AI Search Advertising for B2B: ChatGPT Ads, Microsoft AI Max, and Google AI Max Compared

In short

AI search advertising for B2B means buying placements inside AI-generated answers, and in September 2026 there are three inventories worth your budget: ChatGPT Ads, Microsoft AI Max (Copilot), and Google AI Mode reached through your existing Search, Shopping, and Performance Max campaigns. Start with Microsoft AI Max, because it is a toggle in an account you already have, it is the cheapest AI inventory on the market, and it makes your ads eligible for Copilot’s conversational searches.

Skip Perplexity: it closed its advertising program in February 2026, according to the Financial Times. Measure the program on Share of Voice in answers and assisted conversions, not clicks, because AI assistants frequently recommend a brand without linking to it. A 90-day rollout gets you from zero to a defensible cost per qualified lead across all three channels.

Three AI ad inventories you can actually buy

For a B2B team with a real budget, only three AI inventories are buyable today.

1. ChatGPT Ads (OpenAI). Live since February 9, 2026, self-serve at ads.openai.com since May 5, 2026, and at a $1 billion annual run-rate as of August 31, 2026, according to OpenAI and reporting by TechCrunch and Axios. Sponsored answers and cards appear inside the conversation, labeled as ads, targeted by the topic and intent of the chat. OpenAI’s help center recommends a starting bid of $3 to $5 and a minimum budget of $25 a day.

One advertiser’s campaign saw a 0.91% click-through rate versus a 6.4% benchmark for Google Search ads in the same sector, according to a March 2026 Campaign report citing Adthena: a single client’s result, not a platform-wide average. New accounts can typically advertise only in their home country at first; other supported countries open automatically once identity verification is approved and a required spend in the home country is reached, though the amount is not published. No agency manager accounts: your company verifies its own legal entity, country and currency lock permanently, and your agency is invited in as a user.

2. Microsoft AI Max (Copilot). An optional setting inside your existing Microsoft Ads Search campaigns, rolled out globally to all advertisers by August 2026, according to Microsoft Advertising. It makes your ads eligible for more complex, conversational searches on Bing and Copilot, though Microsoft’s own announcement does not claim advertisers see the full text of the Copilot queries that triggered them. It is the cheapest AI inventory available and the lowest-friction entry point for any market.

3. Google AI Mode and AI Overviews. You cannot buy a placement here directly, and there is no separate opt-in: text and Shopping ads from your existing Search, Shopping, and Performance Max campaigns become automatically eligible to show in AI Overviews and AI Mode, and you cannot opt out or target that placement directly, according to Google Ads Help. Reporting stays inside those campaigns rather than as a standalone AI line.

Not on the list: Perplexity. It closed its advertising program in February 2026 and still shows up in decks written before that date. If a proposal includes a Perplexity ad line, the rest of it is probably stale too. Reddit is the most-cited domain in AI answers, according to Peec AI’s analysis of 30 million citations, and it actively removes spam and fake accounts, so it is a place for honest participation, not a media buy.

Why Microsoft AI Max is the right first step

If you only have budget and attention for one AI channel this quarter, choose Microsoft AI Max. Three reasons.

It is already in your account. No new verification, billing entity, or interface. You flip a switch and existing campaigns become eligible for Copilot. It is the shortest path from decision to data in this category.

It surfaces conversational intent. Microsoft Ads reports the search terms behind your campaigns, now extended to more complex, conversational queries on Bing and Copilot. That is a valuable signal for a B2B marketer: buyers describing their problem in their own words. It feeds your ChatGPT topic map, your landing pages, and your sales enablement.

It runs on budgets you already have. You learn which intents convert inside campaigns you already run, then move money to ChatGPT Ads only for what already worked.

The sequencing is simple: use AI Max to discover, use ChatGPT Ads to scale what AI Max proved, and let Google AI Max ride along inside your existing search and PMax structure.

How AI answers compress the B2B funnel

A classic B2B search funnel runs awareness search, comparison search, vendor-name search, demo request, then months of sales cycle. Each step is a separate query and a separate chance to be found or lost.

