ChatGPT Ads for Agencies From Client Setup to Reporting

ChatGPT now shows ads, and your clients have noticed. Some want to know if they should be there. Others want to know why a competitor’s name just showed up under an answer about their category. Either way, your agency needs a confident answer, not a shrug.

The short answer: ChatGPT Ads are sponsored placements that appear alongside ChatGPT’s answers for adult users on the Free and Go plans. Advertisers buy them through OpenAI’s self-serve Ads Manager on a cost-per-click, cost-per-thousand-impressions, or conversion-optimized basis, with a minimum daily budget of $25. Agencies can’t open accounts on a client’s behalf. The client creates the account and invites the agency in. For most clients, the right move is a structured test with clear success criteria, not a budget shift away from search.

The rest of this guide covers everything an agency needs to run the channel well: how ads get matched to conversations, what clients actually pay, how to manage access across many client accounts, how to measure and report results, what the real performance data says, and the mistakes that burn test budgets fastest.

What ChatGPT Ads Are and How They Actually Work

Here’s the simplest way to picture it. Imagine asking a knowledgeable store clerk for advice. The clerk gives you an honest answer, and next to the counter sits a clearly labeled “sponsored” display related to what you just asked about. The display doesn’t change the clerk’s answer. It just sits beside it, hoping to be useful.

That analogy gets you most of the way. It breaks down in one important place, and that place is where agencies win or lose. There are no aisles to bid on. You don’t buy a keyword or a demographic. You describe the situations where your client belongs. The system decides when a conversation matches. The sections below break down where ads appear, how matching works, and what you can and can’t see as an advertiser.

Where the Ads Appear Inside ChatGPT

According to OpenAI’s launch announcement, ads are labeled as sponsored and visually separated from the organic answer. The standard unit is a card with a title, a short description, a square image, a logo, and a link. Retail clients can also run product-feed ads that pull in product details from a catalog.

OpenAI is also testing a visual format that appears while users generate images. TechCrunch reported that this format is starting in the US with a limited group of advertisers, so treat it as an opportunity to watch rather than one to sell clients on yet.

How ChatGPT Decides Which Ad to Show

Two things happen every time an ad is eligible to appear. First, the system figures out which ads are relevant to the conversation. OpenAI says it matches ads using the topic of the current conversation and, depending on the user’s country and settings, signals such as past chats and past ad interactions. On your side, the inputs are your ad group’s context hints, the ad copy, and the landing page.

Second, relevant ads compete in an auction. The OpenAI Help Center describes it as a relevance-weighted, second-price auction. In plain terms, the highest bid doesn’t automatically win, because a more relevant ad can beat a richer one. And when you do win, you pay roughly what it takes to beat the next competitor, not your full maximum bid. If you’ve managed Google Ads Quality Score, the logic will feel familiar.

What Advertisers Can and Can’t See

This is the adjustment that surprises most PPC teams. OpenAI states that advertisers never see users’ conversations, chat history, or personal details, and that ads don’t influence the answers ChatGPT gives. You get aggregate performance data, such as impressions, clicks, spend, and conversions.

That means there’s no search term report. You won’t see the exact prompts that triggered your client’s ad. You can’t mine queries the way you would in Google Ads. Your optimization loop runs on context hints, creative, landing pages, and conversion data instead.

Why It Isn’t Search Ads or Social Ads

It helps to place ChatGPT Ads between the two channels your team already knows.

The practical takeaway: your client’s ad shows up when the conversation is about their problem, not when a specific phrase appears. That rewards agencies that understand the client’s buyer deeply. It punishes agencies that paste in a Google keyword list and walk away.

Who Sees ChatGPT Ads and Who Never Will

The audience is huge, but it isn’t the whole ChatGPT user base. Knowing exactly who is in and who is out will save you from pitching the channel to the wrong client.

The Eligible Audience

OpenAI reports that ChatGPT reaches 1.2 billion people every week. Ads are shown to adult users on the Free and Go plans. Users on Plus, Pro, Business, Enterprise, and Education plans don’t see ads, according to OpenAI’s announcement.

Not every eligible user sees ads every day, either. Similarweb’s panel data found that by September 2026, 43.4% of eligible US users had seen at least one ChatGPT ad, compared with 45.7% in the UK and 27.8% in Mexico.

