For most of the last decade, B2B content marketing followed a familiar path. A team published useful pages, search engines ranked them, buyers clicked through, and forms captured the interest. That path assumed the buyer would reach the page. Increasingly, the buyer reaches an answer first, and your page is one of several sources behind it.
This article looks at how the rise of AI Search Engines changes content marketing for B2B teams. It covers how buyers use these tools, what the click and traffic studies show, what Google says to do and not to do, and which content gets cited. It also covers why freshness matters, where AI-assisted production helps or hurts, and how to measure content against qualified opportunities and pipeline instead of pageviews.
It draws on Google's own guidance and on buyer surveys from Forrester, G2, and 6sense. It also draws on click and traffic studies from Pew Research Center, Ahrefs, SE Ranking, and Conductor, and on citation research from Wix Studio and Seer Interactive. Many figures come from software vendors, agencies, or self-reported surveys, so it notes where research has limits. It reflects what was published as of November 13, 2026, and this area changes quickly.
What content marketing was built to do
B2B content marketing exists to build preference before a sales conversation. It educates a buying group, shows expertise, answers hard questions, and gives sales something worth sharing. Search was the main delivery route, and traffic was the usual proof that the route worked.
That proof is now less reliable. An AI Search Engine can answer a question, compare vendors, and suggest a shortlist without sending the buyer to any single page. The content still matters, because the answer is built from sources. What changes is where the content does its work, how visible that work is, and how a team knows whether it paid off.
Four shifts matter most, and they pull in different directions.
- Buyers start earlier in AI-generated answers. Research that used to happen on a results page now often happens in a conversation.
- Clicks are less certain. Some studies show fewer visits when an AI-generated summary appears, and Google disputes the aggregate picture.
- The quality bar is rising. Google says unique, valuable content is likely the biggest long-term influence on its generative AI features, and buyers lean on experts to check what AI tells them.
- Measurement is harder. Influence can happen with no visit, so sessions and form fills alone understate it.
The rest of this article takes each shift in turn.
How B2B buyers now use AI Search Engines
Buyer surveys agree on the direction, though not on the numbers. Forrester's analysis of its 2025 Buyers' Journey Survey reports that 94% of business buyers used AI in their buying process, up from 89%. It also says twice as many buyers named generative AI or conversational search as the more meaningful source than named any other source. Forrester also reports that 61% use private AI tools provided by their organizations. That detail matters for measurement, because a visit from a company-managed tool may not be labeled in your analytics. That is an inference, not a Forrester finding.
G2 reports similar behavior among software buyers. Coverage of its March 2026 survey of 1,076 B2B software buyers says 51% now start research with an AI chatbot more often than with Google. G2 also reports that 69% of buyers chose a different vendor than they first planned, and that comparing vendor strengths and weaknesses was the top use at 41%. In G2's list of sources that shape vendor shortlists, AI chatbots ranked first at 54%, ahead of review sites at 43%, vendor sites at 36%, and thought leadership at 28%. G2 sells a software review marketplace, the survey covers software buyers only, and every figure is self-reported.
Buyers do not trust AI alone. Digital Commerce 360's summary of Forrester's report on business buying in 2026 says buyers start with generative AI. They then validate what it tells them through a buying network. The same summary says engagement with providers was triggered more by industry experts than by AI tools. Forrester separately predicted that human expertise will rival generative AI in appeal as buyers seek deeper validation. That is a prediction, not a measured result.
The timing of influence also matters. 6sense's 2025 Buyer Experience Report says that 95% of the time the winning vendor was already on the buyer's Day One shortlist. It also says four out of five deals are won by the pre-contact favorite. 6sense sells revenue intelligence software.
Taken together, these findings suggest that content has to work earlier and less visibly. It has to make the company known, credible, and easy to describe accurately by the time a shortlist forms, which can be long before a form fill. That is an inference from the surveys. None of them measures how much any single piece of content contributes.
What the click studies show, and where they disagree
Studies of clicks and traffic point in different directions, and the differences come from what each one measures.
- Pew Research Center. Its analysis used March 2025 browsing data from 900 US adults. Users clicked a traditional result link in 8% of visits when an AI summary appeared, against 15% when none appeared. They clicked a link inside the summary in 1% of visits. They also ended their browsing session entirely on 26% of pages with a summary, against 16% without one, as Search Engine Land reported. The data covers general consumer browsing on Google, and it dates from March 2025.
