Visibility #27: ChatGPT Ads Get Smarter

Visibility Report #27 | July 28, 2026

ChatGPT ads now support conversion bidding, geographic exclusions, and bulk campaign tools. That makes ChatGPT look less like an experimental placement and more like a performance channel, while the organic side of AI search remains harder to measure.

What matters: Paid AI placements are getting familiar performance controls before publishers have clean organic AI search reporting.

Operator read: separate ad conversions, organic citations, answer mentions, crawler access, and entity accuracy before you trust any blended AI visibility score.

Now brands have to win twice: earn the citation, then understand when a paid placement changes the answer, the click path, or the conversion around it.

ChatGPT Ads Add Conversion Bidding and Campaign Controls

Search Engine Land reports that ChatGPT ads now include conversion bidding, geographic exclusions, and bulk campaign management. Those are not cosmetic controls. They move the channel closer to the operating model paid search teams already know: define an outcome, constrain reach, and manage campaigns at scale.

The measurement question is still unresolved. A ChatGPT ad can sit near an organic answer, influence a recommendation path, or intercept demand that would otherwise have gone through Google, direct, or brand search. Before budgets expand, teams need ChatGPT-specific conversion events and governance rules, not a copied Google Ads account structure.

📎 Search Engine Land​

Google Reports AI Growth but Leaves the Click Question Open

Google's Q2 reporting offered precise revenue figures and broad claims about AI Mode, search engagement, and AI-powered Search. Marie Haynes read the earnings call as another sign that Google is committed to expanding AI answers while protecting search advertising economics.

For operators, those signals matter, but they do not prove publisher traffic or lead quality is holding steady. Treat AI Mode usage and revenue as directional context. The site-level question remains: which answers cite you, which clicks arrive, and which conversions can you verify?

📎 Search Engine Journal | Marie Haynes​

Charging AI Crawlers Can Reduce Citation Reach

Publishers experimenting with paid crawler access face a distribution trade-off. Agents willing or able to pay may keep retrieving content, while other systems lose access. That can affect which products can quote, summarize, or cite a site, even if licensing revenue improves.

The hard part is that crawler access is not one thing. Training, retrieval, search indexing, and user-triggered browsing can follow different paths. Before blocking or charging a bot, document which AI products it serves and compare that access with referral traffic, citations, licensing revenue, and content costs.

📎 Search Engine Journal​

AI Search Cannot Cite a Business It Cannot Verify

Search Engine Land argues that AI systems struggle with businesses whose names, locations, services, and credentials differ across first-party pages and external sources. The implication goes beyond local SEO: conflicting entity data makes it harder for an answer engine to recommend the business confidently.

Teams should audit high-value entities across the website, major directories, review platforms, structured data, and answers from ChatGPT, Gemini, Perplexity, and Copilot. Consistency is not glamorous, but it is much easier to defend than trying to optimize prompts around bad source data.

📎 Search Engine Land​

Ahrefs Finds No Evidence of an AI-Content Penalty

Ahrefs studied 331,000 pages and reports that Google does not appear to punish content simply because AI helped create it. The more useful distinction is quality and intent: content designed mainly to manipulate rankings can still perform poorly or run into spam policies, regardless of how it was written.

So the operator question is not whether a draft contains AI-written sentences. It is whether the page contributes evidence, expertise, clarity, or information that competing pages and answer engines do not already have.

📎 Ahrefs​

Reviews Are Becoming Evidence for AI Recommendations

Amsive examines how online reviews can shape the entities, attributes, and sentiment AI search engines use when describing businesses. Review volume alone is not the full signal. Consistency across platforms, descriptive language, recency, and credible sources may all affect what an engine can verify.

Review programs should capture specific customer experiences without scripting claims or encouraging undisclosed incentives. That gives answer engines better evidence while reducing the risk that reputation management creates structured-data or platform-policy problems.

📎 Amsive | Search Engine Land review markup note​

YouTube May Be an AI Overview Visibility Gap

Search Engine Journal makes the case that YouTube content is an underused source for brands trying to appear in Google AI Overviews. Video can provide demonstrations, expert explanations, transcripts, and named entities that text-only programs may lack.