Contexts Where Ads Never Appear

OpenAI also limits where ads can show, even for eligible users. Its stated principles exclude users predicted to be under 18 and keep ads away from conversations about sensitive topics such as health, mental health, and politics. If your client’s buyers mostly arrive through those kinds of conversations, the available inventory will be thin.

What This Means for B2B and Premium Clients

Here’s the nuance most guides skip. The people most likely to pay for ChatGPT, including many professionals who use it all day at work, are exactly the people who don’t see ads. As Search Engine Journal points out, ChatGPT Ads only reach part of the ChatGPT audience.

That doesn’t rule out B2B. A procurement manager might use a company Enterprise seat at work and a free personal account at night. But it does mean you should expect a different audience mix than the “1.2 billion users” headline suggests. You should validate lead quality early, before scaling spend.

Where You Can Run ChatGPT Ads and How to Buy Them

Availability is broad enough now that most agencies can start a test this week. The two questions to settle first are which of your clients’ markets are covered and which buying route fits your agency.

Countries Where Ads Are Available

OpenAI says ChatGPT Ads are available in more than 60 countries. The Ads Manager availability page lists the markets where self-serve buying is open, including the US, Canada, the UK, most of the European Union, Australia, New Zealand, Japan, South Korea, India, Brazil, Mexico, and several markets in the Middle East and Southeast Asia.

Check that page before every new client kickoff. The country list grows often. An international client may have markets that are open for buying through OpenAI’s sales team before they open for self-serve.

The Three Ways to Buy

Most agencies will use the first route. It’s worth knowing all three so you can answer when a client’s enterprise team asks.

1. Self-serve Ads Manager. Campaigns are built and managed at ads.openai.com. This is the realistic path for independent agencies and small to mid-sized clients.
2. OpenAI’s sales team. Larger advertisers can work with OpenAI directly on managed campaigns.
3. Agency and technology partners. OpenAI named holding companies including Dentsu, Omnicom, Publicis, and WPP, plus tech partners such as Adobe, Criteo, Kargo, Pacvue, and StackAdapt. Havas Media was added later, according to OpenAI.

If your agency already buys through one of those platforms, ask your rep about ChatGPT inventory. It may let you manage the channel alongside your existing programmatic workflow.

How Much ChatGPT Ads Cost Your Clients

ChatGPT Ads has a low barrier to entry and no fixed price. You set a daily budget and a bid, and the auction decides what you pay. The details below cover the minimums, the bidding options, and what agencies are actually paying.

Minimum Budgets and How Pacing Works

The OpenAI Help Center sets the minimum daily campaign budget at $25, with equivalents in other supported currencies. There’s no minimum commitment beyond that.

One pacing rule trips people up. Your daily budget works as an average across a seven-day period. The system can spend up to twice the daily amount on a single day when it sees more opportunity. Weekly spend won’t exceed seven times the daily budget. Explain this to clients before launch so a $50 day on a $25 budget doesn’t trigger a panicked email.

Bidding Options by Campaign Objective

Your bid type follows the objective you pick, based on OpenAI’s campaign setup documentation.

The Help Center recommends a starting maximum CPC of $3 to $5. Treat that as a starting point, not a forecast. As Search Engine Journal explains, that recommendation often gets repeated as a benchmark, even though OpenAI hasn’t published one.

What Agencies Are Actually Paying

Real costs vary widely by market and industry. Here’s what published tests show.

• John Horn, CEO of StubGroup, wrote in Search Engine Land that his team sees average CPCs of $2 to $5 across industries.
• Search Engine Journal found CPCs ranging from about $3 to $22 across publicly shared tests. One multi-country test reported $5.10 in the UK, $10.62 in the US, $17.59 in Australia, and $22.89 in New Zealand.
• A B2B campaign covered in the same SEJ piece paid $9.29 per click and $64.34 per thousand impressions.

A Quick Way to Size a Test Budget

Work backward from the number of conversions you need to make a decision, not from the minimum budget. Say a client’s realistic CPC is $6 and their landing page converts at 2%. Getting 30 conversions, a reasonable floor for judging a channel, takes about 1,500 clicks, or roughly $9,000.

That math explains why so many early tests end in “inconclusive.” A $25-a-day test produces about 4 clicks a day at that CPC. After a month, you’d have around 125 clicks and two or three conversions, which isn’t enough to tell signal from noise. If the budget can’t reach a decision point, either narrow the test to the client’s highest-value offer or measure an earlier signal, like qualified demo requests, and say so upfront.