- Ahrefs. Its update, published in February 2026, compared 300,000 keywords using desktop Search Console data from December 2023 and December 2025. It estimated that an AI Overview correlates with a 58% lower click-through rate for the top-ranking page. The author words the result as a correlation. The estimate rests on a forecast that assumes the affected keywords would have declined like informational keywords with no AI Overview. The author notes that other zero-click features also cut clicks. Ahrefs sells search marketing software.
- Google. In August 2025, Google's head of Search, Liz Reid, wrote that organic click volume from Google Search to websites was relatively stable. She said Google was sending slightly more quality clicks. Search Engine Roundtable noted that Google shared no specific data with the post.
These findings can all hold at once. An aggregate can stay steady while particular pages, query types, and industries lose clicks. The Ahrefs study looks at top-ranking pages on keywords that trigger an AI Overview, which is where a summary is most likely to stand in for a visit. That reading is an inference. The useful question for a B2B team is not which study is right. It is what your own Search Console data shows by query type and page.
Referral traffic from AI chatbots is still small. SE Ranking's study used anonymized Google Analytics data from 101,574 websites, tracked from January 2025 through April 2026. It found that AI referrals made up 0.32% of traffic in the first four months of 2026, about 1 in 312 visits, up from 0.24% in 2025. Organic search accounted for 42.75%. Conductor's benchmark, which used data from May through September 2025, put the average share of AI referral traffic at 1.08%, and the figure for information technology sites at 2.80%. The samples and dates differ, so the numbers do not line up. A vendor study of more than 16,000 sites reported that AI referrals fell at the median SaaS and technology site. It also reported that the overall AI share of sessions rose.
The implication is that visibility and traffic are separating. A buyer can see your company named in an answer, form a view, and never visit. A team that judges content only on sessions will miss influence that never produces one.
What this means for different kinds of content
The evidence does not say all content loses. It says different kinds of content face different pressure.
- Generic informational content is most exposed. Google's guide separates commodity content, such as generic tip lists drawn from common knowledge, from non-commodity content that offers expert or experienced insight. It says content that is unique, valuable, and good is likely the biggest long-term influence on presence in generative AI search. It adds that a first-hand review offers a perspective that a summary of existing content does not. This fits the click evidence, because informational queries are where a summary can replace a visit. That link is an inference.
- Comparison content needs outside support. In the Wix Studio and Peec study, listicles drew 40.86% of citations for commercial-intent prompts. Among the top 1,000 cited URLs in professional services, self-promotional listicles, where a brand ranks itself first, accounted for 19.1% of citations. Third-party listicles on neutral sites accounted for 80.9%. A vendor's own ranked list is therefore unlikely to carry comparison queries alone. Coverage on neutral sites matters too, which is where earned coverage and review programs support content. That reading is an inference from one industry sample.
- Product and solution pages are content too. For transactional prompts, the same study found that product pages drew 24.88% of citations and category pages 14.97%. Pages that explain what you sell, who it is for, and how it works are cited, so they deserve the editorial care usually given to blog posts.
- Original research and expert material has the clearest case. In the Clutch and Conductor survey, proprietary research and original reports ranked as the top written content priority for visibility in AI-generated answers. In the Content Marketing Institute survey, marketers named useful, practical information (64%) and demonstrated expertise (59%) as top trust builders, and original research ranked lower at 27%. Both are marketer opinions. Neither shows which content AI Search Engines cite.
What Google says to do, and what it says not to do
On May 15, 2026, Google published a guide to optimizing for its generative AI features in Search. The page was last updated July 10, 2026. It says retrieval for these features relies on Google's core Search ranking systems. It answers yes to whether optimizing for generative AI search is still SEO, and it treats the newer labels for this work as SEO from Google's perspective. It also describes two techniques. Grounding, also called retrieval-augmented generation, pulls current pages from the index to improve the accuracy and freshness of AI responses, which show links to supporting pages. Query fan-out is a set of related queries the model generates at the same time.
The guide also lists things site owners do not need to do for Google Search.
