Content Marketing Funnel Mapping for B2B SaaS
AI answer engines now control awareness-stage discovery before Google searches even start.
The TOFU-to-BOFU model still describes how buyer intent develops: awareness gives way to consideration, and consideration gives way to a decision. What has changed is the set of channels that carry a buyer through each of those stages, because the model assumed Google sat at the center of every one of them, and that assumption no longer holds. B2B SaaS buyers now finish most of their research before a sales rep ever gets a call, so content itself does the persuading across a buying cycle that can run for months. That cycle rarely involves one person making a single linear trip through the funnel. Buying committees include multiple stakeholders moving at once, a VP spending weeks on evaluation while other sign-offs happen in parallel. The same funnel stage has to serve several different readers with different concerns at the same time. Layered on top of that is a shift in where informational searches even land: AI Overviews and zero-click results now absorb a large share of queries that once sent a visitor to a company's website, so content built only to earn a Google click is already missing a growing slice of the awareness-stage audience. A funnel mapped only to Google organic traffic is a partial one, with the biggest blind spots sitting exactly where buyers are now doing the most research on their own. The fix is a new map, one that accounts for the actual surfaces buyers use before they ever type a brand name into a search bar.
Where B2B SaaS buyers go at each funnel stage
Each stage of the funnel now has its own cluster of channels, and that cluster looks different at awareness than it does at consideration or decision. At the awareness stage, buyers encounter the category through AI answer engines such as ChatGPT, Perplexity, and Google AI Overviews, alongside industry publications, LinkedIn posts from people they trust, and Reddit threads, often before they have settled on a product category, let alone a shortlist of vendors. AI tools have become a primary channel for this kind of early research, so a brand that does not show up in an AI answer is missing from consideration before the buyer reaches any other channel. "Best X" listicles have historically pulled an outsized share of ChatGPT citations, though that share dropped sharply after a late-2026 update to the model, a reminder that awareness-stage discovery through AI already runs on the comparison and category content most teams file under mid-funnel, not top-of-funnel.
By the time a buyer reaches consideration, the channels shift toward proof rather than education. Peer-review platforms such as G2 and Capterra category pages, Reddit product discussions, comparison articles, case studies, webinars, and ROI calculators take over, each one built around social proof and specifics tied to the buyer's actual use case rather than general explanation. Most B2B buyers say they prefer thought leadership and educational material over anything that reads as promotional at this point, where "educational" means depth tied to a real use case and backed by evidence, the kind of depth an AI model cannot already answer in a few sentences. The question a buyer is asking has changed from "what is this category?" to "which one fits my situation better?", and the sources that answer that question honestly, with real tradeoffs laid out, earn more trust than a vendor's own feature page ever will.
Decision-stage buyers land back on the vendor's own site: demo pages, free trial flows, testimonials, and direct vendor comparisons. By the time they arrive there, though, most of the persuading has already happened somewhere else. A buyer who found a brand through an AI answer at the awareness stage and then checked it against reviews on G2 during consideration shows up at the vendor site mostly convinced already, so the job of that bottom-funnel content is to clear away the last few objections rather than make the initial case. This is where the risk compounds for teams that have not adapted: a brand that still ranks well on Google but is invisible in AI answers, absent from peer-review platforms, and quiet on Reddit is winning a shrinking slice of buyer attention while losing the share of attention that is actually growing. The channel map at each stage is the foundation that every content assignment from here forward has to be built against.
Why AI answer engines have become the highest-stakes awareness surface
Among all the surfaces now feeding the funnel, AI answer engines carry the most weight at the awareness stage, because they compress the buyer's search process into a single recommendation before any other channel gets a turn. When a buyer asks an AI model which tools solve a given problem, the model doesn't hand back a list of links for the buyer to sort through. It names a brand directly, and that answer effectively becomes the buyer's starting shortlist before any independent research even begins. Missing from that answer means missing from consideration.