How Agencies Get Access to Client Accounts

This is where ChatGPT Ads differs most from Google and Meta for agencies. There's no manager account sitting above your client accounts. You can't create accounts for clients yourself. Getting the setup right on day one avoids ownership headaches later.

Why the Client Has to Own the Account

The OpenAI Help Center states that an agency can be invited after the client creates its advertiser account. It also says agencies can't create advertiser accounts on a client's behalf. As John Horn noted in Search Engine Land, OpenAI doesn't yet have an equivalent to Google Ads manager accounts or Meta Business Manager.

This is actually good hygiene. The client owns the account, the payment method, and the data from the start. If the relationship ends, nothing has to be transferred. You simply lose access.

The Setup Sequence, Step by Step

Walk your client through these steps on a short call. It usually goes faster than waiting on email back-and-forth.

Choosing the Right Role for Each Person

Ads Manager uses three account roles. Assign them the way you would in any other ad platform, with the least access each person needs.

Keep at least one Admin on the client side at all times. If the only Admin is an agency employee who later leaves, the client can end up locked out of its own account.

Managing Many Clients Without a Manager Account

One login can belong to multiple advertiser accounts. You switch between them inside Ads Manager. The Help Center notes a limit: once a user belongs to 10 or more advertiser accounts, that user can't create new ones but can still be invited to more. Since clients create their own accounts anyway, this rarely blocks agencies.

For larger teams, a few tools make multi-client work more manageable.

• Bulk upload lets you build or edit campaigns from a spreadsheet, following OpenAI's launch campaign guidance.
• The ChatGPT Ads Manager plugin lets you review and manage campaigns from inside ChatGPT or Codex, according to the Help Center.
• The Global Admin Console gives organizations on an OpenAI enterprise tenant central control over who has Ads access, as described in OpenAI's admin documentation.

Offboarding Without Drama

Because the client owns everything, offboarding is mostly about access. Export a final performance report, document campaign settings and context hints in a handoff file, then have the client remove your team's users. Add these steps to your standard offboarding checklist now, so nobody has to remember them later.

How to Build Your First ChatGPT Ads Campaign

The interface will feel familiar, but the craft is different. Most of your leverage comes from how you group intents and how you write context hints, so spend your time there rather than on bid tweaks.

The Campaign Structure

Ads Manager uses a three-level structure, as laid out in OpenAI's campaign documentation.

• Campaign. Sets the type (standard or product feed), the objective, locations, optional custom audiences, the budget, the conversion event, and the schedule. Location targeting can go as narrow as region, metro area, or postal code.
• Ad group. Sets the bid, the default landing page, and the context hints that describe when your ads should show.
• Ad. Holds the headline, description, image, and destination URL.

The rule of thumb: one ad group per distinct buying situation. If a client sells both payroll software and HR consulting, those are different conversations with different buyers. They belong in separate ad groups with their own hints and landing pages.

Writing Context Hints That Actually Match

Context hints are short descriptions of the topics and situations where your ad belongs. OpenAI's documentation is clear that they guide matching rather than act as exact-match rules, so think in terms of meaning, not phrasing.

Here's how that looks for a hypothetical project management SaaS client targeting agency owners.

Test both formats. John Horn shared a case in Search Engine Land where keyword-style hints outperformed question-style hints, which runs against what many people assume. Run them in separate ad groups so the results don't blur together.

If you qualify, use exclusions as well. OpenAI introduced Negative Phrases for qualifying advertisers, which help keep ads away from conversations that look related but aren't a fit.

Creative Specs and Copy That Earns the Click

OpenAI's launch campaign guidance gives these limits.

Make sure the landing page doesn't block OpenAI's crawlers, since the system reads it to understand relevance.

On copy, don't recycle search or social ads. A Graphite and Similarweb analysis of more than 95,000 US ChatGPT ads found that top advertisers wrote 98% new copy for the channel and used a small set of landing pages. The same analysis found ads that appear early in a conversation tend to earn higher click-through rates. The user is still exploring, so copy that helps them take a next step works better than a hard sell.

A Simple First Test Setup

If you want a default to start from, this setup works for most clients.

1. One campaign on the Clicks objective while conversion tracking proves itself, then switch to Conversions.
2. Two to four ad groups, each built around one buying situation.
3. Five to ten context hints per ad group, mixing keyword-style and situation-style phrasing.
4. Two or three ads per ad group, each testing a different angle, such as outcome, proof, or offer.
5. One dedicated landing page per ad group that continues the conversation the user was having.