- Special files or markup. Google says no new machine-readable files, AI text files, markup, or Markdown, including llms.txt, are needed to appear in Google Search. It says creating them neither helps nor hurts Search visibility.
- Chunking. Breaking content into tiny pieces is not required, and there is no ideal page length.
- Rewriting for AI. The systems understand synonyms, so there is no need to chase long-tail keyword variations.
- Inauthentic mentions. Google says these are not as helpful as they seem.
- Structured data for its own sake. It is not required for generative AI search, though it still helps with rich result eligibility.
Two limits apply. The guide covers Google's features only, and other AI Search Engines may behave differently. The guide also says indexing and serving are not assured, even for pages that meet its requirements.
Google's guidance on generative AI content adds a boundary on volume. The page, last updated October 1, 2026, says that using AI tools to generate many pages without adding value for users may violate the scaled content abuse policy. It says to manually fact-check and review AI-generated content before publishing. A reasonable reading is that the pattern to avoid is a thin page for every prompt variation. That is an inference.
What gets cited, and the limits of that evidence
A citation study from Wix Studio's AI Search Lab, which used data from Peec AI, shows what these systems cite. It analyzed 75,000 AI answers and 1,056,727 citations from ChatGPT, Google AI Mode, and Perplexity. It used non-branded prompts across five industries that included SaaS and professional services. Listicles drew 21.9% of citations, articles 16.7%, and product pages 13.7%. The clearest pattern was that the cited format changed with the intent of the prompt.
- Informational prompts. Articles led at 45.48% of citations.
- Commercial prompts. Listicles led at 40.86%.
- Transactional prompts. Product pages led at 24.88%.
The study has limits. All prompts were non-branded, and the author notes that real transactional prompts are often branded. The number of prompts is not stated. The self-promotion comparison covers one industry. The author works for Peec AI, which supplied the data. And the study covers three systems, not all six AI Search Engines.
A smaller agency study, published in September 2026, used Search Console data from seven B2B software companies that the agency works with, from March through August 2026. It reported that AI Overviews and AI Mode mostly cited informational pages. It also reported that ChatGPT slightly preferred product pages (18.9%) over listicles (17.7%), and that Gemini and Perplexity leaned toward listicles. The two studies agree that the cited format depends on the question and the system, though they measure different things. The sample is small, and the agency sells search services.
The practical reading is to match format to the buyer's question instead of chasing one template. A buyer asking what a term means needs an explanation. A buyer asking which vendors fit needs a comparison. A buyer ready to evaluate needs a page about the product. The studies show associations between format and citation. They do not show that changing a page's format would earn a citation.
Freshness changes how content is maintained
Several studies find that pages cited in AI-generated answers tend to be fresher than pages in organic results. Ahrefs analyzed 17 million citations across seven AI search platforms. Its study, published in July 2025, found that cited pages were 25.7% fresher on average than pages in organic results, measured from the publication date. Measured from the last update, the gap was smaller, at 13.1%. The average cited page was still about 2.9 years old. ChatGPT leaned most toward newer pages, while Google's AI Overviews and organic results were the most likely to cite older ones.
Seer Interactive's 2026 study examined 7,683 cited pages that carried 47,097 citations across ChatGPT, Gemini, and Perplexity from March through June 2026. It reported that about three in four cited pages had been updated in the last year, but only about 42% had been published that recently. In its reading, freshness came from updates to existing pages, not from new publishing. It also found engine differences. Of the content Gemini cited, 78% had been updated within a year, against 73% for ChatGPT and 65% for Perplexity, which still cited older evergreen content.
Three cautions apply. According to an agency's summary of the study, Seer could date only about two thirds of the pages. The sample spans four industries, retail energy, pet retail, travel, and commercial banking, so B2B categories may behave differently. And the finding is correlational, because maintained pages may simply be the pages a team cares about most. That last point is an inference.
For a B2B team, the implication is a change in operating model, not a new tactic. Content becomes a library that needs review, not a stream that ends at publication. Pages tied to pipeline, such as comparison, solution, and evaluation content, are the first candidates for scheduled review. A changed date without a substantive revision adds nothing for the buyer, and the studies do not show that it earns a citation.