An AI conversation collapses the first three steps into one session. A VP of Operations opens ChatGPT or Copilot and writes: “We run a 120-person logistics company on spreadsheets and a legacy TMS, we need something that integrates with NetSuite, budget is under $60k a year, what should we look at?” The assistant produces a shortlist. The buyer is now in vendor-evaluation mode after one exchange, and if you were not in that shortlist, you did not lose a click. You lost the deal before you knew it existed.

Three implications for a B2B advertiser:

1. The shortlist is the battlefield. Your job is to be named in the answer, whether through a paid placement or organic visibility. According to a 2025 AirOps analysis of 21,311 brand mentions across ChatGPT, Claude, and Perplexity, 85% of brand mentions came from external, third-party domains rather than the brand’s own website: media coverage, directories, reviews, forums, and YouTube. Brand mentions on third-party sites correlate with appearing in Google AI Overviews about three times more strongly than backlinks do, per an Ahrefs study of 75,000 brands (0.664 vs 0.218).
2. Positions are not stable. Only 30% of brands stay visible from one AI answer to the next, according to AirOps’ 2026 State of AI Search report. That is why paid placements matter: they are the one part of the shortlist you can influence directly.
3. The click is late in the process. By the time a user clicks through to your site, they have already been qualified by the conversation. This is why AI traffic converts so much better than organic: Opollo’s 2026 benchmark of 312 B2B tech firms found 14.2% versus 2.8% for organic traffic, a July 2025 Semrush study found AI visitors 4.4 times as valuable based on conversion rate, and Ahrefs found 23 times more sign-ups per visit from its own site traffic.

Fewer, warmer visitors is exactly the profile a six-figure deal wants.

Measuring AI search advertising: Share of Voice, not clicks

Open the ad platform, compare clicks and CPC against Google Search, and you will kill a working program in month two. ChatGPT often names a brand without linking to it; the influence happens, the click does not. You need three layers.

Layer 1: Share of Voice in AI answers. Build a list of the 30–50 prompts your buyers actually use (your AI Max search-term reports are the best source). Track how often your brand appears in answers to those prompts, on ChatGPT, Copilot, and Google AI Mode, and how that share moves month over month. Trackers exist for this: Otterly from $29 a month, Peec at €89–199 a month, Profound from $99 a month. Pick one and baseline before you spend a dollar on ads.

Layer 2: Assisted conversions. Count closed deals with an AI-referred session anywhere in the path, not just last click. Tag AI referrers cleanly and review weekly.

Layer 3: Cost per sales-accepted lead by channel. The number your CFO will ask about. Compare ChatGPT Ads, AI Max, Google Search, and LinkedIn on one definition: a lead sales actually accepted. Not MQLs, not form fills.

What you should not put in the contract with your agency: traffic targets or positions. Nobody can promise a position in a ChatGPT answer. Contract on Share of Voice and assisted conversions.

Channel comparison table

A 90-day rollout for B2B teams

This assumes you already have Google Ads and Microsoft Ads accounts and some form of CRM attribution.

Days 1–14: baseline and access.

• According to Otterly.ai’s 2026 AI Citations Report, 73% of sites have technical barriers, including robots.txt blocks, CDN restrictions, and JavaScript rendering issues, that keep AI crawlers like GPTBot, OAI-SearchBot, ClaudeBot, and PerplexityBot from accessing their content. Check your own robots.txt first and unblock what you can. Confirm the site is in Bing’s index: ChatGPT search draws on Bing as a data provider alongside OpenAI’s own crawler, OAI-SearchBot. Set up Bing Webmaster Tools with IndexNow to help Bing discover your pages faster. The full checklist is in how to get your site indexed by ChatGPT.
• Build the prompt list and baseline Share of Voice with a tracker.
• Start ChatGPT Ads entity verification now; it involves legal and finance and takes time.
• Flip on Microsoft AI Max in existing campaigns.

Days 15–45: learn on AI Max.