Measuring and Reporting ChatGPT Ads for Clients

Measurement is where agencies earn their fee on this channel. The platform gives you less diagnostic data than Google Ads. Your tracking setup and your reporting story carry more weight. Set both up before the first dollar is spent.

Conversion Tracking With the OpenAI Pixel and Conversions API

OpenAI offers two ways to send conversion data back to Ads Manager: the OpenAI Pixel, which runs in the browser, and the Conversions API, which sends events from the server. Use both when you can. The server-side feed catches conversions the browser misses because of ad blockers or privacy settings.

The partner ecosystem is already deep. OpenAI lists data integrations with Hightouch, Tealium, and LiveRamp, attribution partners including AppsFlyer, Triple Whale, Northbeam, DV Rockerbox, Branch, Adjust, and Singular, and incrementality testing partners Haus, Measured, and WorkMagic. If a client already uses one of these tools, ChatGPT Ads can usually slot into their existing attribution view.

Separating Paid ChatGPT Traffic From Organic ChatGPT Traffic

This is the reporting trap that catches almost everyone. ChatGPT also sends organic referral traffic when it mentions or links to a brand in its answers. If paid clicks aren't tagged, analytics tools can lump them in with that organic traffic, and your client's report will understate the ads and overstate organic AI visibility.

The fix is simple: add UTM parameters to every ad URL. A Graphite study found that with UTMs in place, 92% to 96% of ChatGPT Ads clicks registered as paid sessions in Google Analytics 4. A clean convention looks like this:

utm_source=chatgpt&utm_medium=paid&utm_campaign={campaign_name}&utm_content={ad_group_name}

Then create a custom channel group in GA4 that routes chatgpt plus paid into a "Paid AI" channel, kept separate from organic AI referrals.

Know What the Platform Won't Tell You

Set expectations before the first report goes out. You won't get prompt-level or search-term data. Search Engine Land notes that reporting breaks down by country even when your targeting is narrower. There's no equivalent to Google's Auction Insights or impression share either.

That makes on-site behavior more important. Engagement time, pages per session, form completion, and downstream lead quality in the CRM become your main diagnostic signals.

Pulling ChatGPT Ads Into Cross-Channel Client Reports

Clients don't want another login. They want ChatGPT Ads in the same report as Google, Meta, and LinkedIn, so they can compare it with everything else.

Swydo connects to ChatGPT Ads natively. Add OpenAI Advertising from a client's data sources, and impressions, clicks, cost, CPC, CTR, and CPM flow into the same reports and dashboards as Google Ads, Microsoft Ads, Facebook Ads, and LinkedIn Ads. The integration ships with a preset ChatGPT Ads report template, so your first client report doesn't start from a blank page.

For conversion results, add GA4 to the same report. Your UTM-tagged ChatGPT Ads sessions and conversions already land there, so clients see delivery and outcomes side by side.

Asking About ChatGPT Ads Performance Inside ChatGPT or Claude

There's a fun twist here: the quickest way to review ChatGPT Ads results might be asking an AI assistant. The Swydo MCP connector links your Swydo account to ChatGPT, Claude, and other tools that support MCP. You can ask about one client or your whole portfolio in plain language and get answers from live reporting data, not a pasted export. Permissions carry over, so the AI only sees clients you can already access in Swydo.

A few prompts that work well for this channel:

• "Which clients had a ChatGPT Ads CPC above $8 last week, and how does that compare with their Google Ads CPC?"
• "Compare ChatGPT Ads CTR with Microsoft Ads CTR for [client] over the last 30 days."
• "Write a three-sentence summary of ChatGPT Ads performance for [client]'s monthly report."

This is most useful during a test, when you're checking spend pacing and early signals across several clients at once.

The KPIs to Put in Front of Clients

Report in layers, with business outcomes on top and platform metrics underneath as diagnostics.

Lead with the top two layers. A strong click-through rate paired with zero qualified leads isn't a win, and showing the layers together keeps that honest.

For formulas, benchmarks, and the one metric each campaign objective should be judged on, see our ranking of the most important ChatGPT Ads metrics.

ChatGPT Ads Benchmarks and What They Really Mean

Start with the most honest fact in this whole guide. OpenAI's own FAQ says the platform doesn't yet have performance benchmarks across advertisers, industries, or campaign types. Every number below is either panel data, a single advertiser's result, or a vendor-published case study. Each one is useful, as long as you know which kind you're looking at.