AI-assisted production and the quality gap
Adoption is widespread, and the benefits are uneven. In the Content Marketing Institute's survey, fielded from June through August 2026, 88% of B2B marketers said their organizations use AI content creation tools. They credited AI mainly with speed (78%), efficiency (59%), and volume (44%). Quality ranked low at 17%. The top challenges were accuracy and quality control, each at 54%. Among marketers who rated their marketing effective, content relevance and quality was the top contributing factor at 47%, while AI-powered marketing applications ranked at 14%. CMI notes that these results show correlation, not causation.
The Clutch and Conductor survey of more than 450 content marketers, fielded in January 2026, found that 75% use AI-powered tools in standard content creation. Nearly a quarter said large language models are the primary audience for most of their content, rising to 32% among enterprises. Writing mainly for machines is a risk, given Google's position that special rewriting is not needed. That is an inference. A page written to be extracted can become a page that says nothing new.
Ahrefs tested whether AI-flagged pages rank. In its July 2026 study, it pulled 1 million pages from the top 10 results for 100,000 searches in June 2026. About 150,000 had enough text for AI detection. Among pages in positions one through three, 5.3% scored as 100% AI-generated, and pages scoring under 50% made up 82.2%. According to summaries of the study, pages with low or moderate AI scores earned two to three times the impressions of pages with high or very high scores. Ahrefs used its own detector, which it describes as imperfect and different from any system Google may use. The study shows association, not cause.
These sources are consistent with treating AI as a production aid, though none proves it. What separates one page from another is what it adds that the others lack, such as first-hand experience, specific detail, and review by someone who knows the subject. Fact-checking is not optional, because Google says to review AI-generated content before publishing.
How marketers say they are measuring content
Demand Gen Report's coverage of a 10Fold survey describes how 400 B2B technology marketing decision makers in the United States and Europe answered questions about content. Sapio Research conducted the survey, and it was released in May 2026. It found that 52% ranked AI-generated search and answer engines as their top content distribution channel. AI search visibility was the most cited success metric at 40%, ahead of marketing-qualified leads at 33%. 10Fold sells content and communications services, and the figures are self-reported.
Visibility is a useful signal. It is not the outcome. A team that reports visibility without qualified opportunities has measured an input.
How to measure content against pipeline
A revenue-first view keeps visibility metrics in their place. It uses three layers.
- Layer 1: activity and infrastructure signals. Impressions, Search Console queries, AI referral sessions, and sampled mentions or citations in AI-generated answers. Use them to diagnose reach and technical health, not to judge the program.
- Layer 2: conversion performance. Qualified opportunities and customers by content touched, entry page, and source. This is where attention becomes a sales conversation or does not.
- Layer 3: business impact. Pipeline, customers, and recurring revenue, attributed only where reliable data supports it.
Three practices help when platforms report little.
- Ask buyers how they found you. Add an open-text field to forms and a standard question to the first sales call. Record mentions of ChatGPT, Claude, Perplexity, Grok, Gemini, and Google AI Overviews in the CRM. Buyers who researched in a company-managed tool may appear as direct visits, so their own words fill the gap.
- Track content touches in the CRM. Tie entry pages and assets to opportunities so Layer 2 does not depend on last-click reports.
- Set a baseline and compare over quarters. A long sales cycle means a month of data says little about pipeline. Change one thing at a time where possible, and compare cohorts across several quarters.
If visibility rises and qualified opportunities do not, diagnose the conversion pathway before producing more content. Check whether the pages that earn attention explain the offer clearly and whether forms ask for too much. Check whether leads route to the right person quickly and whether follow-up matches what the buyer read. Do not attribute a change in pipeline to content unless the data supports it.
Where content fits in the larger system
Content does not work alone, and AI Search Engines make that clearer.
- AI Search Readiness + SEO. Content has to be crawlable, accurate, and well organized to be considered at all. Google says SEO fundamentals still apply.
- Digital PR. Comparison queries lean on third-party sources, so earned coverage supports what a company says about itself.
- Conversion Rate Optimization (CRO). Attention only becomes pipeline if the page converts it.
- RevOps and Account-Based Marketing. Routing, tracking, and target-account focus decide whether content results reach Layer 2 and Layer 3.
- GTM Strategy. Positioning decides what the content says and who it is for. Without it, volume produces generic pages.
