Reading time: 16 minutes
To do content marketing in the age of AI, start with the questions customers ask before they buy. Answer each one on a page with your own data or experience. Then distribute it off-site and measure your brand’s presence in AI answers, not just traffic. This guide covers each step, from plan to measurement.
In short
- In the US, 68% of Google searches ended without a click in early 2026, according to SparkToro.
- In a 2026 study of more than 1 million citations from three AI platforms, lists, articles and product pages made up more than half.
- In CMI’s survey of B2B marketing professionals, mostly in North America, 87% of those who create content with AI gained productivity, but only 39% gained performance.
- A content marketing plan measures three layers: presence in AI answers, traffic and business results.
What is content marketing and what changed with AI answers?
Content marketing means attracting and retaining a defined audience with useful, consistent content until some of that audience becomes customers. With AI, however, the place where the answer appears has changed. Google and assistants like ChatGPT answer on their own, so content needs to become the source of those answers.
The Content Marketing Institute defines the practice as a strategic approach. A company creates and distributes valuable, relevant and consistent content for a clearly defined audience. The end goal is to drive profitable customer action. In other words, the goal hasn’t changed, but the path between the question and the website has shrunk.
Meanwhile, in Brazil, fewer companies use the strategy. In RD Station’s Panorama de Marketing e Vendas 2026 (Marketing and Sales Panorama), 58% of companies used content marketing in 2025, down from 71% in 2022. Similarly, SEO and organic traffic fell from 55% to 41% over the same period. That question was answered by 1,351 professionals.
How does content marketing relate to inbound, SEO and AI?
These four terms are close, but each one measures something different.
| Term | What it is | What it usually measures |
|---|---|---|
| Content marketing | Useful content to attract and retain a defined audience | Leads, sales and retention |
| Inbound marketing | Attracting customers with content, without interrupting them with ads | Leads at each stage |
| SEO | Helping search engines understand the site and users find it | Rankings and clicks |
| AI visibility | Showing up and being cited in AI answers | Mentions and citations |
The definitions of content marketing and inbound follow CMI and HubSpot, while the metrics column is a suggestion from this guide.
Inbound is the method, and content is what feeds it. According to the same RD Station survey, from a company that sells marketing automation software, inbound shows up in 76% of operations. SEO helps searchers find the content. AI visibility also adds a new question to content marketing: does the answer cite your brand?
Does content still pay off with zero-click search?
It still pays off, but through a different path. After all, fewer searches end in a click, and part of the decision happens inside the answer. Content marketing now earns its value from what it teaches Google and AI about the brand and from bottom-of-funnel leads. As a result, the volume of blog visits matters less.
In the US, 68.01% of Google searches ended without a click from January to April 2026, according to SparkToro, using Similarweb data. In 2024, the figure was 60.45%, but from a different browsing panel. Rand Fishkin, SparkToro’s cofounder, announced a book about zero-click marketing. He recommends investing in influence and brand without relying on website visits. Still, in his view, a site’s content keeps shaping what AI answers, even as traffic falls.
See how AI Mode and AI Overviews affect clicks, with the studies and Google’s response.
Can content marketing bring in leads through AI itself?
The return can come through AI itself. Grow & Convert is an agency that sells bottom-of-funnel content. In May 2025, it reported that 71% of its new leads in the past month came through ChatGPT. That is one company’s data, from a single month. Still, it shows the path: the buyer asks AI, and AI may point to whoever published specific content about their problem.
In practice, the question changes: instead of measuring how much traffic the blog brings, decide which content is still worth producing when search doesn’t generate a click.
How to do content marketing: step by step
The process has eight steps, from goal to update. With AI, three points change from traditional content marketing. First, the question list includes what customers ask AI. Second, each page brings something only the company knows. Finally, measurement also includes the brand’s presence in AI answers.
- Set the goal and the metric. Leads, opportunities, revenue or presence in AI answers. Without a metric, the calendar turns into a list of topics.
- Describe who buys. Talk to sales and support. Then write down job titles, pains, objections and the words customers use.
- List the real questions. Use Search Console, “People also ask” and the sales team. Include the questions customers would ask an AI assistant.
- Classify each question by buying stage and choose the format. The next section shows how.
- Build the calendar from the bottom of the funnel up. Start with comparisons, pricing, use cases and pains.
- Produce with sources and review. Each page answers the question at the start and brings data, a case or an opinion of its own.
- Distribute off-site. LinkedIn, newsletters, the press and communities.
- Measure and update every quarter. Compare AI presence, traffic and leads with the starting month.
How do you plan content marketing by funnel stage?
Plan from the bottom up. After all, for broad questions like “what is,” AI explains the topic without needing to recommend a company. On the other hand, comparison and purchase questions still lead to brands and websites. That is why most of the calendar should cover pains, comparisons, pricing and use cases. The top of the funnel is for building authority.
