Skip to main content
AI in ASIA
ChatGPT Memory
Business

ChatGPT Just Quietly Released "Memory with Search" - Here's What You Need to Know

ChatGPT's new Memory with Search feature personalizes results using your conversation history and preferences, threatening traditional SEO.

Intelligence Deskโ€ขโ€ข4 min read

AI Snapshot

The TL;DR: what matters, fast.

ChatGPT reached 365 billion annual searches in 2024, 5.5x faster than Google's adoption rate

Memory with Search personalizes results using conversation history and explicit user preferences

Feature threatens traditional SEO as AI-driven personalization becomes standard practice

Advertisement

Advertisement

OpenAI Quietly Launches Memory-Powered Search That Knows You Better Than Google

OpenAI has rolled out Memory with Search for ChatGPT, a feature that fundamentally changes how AI handles your queries. Instead of treating each search as isolated, the system now leverages your saved preferences and conversation history to rewrite prompts into more personalised search results.

The feature works by accessing two key data sources: details you explicitly ask ChatGPT to remember about yourself, and insights the AI infers from your past conversations. When you submit a search query, ChatGPT quietly rewrites it using this contextual knowledge before fetching results.

For businesses across Asia, this represents a seismic shift in how customers discover products and services. Traditional SEO strategies may need fundamental rethinking as AI-driven personalisation becomes the norm rather than the exception.

By The Numbers

  • ChatGPT reached 365 billion annual searches in 2024, a milestone Google took 11 years to achieve
  • The platform now serves 831 million users with 5.7 billion monthly visits as of March 2026
  • ChatGPT holds 60.4% of the AI search market share, excluding Microsoft Copilot
  • 46% of ChatGPT queries now use the web search feature according to Semrush analysis
  • User sessions per person grew 195% in 2024, from 0.43 to 1.27 sessions
"ChatGPT has reached 365 billion annual searches in 2024, a milestone that took Google 11 years to achieve in 2009. That's a 5.5 times faster rate of adoption for AI-based tools."
Mary Meeker, Trends - Artificial Intelligence Report

How Memory Changes Your Search Experience

Memory with Search transforms ChatGPT from a stateless question-answering tool into a personalised research assistant. When you search for "best restaurants," the AI might factor in your previously mentioned dietary restrictions, location preferences, or budget considerations.

The system operates on two levels of memory storage. Explicit memories include details you directly tell ChatGPT to remember, such as your role, interests, or personal circumstances. Implicit memories are patterns the AI identifies from your conversation history, like your communication style or recurring topics of interest.

This creates a fundamentally different search experience compared to traditional engines. Rather than returning generic results, Memory with Search provides contextually relevant information tailored to your specific needs and preferences. Users who frequently discuss sales strategies might receive more detailed workflow suggestions, whilst those focused on professional development could see career-oriented resources.

"With SearchGPT enabled, the intent distribution more closely mirrored that of Google, with a noticeable increase in navigational, commercial, and transactional searches."
Semrush Analysis Team, Search Behaviour Report

What This Means for Asian Businesses

The implications for brands operating in Asia-Pacific markets are profound. Traditional SEO strategies built around broad keyword targeting may lose effectiveness as AI systems deliver increasingly personalised results based on individual user memories.

Companies need to consider how their content aligns with personal preferences rather than just search algorithms. This shift mirrors broader changes in how AI is quietly redesigning digital experiences across various sectors, from food delivery to e-commerce.

The following table illustrates how Memory with Search differs from traditional search approaches:

Aspect Traditional Search Memory with Search
Query Processing Keywords and ranking factors Personal context plus keywords
Result Relevance General population signals Individual preference history
Brand Visibility SEO optimisation Personalisation alignment
User Experience Same results for similar queries Unique results per user

Businesses should focus on creating content that anticipates personal use cases rather than generic search terms. This includes developing materials that address specific professional roles, industry challenges, or personal interests that users might have shared with ChatGPT previously.

Privacy Controls and Strategic Content Implications

OpenAI has included privacy controls that allow users to manage their memory preferences. The system provides granular control over what information gets stored and how it influences future interactions.

