Roughly one in five consumer messages sent to ChatGPT are now requests to find information — a share that's crept up to nearly match the tool's original bread and butter: writing help. The growth has slowed compared to earlier snapshots, but the trend line is clear. People are starting to treat ChatGPT less like a drafting assistant and more like a search bar.
The data comes from usage pattern analysis rather than a single press release, and it tracks what categories of questions and tasks people are actually typing into the consumer version of ChatGPT. Writing assistance — emails, essays, social posts — has long been the dominant use case. Information lookup, the kind of task that used to send people to Google, is now running close behind it.
This isn't a sudden jump. Earlier measurements showed information-seeking climbing steadily as a share of total usage, and the latest numbers show that climb leveling off rather than accelerating. In plain terms: the shift toward using ChatGPT as a research and lookup tool has largely already happened. What's left is consolidation, not a new wave.
OpenAI hasn't framed this as a strategic pivot in public statements, but the usage data lands at a moment when the company has been steadily adding search-like features — real-time web browsing, shopping results, and more structured answer formats. The behavior shift and the product changes appear to be reinforcing each other.
Why it matters
This fits a pattern that's been building across the AI tool landscape for roughly two years. Perplexity built its entire pitch around being an answer engine rather than a chatbot. Google rolled out AI Overviews directly into search results, partly in response to exactly this kind of behavior migration. Microsoft folded Copilot into Bing for the same reason. The common thread: AI chat tools and search engines are converging into the same product category, and each company is racing to make sure it ends up owning that category rather than losing traffic to a competitor.
What's genuinely new here isn't that people use ChatGPT to look things up — that's been true for a while. It's that the behavior has matured enough to show up as a stable, measurable chunk of usage rather than a growing trend. That stability is usually the signal that precedes monetization moves: ads, sponsored results, or paid placement tend to follow once a usage pattern proves durable rather than novel.
What this means for small businesses
If a meaningful slice of consumers are asking ChatGPT questions instead of Googling them, the old playbook of ranking on page one of search results covers less ground than it used to. Business owners who rely on SEO for visibility are now also competing to be the source an AI model cites or paraphrases when someone asks it a question about local services, product comparisons, or recommendations.
There's no established toolkit for this yet, the way there is for traditional SEO. But the raw material is similar: clear, well-structured information on your website, accurate business listings, and content that directly answers common customer questions tend to get pulled into AI-generated answers more often than vague marketing copy.
The trade-off is uncertainty. Unlike Google rankings, there's no dashboard showing how often your business gets mentioned in a ChatGPT response, and OpenAI doesn't publish referral data the way search engines historically have. Businesses are optimizing somewhat blind right now.
What to watch
Watch for two things: whether OpenAI introduces advertising or sponsored placement within ChatGPT's answers, and whether any analytics tool emerges that tracks AI-driven referral traffic the way Google Search Console tracks search traffic. Either development would signal this usage shift is moving from a behavioral trend into a monetized channel businesses need to actively manage.
The bottom line
ChatGPT's role as a lookup tool is no longer a fringe use case — it's now roughly on par with its original function as a writing aid. Businesses that depend on being found online have one more channel to think about, even though the tools to measure and optimize for it are still catching up to the behavior itself.