Click-Through Rates by Market

Similarweb's analysis, built with Graphite, measured clicks per ad-bearing ChatGPT response across all advertisers in each market.

Two cautions before you quote these to a client. First, this is a market-wide average across every advertiser, so a well-targeted campaign can beat it and a sloppy one can fall short. Second, it isn't comparable to a Google Search CTR, because the user didn't type a query looking for a vendor. The useful insight is relative: markets like Brazil and Mexico show noticeably higher engagement than English-speaking markets, which matters for clients with international reach.

Reported Results From Real Campaigns

The published case results so far are mixed, which is exactly what you'd expect from a young channel.

Notice the pattern. The strongest results come from advertisers with mature attribution setups and enough spend to measure properly. Many of them are published by OpenAI or its measurement partners. The weakest come from small or short tests and from B2B lead generation, where audience quality is the sticking point. Australian agencies told AdNews they saw sub-1% CTRs and lower conversion rates than other channels.

How to Use Benchmarks With Clients

Don't promise a number. Instead, agree on the client's own break-even point, such as the maximum cost per acquisition they can afford, and judge the test against that. Benchmarks are for sanity checks, like spotting a CTR far below the market average that points to a targeting problem. They aren't a forecast for any individual account.

Ad Policies, Restricted Industries, and Brand Safety

OpenAI's ad rules are stricter than most platforms in some areas and still being refined in others. Read the current OpenAI Ad Policies before you pitch any client in a regulated industry, because what's allowed depends on both the category and the country.

Regulated Industries Need a Closer Look

The categories agencies ask about most are finance, health, and legal. Under OpenAI's policies, ads in these areas are reviewed case by case in the US and are generally prohibited outside the US. Legal services can advertise in the US when the provider is licensed in the relevant jurisdiction. OpenAI also reserves the right to decline ads for reasons tied to its own competitive position. A client building a competing AI product should expect extra scrutiny.

If a client sits in one of these categories, build approval time into the launch plan. Don't promise a go-live date until the first ads clear review.

Brand Safety and Suitability Controls

Some brand safety protections are built into the product. Ads don't appear near conversations about health, mental health, or politics. They aren't shown to users predicted to be under 18, according to OpenAI.

You also get some controls of your own. OpenAI has announced Negative Phrases for qualifying advertisers, plus brand suitability pilots with DoubleVerify and Integral Ad Science. For clients whose brand teams require third-party verification on every channel, ask whether they can join those pilots before committing budget.

10 Mistakes That Waste Client Budgets on ChatGPT Ads

Most failed tests on this channel fail for predictable reasons. Here are the ones we see most often, and how to avoid each.

1. Treating the $3 to $5 bid guidance as an expected CPC. It's a starting bid. Forecast with the client's market and industry in mind, and expect real CPCs to land higher in some countries.
2. Creating the account in the agency's name. It isn't supported, and it creates an ownership problem the day the relationship ends. The client creates and owns the account.
3. Launching without UTMs and conversion tracking. Untagged clicks blend into organic ChatGPT referrals, and the ads look like they did nothing. Tag every URL and install the Pixel before launch.
4. Pasting in a Google keyword list as context hints. Hints work on meaning. Describe the buyer's situation and decision moment, and test keyword-style phrasing against it rather than assuming.
5. Cramming several intents into one ad group. The system can't match well when one ad group covers three different products. Split by buying situation.
6. Reusing search or social ad copy. People are mid-conversation, not mid-search or mid-scroll. Write copy that helps them take the next step in what they were already doing.
7. Judging the channel on a tiny test. A few hundred dollars over two weeks can't produce enough conversions to decide anything. Size the budget to reach a decision point, or measure an earlier signal.
8. Comparing panel CTR to a campaign's CTR or to Google Search CTR. Market averages, single campaigns, and search ads measure different things. Compare your client against their own break-even targets.
9. Ignoring lead quality in B2B. Clicks and form fills can look healthy while the companies behind them are the wrong size or industry. Check leads against the ideal customer profile in the CRM within the first two weeks.
10. Promising that ads will boost organic AI mentions. OpenAI says ads don't affect ChatGPT's answers. Sell paid and organic AI visibility as separate efforts with separate goals.

Is ChatGPT Advertising Worth It for Your Clients?