Grow & Convert, an agency that sells this type of content, has advocated this order since 2018. With AI search, it sees even less value at the top. Based on its experience, bottom-of-funnel keywords convert 10 to 25 times better than top-of-funnel ones. However, that is the agency’s estimate, not a study with a published method.
At NotHuman, for example, we follow this order. The plan of about 150 articles for the NotHuman blog started from the searches our audience types. We used Google autocomplete in Brazilian Portuguese for each pain point of our persona and organized the topics into a pillar, hubs and specific articles. Decision and solution articles come first; discovery takes up at most about 25% of each branch.
Which format does AI cite at each funnel stage?
The stage of the question also signals which format AI tends to cite. In March 2026, the Wix AI Search Lab analyzed 75,000 answers from ChatGPT, Google’s AI Mode and Perplexity to questions without a brand name. The author is a researcher at Peec AI, which sells AI brand monitoring, while Wix sells a website builder. The intent of the question predicted the cited format better than the industry or the model did.
| Stage | Typical question | Most cited format |
|---|---|---|
| Discovery | “What is,” “how does it work” | Article (45% of citations) |
| Consideration | “Best X,” “X or Y” | List (41%) |
| Decision | “X price,” “hire X” | Product page (25%) |
The table is our interpretation, since it maps the study’s intents (informational, commercial and transactional) to funnel stages. The study does not report the country or the language of the prompts.
Which formats does AI cite most: articles, FAQs, comparisons or original research?
Lists, articles and product pages lead AI citations: lists had 21.9%, articles 16.7% and product pages 13.7%. In the Wix AI Search Lab study, the three made up more than half of over 1 million citations. In contrast, comparison and alternatives pages stayed below 3%. Each format answers a type of question and needs a different kind of care.
Blog articles: the answer up front and something of your own
The article AI cites answers the question right at the start of each section. It also brings something that isn’t on other sites. It can be data, a test, a customer case or a signed opinion. The rewriting tactics with the most evidence, such as quotes, statistics and source citations, are in the guide to what GEO is.
A good blog article structure that Google and AI cite has a clear title, the answer up front and self-contained sections. Original research and customer cases fit here as the data only your brand has.
Comparisons and alternatives: when AI recommends the competitor
In the comparison phase, AI cites a lot of lists. In fact, for commercial questions, lists had 41% of citations in the same study. However, the study only split list authorship for professional services, among the 1,000 most cited URLs. There, third-party lists had 80.9% of list citations, and brands’ own lists had 19.1%. Therefore, in that sector, appearing on another site’s list tends to be worth more than publishing your own. Before you publish yours, see the risks and criteria of a self-promotional listicle.
Your own comparison page is still useful for buyers who reach your site. An honest competitor comparison page shows criteria, pricing and who each option is for. State clearly which type of customer your offer is the best choice for, and which it isn’t.
FAQ, About us and product pages: the official source
When people look for a website or a service in their area, AI cites company pages more. In the same study, for navigational or local questions, product pages, category pages and homepages made up 54% of citations. So pricing, offer, service area and differentiators should appear on the site, up to date and consistent on every page.
An FAQ on your site that answers pricing, timelines and coverage gives AI an official source to talk about the company. The About us page does the same for the company’s history, team and numbers.
How does content marketing build authority with Google and AI?
Content builds authority when it shows who wrote it, how they know what they say and why the page exists. Google sums up these signals as E-E-A-T: experience, expertise, authoritativeness and trustworthiness. For AI, however, authority also comes from outside, from what the press, customers and experts say about the brand.
According to Google, E-E-A-T isn’t a specific ranking factor, and trust is the most important of the four aspects. Google suggests evaluating each piece of content in terms of “Who, How, and Why”: who created it, how it was made and for what purpose. It also warns that fabricating authors, with AI-generated headshots or false credentials, is a form of deception.
In practice, this calls for a real author or an identified team and a visible update date. It also calls for a source for every number and an account of how the company tested what it recommends. First-hand experience, clear authorship and verifiable sources are the foundation of E-E-A-T in content.
How do you produce content with AI without losing quality?
Use AI in content marketing for research, structure and the first draft. Then leave to people what AI doesn’t have: experience, proprietary data, fact-checking and brand voice. Google doesn’t penalize the use of AI in itself. The problem, however, is pages produced at scale with no value for the reader.
The “Insights for 2026” edition of the Content Marketing Institute survey covered 1,015 B2B marketers, mostly in North America. The survey ran from June to August 2025. CMI sells content marketing courses and events, and Storyblok, which sells a content management system, sponsored the study. Almost 9 in 10 use AI to write marketing content. Among them, 87% say productivity improved. However, only 58% saw quality improve, and 12% saw it get worse. Content performance improved for 39%.
Robert Rose, CMI’s chief strategy advisor, summed up the result:
“As it stands, for the most part, AI helps marketers type faster, not think better.”
Robert Rose, Content Marketing Institute (2025)
How do you use AI without breaking Google’s rules?
