OpenAI has introduced cost-per-click (CPC) ads inside ChatGPT, a pricing model shift that, on paper at least, places it in the same conversation as Google Search and Meta. Advertisers can now bid between $3 and $5 per click, moving beyond the CPM model that launched the pilot. But unlike a search result or a social media ad, no one is quite sure what a click inside a chat interface actually signals or whether it will be worth the spend.
The announcement lands at a pivotal moment for performance marketing. India’s digital ad market grew 19% in 2025 to reach Rs 71,621 crore and is projected to hit Rs 98,034 crore by 2027, accounting for nearly 70% of total advertising spend, per Dentsu’s Digital Advertising Report 2026. WPP’s This Year Next Year report puts digital at 68.1% of India’s total ad revenue in 2026, with commerce-led formats projected to grow 24.2%. Performance advertising is the engine behind much of this growth, and the question is whether ChatGPT can eventually become part of it.
The death of the classic marketing funnel
On Google, CPC works because intent precedes the click. A user searches for something specific, clicks a result, and the advertiser pays for that moment. Google Search campaigns average around $2.69 CPC across industries. On Meta, the average CPC sits considerably lower, around $0.62, based on data. This is because users are browsing a feed, not looking for something specific. That gap in intent is what determines platform value, and it is where ChatGPT’s CPC bids of $3 to $5 will eventually be tested against actual outcomes.
Chandan Bagwe, Founder and Director of C Com Digital, believes, “OpenAI is basically saying they’re ready to compete with Google and Meta in the big leagues by moving to CPC. It’s not just about ‘cool AI tech’ anymore; it’s about getting a return on investment.”
In fact, OpenAI’s ad pilot’s US rollout crossed $100 million in annualised revenue within six weeks, with over 600 advertisers signed up, reports suggest. But initial CPM rates dropped from roughly $60 at launch to as low as $25 in some cases. When impression prices fall, revenue per impression follows. CPC gives OpenAI a path to ad revenue growth that does not depend on holding CPMs up.
Vishal Kumar, VP Digital Business, North & East at Madison Media, points out what makes the format different from both Google and Meta. “A click inside ChatGPT is much more valuable than a click on Google because the hard work of researching and selecting options has already been done by the language model; at the point of the click, the user isn’t discovering something but purchasing it.”
According to Kumar, this signals “the death of the classic marketing funnel, all condensed down into a single chat session; a point advertisers are well aware of, as those that recognise the value won’t hesitate to pay a premium.”
If that holds in practice, awareness, consideration, and conversion collapse into a single conversational exchange, which is a shift from how performance budgets are currently spread across channels.
Vaishnavi Bala Kumaran, Co-founder of BrandStory Digital, acknowledges the opportunity but flags what makes it difficult to quantify. “Unlike Google, where intent is direct and keyword-driven, conversational AI captures intent as it unfolds. This makes it more nuanced and in many ways more valuable, but also harder to define in conventional performance terms.”
What proof points will actually move budgets
The advertisers currently testing ChatGPT ads include Williams-Sonoma, Target, The Knot, DSW, and BEHR, per reports and are largely in the exploratory phase. OpenAI described its rollout as intentionally cautious while emphasising relevance and privacy safeguards.
Before meaningful budget reallocation happens, advertisers will need more than early pilot signals. Pulkit Dubey, Director and Co-founder of PROHED, comments, “If ChatGPT places the right contextual ads, it could be less chaotic than Google, and the ad clicks might be very premium.” He draws on post-click behaviour as an example. Meta traffic campaigns often see users spend 15 to 20 seconds on a landing page, while Google Search users tend to stay 35 to 40 seconds or more, a difference that reflects the intent quality behind each click. Well-placed conversational ads, he argues, could match or exceed that threshold.
Kumar’s compressed funnel argument also points out that if the model has done the research by the time a user clicks, a poorly placed or irrelevant ad becomes more disruptive than a misplaced banner, not less.
Kumaran outlines what brands will need to see before shifting spend.
“Brands will look for clear indicators such as quality of conversions, assisted impact across the journey, and consistency in outcomes over time.” Most advertisers, she notes, will start with a test-and-learn approach rather than committing real budgets upfront. This is a similar pattern to the early adoption of both Google Performance Max and Meta’s Advantage campaigns, where measurement infrastructure lagged behind adoption by months.
Bagwe adds, “Budgets won’t really change until we can show that these clicks convert at a higher rate than regular search.”
Attribution is still being built from scratch
OpenAI is reportedly hiring for its first advertising marketing science leader, a role that will own the company’s full measurement strategy, from attribution models to incrementality testing and geo experimentation. The scope of the role, described in a public job posting, also includes building self-serve reporting infrastructure and working with third-party measurement partners and clean-room providers.
Kumaran identifies the core concern clearly. “The absence of mature attribution models and incrementality frameworks will make advertisers cautious, at least in the early stages, with most adopting a test-and-learn approach before scaling spend.”
Dubey suggests, “I’m certain OpenAI will work on a proper 7-day attribution window for click-throughs. If not initially, some seasoned attribution systems like Meta (with Conversion API) and Google (with enhanced conversions) might cross-attribute on a last-click basis with UTMs. However, the quality will be visible with initial metrics almost immediately, such as average time on site and bounce rates, as basic referral tagging works with ChatGPT even now.”
That interim path gives performance marketers something operational while native attribution is built out.
The longer-term question is whether ChatGPT competes with search for performance budgets or carves out a new category between them. Kumaran’s view suggests that conversational AI platforms are unlikely to replace search but will occupy “a distinct space between discovery and intent while actively shaping demand, which could make them a powerful addition to the performance marketing mix.”
Dubey, on the other hand, believes that it will depend largely on whether it leans into recommendations built on past user data or stays anchored to the intent of a current query.
As of now, ChatGPT’s CPC ads are still in pilot, and it will take the next few months of data to show whether the clicks are worth what advertisers pay.