If you’ve noticed your Google leads getting more expensive or your organic traffic softening over the past year, you’re not imagining it, and you’re not alone. The search landscape has fundamentally shifted, and ChatGPT is at the center of it. Now OpenAI has launched a paid advertising platform directly inside ChatGPT, and it opens up a channel that didn’t exist twelve months ago.
ChatGPT Didn’t Kill Google. It Changed How People Use It.
Google’s total search volume hasn’t collapsed. In fact, the number of global digital queries grew. What changed is how people behave before they ever get to Google, and that shift is costing businesses real visibility.
Here’s how the data actually breaks down:
1. Raw Query Volume (The 12% to 20% Footprint)
ChatGPT handles roughly 2.5 billion prompts per day, while Google processes an estimated 13.7 billion to 14 billion daily searches. According to a structural analysis by Ahrefs and Harvard/NBER data, roughly 65% of ChatGPT prompts are “search-equivalent”: queries where a user is looking for information, advice, or facts they traditionally would have Googled.
When isolating just those search-like queries, ChatGPT’s volume sits at roughly 12% of Google’s total. When factoring in mobile app usage, independent digital traffic analyses from Graphite and KIME indicate ChatGPT captures roughly 17% to 19.5% of global search-and-discovery traffic, making it the second-largest information discovery platform in the world.
2. The “Search Behavior” Shift (-20% Fewer Searches)
ChatGPT hasn’t necessarily stolen users away from Google entirely. Most people use both. However, clickstream data from SparkToro and Datos reveals a massive hidden impact: the number of Google searches per active user has dropped by roughly 20%.
Instead of performing five different sequential Google searches to research a complex topic, users go to ChatGPT once to get a synthesized answer, eliminating the need for those extra Google queries.
3. The Website Traffic Paradox (190x Difference)
While ChatGPT is aggressively taking away query volume, it is not taking away the downstream traffic that Google sends to websites.
| Platform | Global Search/Discovery Share | Share of Outbound Traffic Sent to Websites |
|---|---|---|
| ~71% – 80% | ~40.0% | |
| ChatGPT | ~12% – 19.5% | ~0.21% |
Google is built to act as a traffic cop, routing users out to external web links. ChatGPT is built to keep users in a closed conversation. Because ChatGPT’s Click-Through Rate (CTR) is about 96% lower than Google’s, Google still sends 190 times more traffic to actual websites than ChatGPT does.
The Takeaway: ChatGPT has successfully captured nearly a fifth of the world’s informational and research queries, severely threatening Google’s historical monopoly on intent. However, until conversational shopping and outbound link clicks become the norm, Google remains the undisputed king of web traffic.
That’s the gap the OpenAI Ads Manager was built to close.
Suggested image idea: side-by-side screenshot of a Google search results page vs. a ChatGPT conversation answering the same query.
What OpenAI Ads Manager Actually Is
OpenAI Ads Manager Beta is a self-serve platform that lets businesses run paid campaigns directly inside ChatGPT conversations. If you’ve used Google Ads or Meta Ads, the campaign structure will feel familiar: it uses the same Campaign → Ad Group → Ad hierarchy. But the way targeting works is completely different, and that difference matters.
The OpenAI Ads Manager is best used for capturing high-intent users during the consideration and research phases. Because ads appear inside ChatGPT conversations, often after a user has spent several prompts narrowing down what they want (e.g., “Compare the top three CRMs for a small agency”), the traffic it sends is highly qualified. Early performance data shows that while Click-Through Rates (CTR) are lower than traditional Google Search (around 0.6% to 1.5%), the conversion quality is roughly 1.5 times higher once users land on a website.
Key Features
Familiar Interface
Structurally similar to Meta or Google Ads, featuring step-by-step campaign creation and bulk uploads.
Product Feed Integration
Merchants can upload product catalogs via SFTP, enabling ChatGPT to serve direct product ads during shopping queries.
Bidding & Tracking
Supports both CPM and CPC models, alongside pixel tracking and a Conversions API to measure ROI.
Ad Integrity
Ads are clearly labeled as sponsored and kept distinct from organic AI responses. Merchant data is used solely for paid placements, not model training.