For the right client, yes, as a structured test. For the wrong client, it's an expensive way to learn something you could have predicted. The scorecard below helps you sort clients quickly. The test plan after it gives you something concrete to pitch.

A Quick Fit Scorecard

The more boxes a client checks in the left column, the stronger the case for a test.

How to Structure a Test Clients Will Trust

A good test answers one question: does this channel produce customers at a cost the client can accept, beyond what they'd get anyway? Build it like this.

1. Agree on the success metric upfront. Use a business outcome, such as cost per qualified lead or cost per acquisition, with a clear target.
2. Size the budget to the decision. Use the back-of-the-envelope math from the cost section to estimate the spend needed for roughly 30 conversions.
3. Run long enough to learn. Four to eight weeks gives the system time to learn and gives you a few rounds of creative and hint changes.
4. Measure incrementality when budgets allow. A geographic holdout, where some regions see ads and comparable regions don't, shows whether ChatGPT Ads drive new demand or simply claim credit for demand you'd have captured anyway. Partners such as Haus, Measured, and WorkMagic support this kind of test, according to OpenAI.
5. Set a go or no-go review date. Decide in advance what result means scale, adjust, or stop.

A Pitch You Can Adapt

Keep the pitch short and honest. Something like this works well:

"ChatGPT now shows ads next to its answers, and people in your category are asking it for recommendations. We'd like to run a focused six-week test with a budget of [amount], tracked against your target cost per acquisition of [amount]. If it beats that target, we scale. If it doesn't, you've spent a capped amount to learn whether this channel works for you before your competitors figure it out."

Clients respond well to a capped downside and a clear decision rule. It positions your agency as the team that tests new channels carefully, not the one that chases every shiny new platform.

How ChatGPT Ads Compare With Google and Other AI Ad Platforms

Clients will ask how this fits next to what they already run. The honest answer is that ChatGPT Ads complements search rather than replacing it. It's one of only a few places where you can buy ads inside an AI assistant today.

ChatGPT Ads Versus Google Search Ads

John Horn's comparison in Search Engine Land shows the difference in practice: ChatGPT Ads visitors spent less time on site and converted at a much lower rate than Google Ads visitors on the same site. That doesn't make the channel useless. It means it often reaches people earlier, so judge it on that basis and give it a fair attribution window.

For the full side-by-side on costs, click-through rates, targeting, and tracking, plus a 90-day plan for testing ChatGPT without hurting the campaigns that already work, see our guide to ChatGPT Ads vs Google Ads.

Other AI Platforms Agencies Ask About

The AI ad landscape is smaller than the headlines suggest.

• Google AI Mode. Google is testing new ad formats inside AI Mode, including conversational ad experiences. These run through existing Google Ads campaigns, according to Search Engine Land's coverage of Google Marketing Live. For many clients, this is the easiest AI placement to access because the account already exists.
• Microsoft Copilot. Ads in Copilot are bought through Microsoft Advertising, so they fit into existing Microsoft search workflows.
• Perplexity. Perplexity stopped testing advertising, as Search Engine Land reported. There's no Perplexity ad product to buy right now.

That makes ChatGPT Ads the largest stand-alone ad platform inside an AI assistant. It's also the only one that needs a new account, new targeting skills, and new reporting setup, which is exactly why clients need an agency for it.

Do ChatGPT Ads Affect Organic AI Visibility?

No, and it's worth being very clear with clients about this. OpenAI states that ads don't influence the answers ChatGPT gives, and John Horn notes in Search Engine Land that running ads doesn't change whether a brand is mentioned organically. Paying for the sponsored card won't earn your client a spot in the answer itself.

Why Paid and Organic Belong in the Same Plan

The two still work well together, just not in the way clients sometimes hope. Organic visibility, often called generative engine optimization or GEO, depends on how the brand shows up across the web: reviews, comparisons, credible coverage, and clear product pages. Ads buy guaranteed placement in conversations where the brand isn't yet showing up organically.

A practical way to combine them is to use organic gaps to guide paid targeting. If ChatGPT rarely mentions your client when people ask about their category, those conversation themes make strong context hints for an ad group. Over time, as organic visibility improves, you can shift budget away from topics where the brand already appears in the answer.

Reporting Both Without Double Counting

Keep the two in separate channels in your reports. Tagged ad clicks go into a paid AI channel, and untagged ChatGPT referrals go into organic AI. Show them side by side so the client sees the full picture of how AI assistants send them business, without either effort taking credit for the other.