The platforms do not all work the same way, and that difference matters for anyone trying to plan for it. Gemini leans heavily on brand-owned websites when it chooses what to cite. ChatGPT draws a meaningful share of its citations from what might be called internet consensus: third-party directories, listicles, and review aggregators, even if that does not make up the majority of what it cites overall. Perplexity, in turn, puts weight on community content, especially Reddit, along with how fresh a page is and how much topical authority it carries. Only a small share of domains get cited by both ChatGPT and Perplexity, so a brand that shows up clearly on one of these platforms can be completely absent from another, with no overlap to fall back on.
Branded web mentions, references to the company by name across the web, drive citation share more than backlinks, the metric SEO teams have optimized for over a decade, and the correlation between those mentions and AI visibility runs well ahead of the correlation between backlinks and AI visibility. A large share of what large language models cite comes from the opening portion of a page, so definitions, authoritative claims, and proof points need to sit early rather than get buried after a long preamble, which affects where credibility signals should sit inside a document. Traditional funnel analytics were built to measure clicks and conversions, but this new layer rewards something different: citation. Letterstory's measurement layer tracks whether ChatGPT, Claude, Gemini, and Perplexity actually name and cite a brand at the awareness stage, which turns citation presence into something a team can measure rather than guess at.
A brand can get cited and recommended by an AI answer engine, only for that same engine to point the buyer toward a competitor when the moment of actual purchase arrives; this failure mode deserves its own name: Ghost Rankings. The brand shows up in the conversation but still loses the sale. Visibility in an AI answer is necessary but not sufficient, and measuring that gap is the next problem to solve.
What content gets cited by AI
Citation by an AI model is not random, and it is not simply a function of how authoritative a domain already is in Google's eyes. Specific structures and signals inside a piece of content measurably raise or lower its odds of being cited. Writing for AI citation is its own discipline, separate from writing to rank on Google. Content built around quotations from credible sources, real statistics, and citations to primary evidence earns a noticeably larger share of AI answers than content that skips those elements, and the lift from each one tends to stack rather than cancel out.
Author attribution turns out to carry real structural weight. Pages without a named author tend to drop in AI retrieval rankings, while sites linking to detailed author bios gain measurable ground in visibility, which puts E-E-A-T signals, the kind SEO teams have often treated as nice-to-have, in a load-bearing role for AI citation specifically. Because a large share of branded AI citations trace back to third-party surfaces rather than the brand's own site, seeding coverage on G2, Reddit, and industry publications matters as much as anything published on the company blog. The ecosystem outside the brand's own domain is the larger part of what these models are actually reading.
Comparison content does double duty in this landscape. An "X vs Y" or "X alternatives" piece matches the kind of high-intent query a buyer types late in the funnel, and it also appears in AI answers when a model is helping someone weigh a short list of options, so a well-built comparison article can earn citation at both the awareness stage and the decision stage at once. The operational split between AEO and GEO is that AEO optimizes a passage so a model can lift it out as a direct answer, while GEO builds the broader authority signals that make a model comfortable recommending the brand at all. Both jobs are necessary, and they call for different content investments, which is the foundation the next section builds its architecture on.
Building a two-surface content architecture: Google organic and LLM citation as separate editorial tracks
A content calendar built for both Google organic and LLM citation needs two separate editorial tracks, each with its own brief, its own success metric, and its own structural rules. Treating the two as one job produces content that underperforms on both fronts at once. The Google organic track covers commercial category pages, high-intent comparison content, and transactional pages, shaped by search volume, keyword competition, and commercial intent, and measured by clicks and pipeline. The LLM citation track covers definitional content, pages that map out how entities relate to each other, brand authority signals, and factual grounding pages, shaped by topical depth, credibility signals, recency, and how closely the content matches the kind of query a model is answering, and measured by whether a model actually retrieves and cites it.