Users can disable memory functionality entirely or selectively remove specific stored information. The privacy settings are accessible through ChatGPT's personalisation menu, where the "Reference saved memories" toggle can be switched off.

Key privacy management options include:

  1. Complete memory disable through the personalisation settings panel
  2. Selective memory deletion for specific topics or conversations
  3. Temporary memory pause for sensitive discussions
  4. Memory review functionality to audit stored information
  5. Export options for personal data transparency

Content creators and marketers face a fundamental shift in how their materials get discovered and consumed. Memory with Search means that the same piece of content might be presented differently to various users based on their stored preferences and conversation history.

Rather than optimising for broad keyword searches, creators need to consider how their content serves specific user contexts and personal situations. Content that demonstrates practical AI implementation strategies may perform particularly well, as users with technical interests stored in their memory profiles receive more targeted recommendations.

Frequently Asked Questions

How does ChatGPT decide which memories to use in searches?

ChatGPT evaluates the relevance of stored memories to your current query, incorporating details that could make the search results more useful and contextually appropriate for your specific situation and preferences.

Can I see what memories ChatGPT has stored about me?

Yes, you can review your stored memories through ChatGPT's settings menu under personalisation options, where you can view, edit, or delete specific memory entries as needed.

Will Memory with Search work for business accounts?

Memory functionality varies by account type and organisational settings. Business administrators can typically control memory features through their enterprise dashboard to ensure compliance with company data policies.

How does this affect my search privacy compared to Google?

Memory with Search stores personalisation data within ChatGPT's system rather than across multiple services, but users should review OpenAI's privacy policy to understand data handling practices fully.

Can Memory with Search replace traditional search engines?

While Memory with Search excels at personalised queries, traditional search engines still provide broader web indexing and real-time information that ChatGPT's system may not capture completely.

The AIinASIA View: Memory with Search represents more than a feature update; it signals the beginning of truly personalised information discovery. Whilst the technology offers unprecedented customisation, we're concerned about the potential for information bubbles and reduced serendipitous discovery. Asian businesses should start experimenting with memory-aware content strategies now, but must also consider the broader implications of AI systems that know users intimately. The key question isn't whether this technology will succeed, but whether society is prepared for search engines that remember everything.

The rollout of Memory with Search fundamentally changes how we interact with information online. As AI systems become more sophisticated at understanding and remembering user preferences, the line between search and conversation continues to blur.

This shift towards memory-enhanced AI interactions extends beyond search into areas like workplace productivity and customer service, suggesting a broader transformation in how we engage with digital tools.

How do you feel about AI systems that remember your preferences and tailor results accordingly? Does the convenience outweigh potential privacy concerns, or are we moving too quickly towards hyper-personalised digital experiences? Drop your take in the comments below.

โ—‡

YOUR TAKE

We cover the story. You tell us what it means on the ground.

What did you think?

Written by

Share your thoughts

Join 3 readers in the discussion below

This is a developing story

We're tracking this across Asia-Pacific and may update with new developments, follow-ups and regional context.

Advertisement

Advertisement

This article is part of the AI in ASEAN Markets learning path.

Continue the path รขย†ย’

Latest Comments (3)

Sarah Chen
Sarah Chen@sarachen
AI
10 January 2026

leverage relevant information from memories to make the query better and more useful." this raises immediate concerns about data provenance and how "useful" is actually defined. what's the control group for these "better" queries?

Nicolas Thomas
Nicolas Thomas@nicolast
AI
9 June 2025

this "memory with search" feature, I'm a bit torn. on one hand, personalising results could be helpful, but on the other, it feels a bit like a black box. "chat history insights"-what exactly is ChatGPT inferring? and how does that influence the search? open-source models, especially what we're seeing from some european labs, often give developers and users more transparency into how these memory layers work. for brands thinking about SEO, it's not just about anticipating prompts, but understanding the underlying mechanisms of these memory functions too, otherwise it's just guessing.

Elaine Ng
Elaine Ng@elaineng
AI
19 May 2025

the idea of "chat history insights" being leveraged for search without explicit user input raises some interesting questions around algorithmic bias and surveillance. how transparent will openai be about what inferences are actually being made and used? this feels like another layer of data extraction that's not fully articulated.

Leave a Comment

Your email will not be published