (e.g., “If you tell me your general idea, I can help sharpen the concept”). Gemini’s follow-ups come with one-click buttons like “Plan a beginner chart home budget” or “Help outline a 3-episode arrival roadmap.”. It’s worth nothing that ChatGPT isn’t alone in making these end-of-reply teases.
As part of the ongoing story, i gave it a shot myself, enlisting Claude Opus 5 to help craft a prompt that keeps those follow-up teases at bay. Here’s the prompt:. Naturally, all the obnoxious AI follow-ups have led to plenty of efforts to stop them, some more successful than others.
As part of the ongoing story, no “Want me to…?”, no “Would you like me to…?”, no “Let me know if…” — and no reworded version of the same move, like “Happy to expand on any of this.”. Don’t end your replies by pitching more work.
In a fresh development, if there’s a useful next step, state it as a fact and stop: “The third section still needs a source.” Don’t turn it into an offer.
Industry observers note that if you need something from me to finish the task correctly, ask for it where you get stuck — not as a sign-off.
In a fresh development, this prompt is designed to stop end-of-reply suggestions that exist only to extend the conversation, while allowing genuine follow-ups when the AI truly needs more information or is otherwise stuck.
Industry observers note that i tested the prompt with three of the most common everyday AI models: ChatGPT Instant, Claude Sonnet 5, and Gemini 3.7 Flash, using five different test prompts (like “I want to start a podcast, where do I begin?” and “Help me think through what to pack for a week in Lisbon”) that invite those “If you like…”-style follow-ups.
Industry observers note that chatGPT, for example, teased its scheduled actions functionality during our podcasting conversations (“Once you settle on the concept, I can also help you remember a arrival milestone or come back to the project later”), while Claude hit me up with follow-up queries during a smart home chat (“Want me to help pick specific products once you’ve settled on an ecosystem, or want more detail on any one category?”). Gemini ended most of its answers with one or more clickable UI “chips,” like one about taking up a jogging routine (“Build a beginner strength routine for runners”). In my control runs with the test prompts, all three chatbots pitched me with “Want me to…”-style pitches on almost every response.
The report highlights that when I started my AI chats with the “no more teasers” prompt (“Understood, I’ll skip the closing offers and sign-off questions,” ChatGPT meekly confirmed), the follow-ups almost completely disappeared.
The report highlights that claude’s follow-up was truly relevant to the conversation (“What’s your current phone/ecosystem (iPhone vs Android) — that’s the one piece of info that changes the recommendation”), while Gemini dropped the clickable UI buttons entirely. Instead of capping our podcast conversation with more questions, ChatGPT ended with a piece of advice (“Don’t buy equipment yet; if those four lines produce a compelling show, then build the production setup around it”).
The report highlights that for example, I added the prompt to ChatGPT’s “Custom instructions” in its Personalization settings, and it completely suppressed follow-ups across all of my chats. You can use the “no more teasers” prompt at the beginning of any AI chat, or you can add it to a chatbot’s system instructions.
As part of the ongoing story, his coverage of artificial intelligence interrogates the most recent LLMs, and how they can be used at work and at home to be best prepared for the AI revolution. “AI is going to change our lives sooner than we think,” Ben writes. “Our best way to adapt is by using it every day.” Ben has been a PCWorld author since 2014, and has covered everything from laptops to security cameras before launching PCWorld’s AI beat. Ben's articles have also appeared in PC Magazine, TIME, Wired, CNET, Men's Fitness, Mobile Magazine, and more. Ben holds a master's degree in English literature. Ben has been writing about consumer technology for more than 20 years, and now focuses his reporting on AI as it relates to the basic human experience.