• Pull the Copilot-eligible search-term reports weekly. Cluster the queries into intents: problem-aware, solution-aware, vendor-comparison.
• Build one landing page per top intent that continues the conversation rather than restarting it.
• Confirm your Google Search, Shopping, and PMax campaigns are eligible for AI surfaces (there is no separate opt-in) and watch lead quality.

Days 46–75: launch ChatGPT Ads on proven intents.

• Use the AI Max intent clusters to build the ChatGPT topic map. Do not import your Google keyword list.
• Run breadth for two weeks, then cut irrelevant topics. The launch mechanics are in how to launch ChatGPT Ads step by step.
• Track cost per sales-accepted lead from day one, alongside CTR, not instead of it.

Days 76–90: decide and rebalance.

• Compare all three channels plus Google Search and LinkedIn on cost per sales-accepted lead and assisted conversions.
• Re-measure Share of Voice against the day-14 baseline and set next quarter’s split using the logic below.

Budget split logic without fixed numbers

There is no correct percentage split; it depends on deal size, your Microsoft Ads footprint, and how fast legal clears ChatGPT verification. There is a correct order of operations.

• Fund discovery before scale. AI Max is where you learn cheaply. Give it enough budget to produce a query export you can actually cluster, which usually means more than a token amount.
• Let ChatGPT Ads earn its budget. Start it only on intents AI Max has already validated, size it so you get a readable number of visits per intent per week, and grow it as cost per sales-accepted lead holds.
• Treat Google AI surfaces as coverage, not a new line. Eligibility is automatic on your existing Search, Shopping, and PMax campaigns, so there is no separate budget to allocate. Monitor lead quality rather than carving out a separate line item.
• Reserve a share for visibility work. Most AI visibility is built off-site, so part of the program should earn honest mentions in the media, directories, review platforms, and communities your buyers trust.
• Rebalance quarterly, not weekly. AI channels have small volumes and noisy weeks. Move money on 90-day evidence.

How MOVE structures an AI advertising retainer

We run AI advertising as an add-on to the SEO and media buying we already do for B2B clients, because the comparison against your existing channels is the whole point. A typical engagement starts with an AI visibility audit from $1,500: crawler access, Bing indexation, Share of Voice baseline on your buyer prompts, and a channel-by-channel plan.

The retainer then covers Microsoft AI Max setup and query analysis, ChatGPT Ads build and management inside your own verified account, confirming Google AI Max and PMax eligibility, landing pages per intent, and a weekly report on Share of Voice, assisted conversions, and cost per sales-accepted lead. Campaign management is priced from 15% of ad spend. We are a Google Partner and Meta Partner with B2B and fintech work behind us, and we do not put positions or traffic targets in the contract.

6 mistakes B2B teams make in AI search advertising

1. Starting with ChatGPT Ads instead of AI Max. You pay premium CPCs to learn what Microsoft would have shown you in search-term reports for a fraction of the cost.
2. Measuring on clicks. AI assistants recommend without linking. On a click KPI the channel looks broken while it quietly fills your pipeline.
3. Putting Perplexity in the plan. Closed since February 2026. Its presence in a proposal tells you when the proposal was written.
4. Ignoring off-site visibility. Around 85% of AI visibility forms outside your domain. Ads on top of zero mentions are expensive.
5. Buying schema markup as a GEO strategy. A May 2026 Ahrefs study that tracked 1,885 pages adding schema markup against 4,000 control pages found no statistically meaningful lift in AI citations on ChatGPT, Google AI Mode, or Google AI Overviews. Treat schema markup and llms.txt as hygiene, not as a paid growth lever.
6. Verifying ChatGPT Ads under the wrong entity. Country and currency are permanent. Decide with legal and finance before anyone clicks “verify.”

Where to start

If your sales cycle is long and your lead is expensive, your buyers are building shortlists in AI answers right now. Flip on Microsoft AI Max this week, baseline Share of Voice, and start ChatGPT verification in parallel.

If you want the 90-day plan run for you, book an AI visibility audit or read how we scope AI advertising for B2B teams. The launch tutorial and indexing checklist live in the blog, and you can contact us with questions before you commit a budget.