There are no keywords. There are context hints.
Instead of bidding on search terms, you write plain-language descriptions, called “context hints,” that tell OpenAI’s AI when your business is relevant. The system reads the underlying meaning of a user’s conversation and decides whether your ad fits, not based on the specific words they typed, but on what they actually need.
When a match is found, your ad doesn’t interrupt the experience. ChatGPT answers the user’s question completely first. Once the response finishes, a clearly labeled “Sponsored” card fades in below the answer: your brand logo, a short headline, a brief text snippet, and a link to your landing page. It reads like a natural next step, not an intrusion.
It’s available now to select advertisers in the US, Canada, Australia, and New Zealand.
How to Get Your Business Into the Beta
Schedule a Free ChatGPT Search Audit
Rather have us handle the setup and strategy? We’ll walk you through whether this platform is the right fit right now.
One thing to know before you start: the business owner must create the account first. Agencies and team members can only be added after the account exists, not during setup.
Suggested image idea: screenshot of the business details entry step (Step 2) at ads.openai.com.
- Access the portal. Go to ads.openai.com and sign in with an existing OpenAI account (use your work email) or create one on the spot.
- Enter your business details (5–10 minutes). Provide your legal business name, website URL, industry type, and a square brand logo at minimum 128×128 pixels. This logo appears directly inside your live ad units, so keep it clean: it renders small, and text overlays become unreadable at that size.
- Submit for verification (processing time varies). Confirm your country, currency, and time zone, select “Business” as your advertiser type, and enter your EIN and physical business address. This kicks off identity and policy compliance review through Persona, OpenAI’s verification partner. Expect anywhere from 2 to 12 days before full access is granted.
- Configure billing (after approval). Once you receive your approval email, log back in to set up a billing profile and link a corporate credit card. Your billing address must match your registration country.
- Add your team or agency (optional). Go to Settings → Users → Invite to bring in any team members or agency partners who’ll be managing campaigns.
One important heads-up: OpenAI has a strict no-refund policy. If the contextual matching misfires and your budget gets spent on the wrong conversations, or if a platform bug causes inefficient delivery, there are no credits. Go in treating your early testing budget as a learning investment, not a performance campaign.
What the Platform Is Good For, and What It Isn’t Yet
The OpenAI Ads Manager is best used for capturing high-intent users during the consideration and research phases.
Because your ad only appears after a user has already spent several prompts narrowing down what they want, think “compare the top three options for a small business like mine,” the people who see your ad are already well-qualified. Early data shows conversion quality running roughly 1.5 times higher than comparable social traffic once users click through. The AI does the qualification work before your ad ever appears.
What it can’t do yet: optimize for conversions.
If you open the dashboard today, you’ll see two campaign objectives: Reach (CPM) and Clicks (CPC). The Conversions objective is locked or labeled “coming soon.” Two reasons for that:
First, the algorithm is still learning. An ad platform needs millions of tracked conversion events across different industries before its machine-learning model can accurately predict who will actually buy. Right now it can reliably optimize for impressions and clicks. That’s it.
Second, OpenAI intentionally doesn’t build deep user profiles. Unlike Google or Meta, which use years of browsing history and demographic data to target conversion-ready audiences, OpenAI matches ads to the live conversation, not to the person. That privacy-first approach makes direct conversion optimization a harder problem. It’s coming, but it isn’t here yet.
Suggested image idea: screenshot of the campaign objective selection screen showing “Reach” and “Clicks” active with “Conversions” grayed out or labeled “Coming soon.”
What Agencies Are Seeing in Early Testing
The early consensus from agencies inside the beta can be summed up in four words: high intent, low control.
The Case for Getting In Now
Despite lower click volumes than Google Search (CTRs are running between 0.5% and 1.2%), the quality of that traffic is strong. Publications like Digiday have reported that users arriving from ChatGPT ads spend more time on-site and bounce less than social traffic. By the time someone sees your ad, the AI has already walked them through their research, and they arrive as a warm lead, not a cold visitor.