Treat this as a testable opportunity, not a guarantee. Use topics where a visual explanation materially improves the answer. Add descriptive titles, clear spoken definitions, accurate transcripts, chapters, and supporting pages so Google has multiple ways to understand the evidence.

📎 Search Engine Journal​

Profound Finds Google AI Mode Citing Google More Often

Profound reports that Google.com has become a prominent cited domain in AI Mode, raising questions about how much citation space remains available to independent sources. The provider also published research comparing variability across Gemini, AI Overviews, and AI Mode.

As vendor research, the findings deserve independent comparison, but the measurement point is real. Track the same topics separately across each Google surface, device context, and location. A blended citation rate can hide whether visibility is improving in Gemini while disappearing from AI Mode.

📎 Profound Google citation research | Profound model variability research​

Lumar Connects Technical Audits to AI Assistants

Lumar published setup guidance for its MCP server, which connects an AI assistant with Lumar technical SEO, auditing, and AI visibility data. The practical value is faster retrieval and analysis without repeatedly exporting reports or rebuilding context in a separate chat.

Access controls and output verification still matter. Start with read-only workflows such as finding crawl anomalies, grouping issues, and comparing AI visibility segments. Keep approval with a human before an assistant changes tickets, priorities, or production systems.

📎 Lumar​

More From This Week

Search & AI Visibility

• ​Google compares Gemini and AI Mode use with daily activities (Search Engine Journal) - Useful adoption context, but the methodology matters before using it as demand evidence.
• ​Audit what AI says about your locations (Search Engine Journal) - Practical checks for incorrect location facts, services, and brand descriptions in generated answers.
• ​Why SEO tests fail (Search Engine Land) - Better experimental design helps separate GEO changes from ordinary ranking volatility.

Paid AI and Measurement

• ​Household income exclusions appear in Performance Max (Search Engine Land) - Reported controls could affect governance and segmentation inside automated paid search.
• ​ChatGPT ads add conversion bidding and bulk tools (Search Engine Land) - New controls move ChatGPT advertising closer to a performance marketing channel.

From the Tool Blogs

• Ahrefs:Five AI search trends backed by Ahrefs data - A data-led look at which changes appear measurable and which remain mostly industry noise.
• Botify:How consumers respond to AI-generated results - Consumer data on AI-result adoption and trust, with methodology to inspect before changing strategy.
• Lumar:Getting started with the Lumar MCP server - Connects crawl and visibility data to assistant-led technical analysis.
• Profound:How ChatGPT Shopping works - Breaks down the systems between a shopping prompt and the product experience shown to a user.
• Profound:FactCheck connects with Noble mention refresh - Ties AI-claim accuracy monitoring to source-mention refresh workflows.
• Profound:Google.com is now AI Mode's second cited domain - Vendor research worth testing against your own prompt set.
• RankZero:Apple AI search in Safari could reduce Google dependence - A reminder to measure visibility across discovery surfaces, not only Google.

Agency & Practitioner Insights

• ​Seven reasons to monitor brand mentions (SparkToro) - Mentions reveal audience language and trusted sources that can influence AI answers.
• ​Google pays Reddit $60 million a year. For what? (Foundation Inc) - Community content remains valuable source material for AI-powered discovery.
• ​AI search visibility starts with one page (Search Influence) - A focused starting point for teams that need practical GEO work before a sitewide program.

So What Do You Do About It?

Run a paid-plus-organic AI visibility audit this week. For your five highest-value topics, record what ChatGPT, Gemini, Google AI Mode, Perplexity, and Copilot say, which sources they cite, whether paid placements appear, whether your crawlers and pages are accessible, and what conversion event you would trust if the interaction turns into a lead.

The goal is not one universal AI visibility score. It is a map showing where your brand is verified, where it is cited, where ads change the path, and where the measurement breaks.

The Visibility Report | Will Scott​
This newsletter is produced collaboratively by Will Scott and Bob, an AI agent. Human oversight, AI efficiency.
​Unsubscribe​