One proof point makes this concrete. A fresh domain registered in November 2025, with essentially no backlinks, was set up as a grounding page under this kind of structure, and within three weeks it was already showing up as a cited source in ChatGPT, Perplexity, and Google Gemini. That result points to the grounding page as a structural piece to build deliberately rather than treat as an afterthought. A grounding page is a purpose-built HTML page, ideally placed at a path like /grounding-page/ or /facts/, that states the brand's facts in a form a model can pull cleanly. It defines the brand rather than functioning as a marketing page or a sales page meant to persuade anyone of anything. The strongest version combines plain visible text as the primary source with JSON-LD structured data behind it, and links from the site's footer sitewide so it functions as a constant authority signal rather than a one-off post. A grounding page also lowers the odds of a model getting basic facts about the brand wrong, since a brand that never publishes its own facts in a retrievable form leaves the model to reconstruct them from third-party sources, where small errors tend to pile up over time.
Off-site seeding matters just as much as this architecture's on-site half. Because branded AI citations come mostly from third-party surfaces such as reviews, listicles, press coverage, and directories, the editorial calendar needs a deliberate, ongoing program for building presence on G2, Reddit, and industry publications alongside a steady stream of posts on the company's own site. This isn't a PR function tacked onto the content plan; it functions as a structural piece of the citation architecture itself, because models are reading that third-party ecosystem more closely than they are reading the brand's owned pages. The gap between the funnel teams think they have and the one that actually shapes buyer research is widest right here, at awareness, where AI answer engines now intercept a buyer's research before any other channel gets a chance. Brands that have published only for Google organic visibility can now check whether they are actually being cited in those AI responses, which shows the real shape of the funnel rather than the one everyone assumed was still in place. The editorial calendar now runs on two separate inputs instead of one: Google opportunity, meaning search volume, keyword competition, and commercial intent, and LLM citation potential, meaning topical depth, credibility signals, recency, and how well the content matches what a model is trying to answer. These two inputs do not need to line up with each other, and most of the visibility teams are leaving on the table comes from trying to force that correlation instead of planning for each one on its own terms.
Assigning content types to funnel stages and discovery surfaces, the full mapping
Mapping a B2B SaaS funnel in 2026 means assigning content not only to a stage of buyer intent but to the specific surface where that intent actually plays out, since the same underlying question looks different depending on whether it shows up on Google, inside an AI answer engine, on G2, or on Reddit, and each of those surfaces calls for its own content form.
At the awareness stage, across AI answer engines, Google organic, LinkedIn, and Reddit, the content mix starts with grounding pages and definitional material built for LLM retrieval, laying out what the brand is and does in terms a model can pull reliably. "Best X" listicles still belong here too, since they remain the single most cited page type in ChatGPT responses and continue to serve high-intent discovery queries on both Google and AI platforms. Thought leadership and content built around original data earn a durable place in this stage as well, since a model has no way to replicate a point of view rooted in proprietary data or direct customer insight, which makes that kind of content one of the few assets in the funnel that genuinely resists being answered away by AI. Alongside all of this, a steady presence on Reddit and LinkedIn helps seed the category-level conversations buyers have before they have settled on a shortlist, work that functions less as brand promotion and more as brand-adjacent credibility building.
Consideration-stage content shifts toward peer-review platforms, comparison content, webinars, and case studies. Case studies need concrete detail, actual numbers, screenshots, and direct quotes from customers, because a buyer at this stage is weighing evidence rather than learning about the category for the first time, and a vague case study simply doesn't clear that bar. Presence on G2 and similar peer-review platforms functions as a structural piece of the citation architecture rather than a side project for the PR team, since models read those pages and buyers check them independently of anything the brand publishes itself. Webinars and live product demos do double work at this stage too, qualifying a prospect while educating them at the same time, since buyers who would rather avoid a sales call will often still show up to a demo event, which lines up with how self-directed most B2B SaaS research has become.
At the decision stage, the vendor's own website, demo flows, testimonials, and vendor comparison pages carry the load, but by the time a buyer reaches them, most of the persuasion has already happened elsewhere. These pages exist to close the remaining gap for someone who has already formed a favorable view of the brand, not to introduce the brand for the first time. Despite that leverage, this is consistently the stage teams underinvest in relative to what it returns: teams that build out awareness and consideration content well, then leave the bottom of the funnel thin, end up losing deals in the final stretch, right where the buyer was already most of the way to saying yes.