According to Google, generative AI can be particularly useful for researching a topic and adding structure to original content. However, generating many pages without adding value for users may violate its policy against scaled content abuse. Google also asks you to check facts manually before publishing. Google’s guidance on AI-generated content extends this review to titles, meta descriptions, structured data and image alt text.
- AI does: initial research, structure, drafting and title variations.
- People do: the angle, proprietary data, examples, checking every number and the final review.
How do you distribute content off-site?
Distribute where buyers and AI look: LinkedIn, newsletters, the press, communities and review sites. After all, most of what AI platforms cite comes from outside the brand’s website. That is why distribution also needs to earn mentions in independent sources.
In B2B, LinkedIn leads. In the same CMI survey, 96% of companies create thought leadership content. For 76%, LinkedIn is among the most effective channels for it. But 37% say fewer than 5% of the company’s experts contribute. It pays to get the people who serve customers, sell and deliver the work to write, because they have the experience the content needs to show. The plan for companies that sell to other businesses is in the guide to B2B content marketing.
For AI, mentions on third-party sites carry weight. See which sources AI cites most, from social networks to the press, and how to earn a presence there without breaking rules. Digital PR means turning data, a study or an expert opinion into a story published by third parties.
When should you update old content?
Update when the data gets old, when the page loses rankings or when AI starts citing a more recent competitor. For bottom-of-funnel pages, review every quarter. In other words, updating means correcting numbers, adding what changed and cutting what no longer makes sense. Changing the date isn’t enough.
In addition, AI assistants tend to cite newer content. In July 2025, Ahrefs analyzed 16.975 million URLs cited by AI assistants and by Google Search. Content cited by ChatGPT, Perplexity, Gemini and Copilot was, on average, 25.7% fresher than content in organic results. However, AI Overviews were the exception, with an age close to that of organic results. Ahrefs sells SEO tools.
Before writing a new post, see whether you can update outdated content on the same topic.
How do you measure content marketing results beyond traffic?
Measure content marketing in three layers: presence in AI answers, traffic and business results. The first counts in how many answers the brand appears and is cited. The second separates visits that come from AI assistants. Finally, the third connects content to leads, opportunities and revenue, which is what pays for the program.
| Layer | What to measure | Where |
|---|---|---|
| AI presence | Mention, citation and recommendation | Fixed prompts, repeated every month |
| Traffic | Visits from AI and from search | GA4 and Search Console |
| Business | Leads, opportunities and revenue | CRM |
In May 2025, Google itself, an interested party, said that clicks from pages with AI Overviews are higher quality, without publishing the numbers. So it recommends looking at sales, sign-ups and engagement, not just the number of clicks.
Since May 2026, GA4 has had an AI Assistant channel, with visits from assistants such as ChatGPT and Gemini. However, clicks from AI Overviews and AI Mode remain in organic search. The protocol to measure brand mentions and citations in AI covers prompts, repetitions and margin of error. To separate these visits in analytics, see how to set up an AI traffic channel in GA4.
What we found when we analyzed 12 Portuguese-language guides on content marketing
We analyzed 12 Portuguese-language pages that showed up in searches for “como fazer marketing de conteúdo” (how to do content marketing) and variations. Only 3 treat AI answers, such as AI Overviews and ChatGPT, as a place where content needs to appear. In addition, six don’t even mention AI in the text.
None includes brand mentions or citations in AI answers in its list of metrics. Similarly, none uses the term zero-click search, and only one mentions the drop in organic traffic from search. Four had not been updated since 2024 or earlier.
Frequently asked questions
Inbound marketing is the method of attracting customers with content, without interrupting them with ads. Content marketing, on the other hand, is what feeds that method: useful, consistent content to attract and retain a defined audience. SEO comes in so the content can be found in search.
Google doesn’t penalize the use of AI in itself. Generating many pages without value for users, however, may violate Google’s policy against scaled content abuse. Google also asks you to check facts manually before publishing a text made with AI.
In professional services, appearing on another site’s list tends to be worth more. In the Wix AI Search Lab study, third-party lists had 80.9% of list citations in that sector, and brands’ own lists had 19.1%. See which sources AI cites most.
Want to know how your brand shows up in AI answers today?
NotHuman is opening spots for agencies and brands to track how they appear in AI answers. Join the waitlist and we’ll let you know when a spot opens up for you.
Main sources
- Content Marketing Institute, B2B Content and Marketing Trends: Insights for 2026 (2025).
- RD Station, Panorama de Marketing e Vendas, channels and operations section, in Portuguese (2026).
- SparkToro, zero-click Google searches in the US (2026).
- Wix AI Search Lab, the content types most cited by LLMs (2026).
- Grow & Convert, bottom-of-funnel content in AI search (2025).
- Google Search Central, guidance on using generative AI content on your website (2026).
- Ahrefs, study on whether AI assistants prefer to cite fresh content (2025).