CPMs vary by category. High-competition verticals like finance and B2B software are seeing $25–$40 CPMs. Broader lifestyle and service categories are running closer to $12–$18. For context, those are competitive but not unreasonable for the intent level.
The Case for Caution
The control problem is real and worth naming plainly. There are no keyword controls and no negative keyword lists. If your budget is going to the wrong conversations, your only move is rewriting your context hint and waiting to see if the AI interprets it differently. You cannot see which prompts triggered your ads. You cannot exclude specific topics. You are trusting the algorithm completely.
Delivery is also inconsistent right now. You might set a $500 daily budget and the system serves a fraction of it one day and spikes the next. The platform is still finding its footing.
The Strategic Frame Most Agencies Are Using
The brands moving earliest are doing two things: protecting their own brand terms so competitors can’t appear when a user asks ChatGPT “what’s the best alternative to [Brand X],” and buying visibility now while competition is low and prices haven’t spiked. Ad auction costs on new platforms almost always rise as adoption grows. That window doesn’t stay open long.
Where the Platform Stands Today
What’s Built (Available Now)
- The Campaign Infrastructure: A traditional Campaign → Ad Group → Ad hierarchy, familiar to anyone who has used Google or Meta.
- Context Hints & Semantic Matching: Advertisers write plain-language “context hints” and OpenAI’s language model handles conversation matching based on intent. No keyword bidding.
- Flexible Bidding: Supports both CPM and CPC bidding models.
- Product Feed Ingestion: Merchants can upload 1,000–2 million SKUs via SFTP to trigger visual product cards during shopping queries.
- Basic Attribution: A Conversions API and pixel tracking capture raw conversion actions downstream.
- The “Answer Independence” UI: Ads render as clearly labeled sponsored cards tucked below ChatGPT’s finalized organic response.
What’s Missing (Current Pain Points)
- Beta Interface Bugs: Significant friction with validation lags, feed processing errors, and inability to delete old product feeds.
- Granular Performance Metrics: ROAS, Cost-per-Conversion, and Conversion Rate are absent from the native reporting dashboard.
- Major Vertical Restrictions: Healthcare, mental health, politics, gambling, cryptocurrency, and various financial and legal services are off-limits.
- No Keyword/Negative Match Controls: You must fully trust the AI’s contextual matching. Manual keyword exclusions aren’t possible.
What’s Coming (The Roadmap)
- Audience Targeting & Retargeting: Customer list uploads, lookalike audiences, and exclusion targeting are in a gated rollout.
- Advanced Retail Integrations (Agentic Commerce): Instant Checkout via Stripe and PayPal partnerships, allowing purchases directly within the chat.
- Global Expansion: Rollouts to the UK, Japan, Brazil, and broader European markets are actively being prepared.
- Ad-Management APIs: Direct API integration will replace manual SFTP uploads, opening the door to third-party ad-tech tools like Pacvue and Skai.
Should You Be on This Platform?
If you’re looking for a channel that replaces your Google Ads program or delivers clean cost-per-lead data out of the gate, this isn’t that, not yet.
If you’re looking for an early foothold inside the platform where your customers are increasingly forming their buying decisions, at a price point that will almost certainly rise as adoption grows, this is worth a serious look.
We’re actively testing OpenAI Ads across our accounts and will share real performance data as it becomes meaningful. In the meantime, if you want a straight conversation about whether this fits your current strategy, we’re here for it.
This Post Will Keep Evolving
OpenAI Ads Manager is still in beta, which means the numbers in this post (CPMs, CTRs, feature availability) are a snapshot, not a permanent record. Betas move fast. Features that are locked today might unlock next month. Stats that look solid right now will shift as more advertisers get in and more data comes in.
We’re tracking this closely, including our own testing across our accounts, and we’ll be updating this post as new information, stats, and platform changes come out.
Ready to Talk Through Your Strategy?
We’ll walk you through whether OpenAI Ads is the right fit for your business right now. No pressure, just a straight conversation.
Tags: Our Favorites • AI Search
Daniel Falco
Daniel is a digital marketing strategist at TheeDigital. He is an expert in digital advertising and small business marketing strategy, and he aims to fully understand his client’s business and industry to find unique competitive advantages.