10spring treats Content Development as one of nine Execution Channels that work together. It performs best when it is coordinated with the others and judged on pipeline, not output.
Practical takeaways
- Diagnose before producing more. Audit which pages earn qualified opportunities, which queries lost clicks, which pages AI Search Engines cite, and which pages are stale.
- Retire or rebuild commodity pages. Generic tip lists that summarize common knowledge are the most exposed.
- Put first-hand expertise on the page. Named experts, original data, and specific detail give a page something others lack.
- Match format to the buyer's question. Explain, compare, or describe the product depending on what the buyer is trying to do.
- Treat comparison content as a team effort. Pair your own pages with coverage on neutral sites that buyers and AI Search Engines already use.
- Maintain the library. Schedule substantive reviews, starting with pages closest to pipeline.
- Use AI as a production aid. Have someone who knows the subject fact-check and edit, and avoid publishing pages in bulk.
- Skip unverified tactics. Google says llms.txt, chunking, and rewriting for AI are not needed for its features.
- Measure in layers. Judge content on qualified opportunities and pipeline, and use visibility as a diagnostic.
- Set a baseline and compare over quarters. One month of data says little in a long sales cycle.
10spring works alongside internal marketing teams. If you want to pressure-test how your content connects to pipeline, Book A Call. Let's accelerate your pipeline together.
Sources and further reading
All sources accessed November 13, 2026.
- Google Search Central. Optimizing your website for generative AI features on Google Search. Documentation. The guide was announced May 15, 2026, and the page showed a last-updated date of July 10, 2026. Used for the points on core ranking systems, grounding, query fan-out, unique and non-commodity content, and the list of things site owners do not need to do. It describes Google's features only.
- Google Search Central Blog. Announcement of the guide to optimizing for generative AI in Google Search. Blog post, read through a search excerpt only. The page excerpt reviewed does not state a date. Used to confirm that Google announced the guide.
- Google Search Central. Using generative AI content. Documentation. The page showed a last-updated date of October 1, 2026. Used for the scaled content abuse boundary and the instruction to fact-check and review AI-generated content.
- Forrester. B2B Buyers Make Zero-Click Number One. Analyst blog post on Forrester's Buyers' Journey Survey, 2025. Read through a search excerpt only. The page excerpt reviewed does not state a date. Used for the 94%, 89%, and 61% figures and the point about generative AI as the more meaningful source. Forrester sells research and advisory services.
- Digital Commerce 360. Coverage of Forrester's report on business buying in 2026. Trade article, January 22, 2026. Read through search excerpts only. Used for buyers validating AI output through a buying network and for experts outranking AI tools as triggers for engagement.
- Forrester. 2026 B2B marketing, sales, and product predictions. Press release, October 28, 2025. Read through search excerpts only. Used for the prediction that human expertise will rival generative AI in appeal. It is a prediction, not a result. Forrester sells research and advisory services.
- G2. The Answer Economy, G2's 2026 AI search insight report. Vendor report, read through search excerpts only. The page excerpt reviewed does not state a date. Used for the ranking of sources that influence shortlists. G2 sells a software review marketplace, and the survey covers software buyers only.
- Demand Gen Report. Coverage of G2's March 2026 survey of B2B software buyers. Trade article, May 4, 2026. The full page could not be fetched, so it was read through search excerpts. Used for the 1,076-buyer sample and the 51%, 69%, and 41% figures.
- 6sense. 2025 Buyer Experience Report. Vendor report, read through a search excerpt only. The report was released November 12, 2025. Used for the Day One shortlist figures. 6sense sells revenue intelligence software.
- Pew Research Center. Google users are less likely to click on links when an AI summary appears in the results. Short read, July 22, 2025, on March 2025 browsing data. Read through search excerpts and secondary coverage only. Used for the 8%, 15%, and 1% figures and the sample.
- Search Engine Land. Google's AI Overviews are hurting clicks: Pew study. Article, read through a search excerpt only. The page excerpt reviewed does not state a date. Used for the 900-adult sample and the 26% versus 16% figures on ending browsing sessions.
- Ahrefs. Update: AI Overviews Reduce Clicks by 58%. Study, February 4, 2026, with a later modified date of October 15, 2026. Used for the method, the 58% figure, and the author's wording and caveats. Ahrefs sells search marketing software.
- Google. AI in Search is driving more queries and higher quality clicks. Blog post, August 6, 2025. Read through search excerpts only. Google's own statement about its own product.
- Search Engine Roundtable. Google clicks stable with AI Overviews. Article, read through a search excerpt only. The page excerpt reviewed does not state a date. Used for the observation that Google shared no specific data.
- SE Ranking. Analysis of Top AI Search Engines: Who Is Catching Up to ChatGPT? Study, June 18, 2026, using aggregated, anonymized Google Analytics data from 101,574 websites. Used for the 0.32%, 0.24%, and 42.75% figures. SE Ranking sells search software, and the authors say the results may not hold for every site.
- Conductor. 2026 benchmarks report. Vendor report, read through a search excerpt only. The data covers May through September 2025. The page describes 1,215 enterprise customer domains for its traffic findings, while press materials cite 13,770 domains, so sample descriptions differ. Conductor sells search and answer engine software.
- Conductor. Information technology industry benchmarks. Vendor report page, read through a search excerpt only. The page excerpt reviewed does not state a date. Used for the 2.80% figure.
- Search Engine Land. Coverage of Conductor's AI referral benchmark. Article, read through search excerpts only. The page excerpt reviewed does not state a date. Used for the 1.08% average and the data period.
- SEOcrawl. AI traffic versus organic search study. Vendor study, read through a search excerpt only. The page excerpt reviewed does not state a date, and the method is not described in what was reviewed. Used for the statement that AI referrals fell at the median SaaS and technology site.
- Wix Studio AI Search Lab. The content types most cited by LLMs. Study, March 23, 2026, by a researcher at Peec AI, which supplied the data. Used for the sample, the citation shares by format and intent, and the stated limit on non-branded prompts. The number of prompts is not stated.
- Position Digital. What B2B SaaS Content Works Best in AI Search? (New Study). Agency study, September 9, 2026, using Search Console data from seven B2B software companies the agency works with. Read through search excerpts only. Position Digital sells search services.
- Ahrefs. New Study: AI Assistants Prefer to Cite Fresher Content (17 Million Citations Analyzed). Study, July 28, 2025. Read through search excerpts only. Used for the 25.7% and 13.1% figures, the 2.9-year average age, and the platform differences. The data comes from Ahrefs' own Brand Radar tool, and Ahrefs sells search marketing software.
- Seer Interactive. Study: Content Recency's Impact on AI Visibility in 2026. Study, read through search excerpts only. An agency summary dates it July 24, 2026. Used for the sample, the updates finding, the industries covered, and the engine figures. Seer sells search marketing services.
- Something Inc. Content recency and AI citations: refresh beats publishing new. Agency blog post summarizing Seer's study, read through search excerpts only. The page excerpt reviewed does not state a date. Used only for the date of Seer's study and the point that about two thirds of pages could be dated. It is an agency's reading of the study, not Seer's own.
- Content Marketing Institute. Faster and Cheaper, But Not Better. Article dated November 4, 2026, on a survey fielded June 23 through August 11, 2026, with 861 B2B marketers reported. A search excerpt listed the same address under an October 2025 edition of the annual report, so confirm the current page before publishing. Used for the AI adoption, benefit, challenge, and effectiveness figures and for the trust builders.
- Clutch and Conductor. New Clutch and Conductor Data Reveals 87% of Content Marketers Increasing Budgets in 2026 as SEO Expands Into AI Search. Press release, February 17, 2026, on a January 2026 survey of more than 450 content marketers. Used for the research priority, the AI tool use, and the audience figures. Clutch runs a marketplace of B2B service providers, and Conductor sells search software.
- Ahrefs. Google Doesn't Punish AI Content; It Punishes Bad Content (331k Pages Studied). Study, July 27, 2026. Read through search excerpts only. Used for the sample, the 5.3% and 82.2% figures, and the detector caveat. The headline page count covers several samples. Ahrefs sells search marketing software and uses its own detector.
- Demand Gen Report. Coverage of 10Fold's survey of B2B technology marketers. Trade article, June 2, 2026, on a study released May 12, 2026, by 10Fold with Sapio Research. Read through search excerpts only. Used for the sample, the 52% figure, and the 40% and 33% figures. 10Fold sells content and communications services.
