An SEO growth strategy now has to earn two kinds of visibility from the same pages, a ranking in search results and a mention or citation inside an AI answer. Run both as one compounding channel, judge it on pipeline and customer acquisition cost (CAC), and treat AI answer presence as one of its leading indicators.
Most content teams we talk to run these as separate projects. One person owns keyword rankings in Ahrefs or Semrush. Someone else bought an AI visibility tracker after leadership asked whether ChatGPT recommends the company. Neither report connects to the pipeline number the CMO actually reviews, even though both teams are working on the same pages.
Where SEO fits in a growth strategy
SEO is the compounding channel in a growth mix. It ramps slower than paid media, but while a page stays visible in rankings and AI answers, it can keep attracting visits without an additional click charge. Give it its own pipeline and CAC targets in the plan, the same way you would for paid search.
Paid clicks can arrive within days and stop when spend stops. Organic visibility usually takes months to develop. Teams that fund both should evaluate them against the same pipeline and CAC targets, so neither channel is judged by a different yardstick.
Search rankings and AI answers share technical foundations, and each public page serves buyers, AI engines and training bots.
Why AI visibility and SEO are one strategy
Google's AI features draw on the same index as classic search. To appear as a supporting link in AI Overviews or AI Mode, a page has to be indexed and eligible for a snippet.
Ranking well helps, but it doesn't guarantee a citation. About 38% of pages cited in AI Overviews also rank in Google's top 10, based on 863K keyword results pages analyzed in March 2026.
Citation presence correlates with higher click-through rates. Across 53 brands tracked from January 2025 to February 2026, organic CTR ran about 2.1% when the brand was cited in an AI Overview and about 0.9% when it wasn't, against roughly 3.3% on queries with no AI Overview.
Clear intent, authority and well-structured pages support rankings. Those foundations can also support citations, though the CTR comparison doesn't establish that citations cause the difference.
The vocabulary is still settling, so it helps to pin it down:
- Answer engine optimization (AEO) is the work you do on pages and off-website signals so engines like ChatGPT, Perplexity, Gemini and Google AI Overviews use your content in their answers.
- Generative engine optimization (GEO) describes the same work and comes from a 2023 academic paper on optimizing content for generative engines.
- AI visibility is the outcome you measure. It captures whether the brand appears, and how it's described.
Inside AI visibility, a mention and a citation are different events. A mention is your brand named in the answer text. A citation is your URL used as a source.
The two move independently. An engine can name you based on review sites and forum threads without citing a page you own. It can also cite your guide while recommending a competitor.
One website for buyers, AI engines, and training bots
Every public page now has three readers. Buyers read it to make a decision. AI engines retrieve it while they compose an answer. Training bots crawl it, and what they collect may inform future models.
Write each page so a person can act on it, and as an editorial practice, write each section so it still makes sense when read on its own. Put every fact you want repeated in visible text. Keep the buyer-facing information on your public pages rather than duplicating it into separate AI-only pages.
Start from business outcomes and leading indicators
Work backward from the number leadership already tracks. If the goal is pipeline at a target CAC, SEO's job is to produce qualified demos, trials or purchases from organic and AI-referred visits, and everything upstream of that is a leading indicator. The mapping below gives you a starting set:
| Business goal | Lagging measure | Leading indicators |
|---|---|---|
| Revenue | Closed-won deals with an organic or AI touch | Pipeline sourced or influenced by organic |
| Pipeline | Demos, trials and form fills from organic and AI referrals | Conversions per money page, branded search growth |
| CAC | Blended CAC including content and tooling cost | Cost per organic conversion, trended quarterly |
| Visibility | Clicks to money pages | Impressions, rankings on buyer-intent queries, mention and citation presence in AI answers |
Set one target per row before anyone writes a brief. Leading indicators move months before pipeline does, and a reported trend line on them keeps the program funded through that gap.
How SEO works alongside paid, social and email
Each channel solves a different timing problem, so each one has a different job in the plan:
| Channel | Ramp time | Marginal cost | Best use |
|---|---|---|---|
| SEO and AI visibility | Months, and Google's own documentation says some changes take several months to register | Falls as pages keep earning without new spend | Capturing and shaping demand on buyer questions |
| Paid search | Days to weeks, and Google Ads guidance evaluates Smart Bidding over two conversion cycles or two weeks | Steady or rising per click, and it stops when spend stops | Harvesting existing demand, testing which terms convert |
| Paid social | Meta's learning phase usually ends after about 50 results in the week after the last significant edit, not a fixed performance ramp | Rises as audiences saturate | Creating demand and distributing content |
| Immediate to an existing list, weeks to warm up a new sending domain | Near zero per send, capped by list size | Nurturing, and sending new pages to people who already know you |
Use paid search as the research arm of SEO. Run bids on candidate keywords before you write for them, and only commission pages for terms that produce pipeline. Once an organic page earns the top spot on a term, pull the bid down and watch whether total conversions hold. That tells you how much of the paid spend was buying clicks you'd have gotten anyway.
How your business model changes SEO priorities
The same strategy weights different page types and KPIs depending on how you sell:
| Model | Priority page types | KPIs that matter |
|---|---|---|
| B2B SaaS | Product, comparison, alternatives, integration and use-case pages, backed by educational clusters | Demo and trial conversions, pipeline influenced, citation presence on category prompts |
| Ecommerce | Category, subcategory and product pages, plus buying guides | Organic revenue, product page impressions, structured data eligibility |
| Local | Google Business Profile, location pages and service pages | Calls, direction requests, bookings, local pack position |
Track pipeline influence and ask how buyers found you at the demo stage, because B2B SaaS deals run long and last-click attribution hides most of SEO's contribution.
Ecommerce depends on structure. Google's ecommerce guidance asks you to link menus to categories, categories to subcategories, and subcategories to every product, so crawlers can reach each one. Use Merchant listing markup on pages where shoppers can buy, and Product snippet markup where they can't.
Local runs on the profile as much as the website. Google ranks local results on relevance, distance and prominence, and you can't pay for a better local ranking. Complete profile details feed relevance, while reviews and links to your website feed prominence.
Build on search intent and keyword research
Rank keywords by commercial value and intent first, and volume after that. A low-volume query from buyers comparing vendors in your category is worth more than a high-volume definitional query that gets answered in an AI Overview and never clicked. CXL's SaaS keyword framework makes the same point by modeling the value of traffic through to the sale instead of maximizing volume.
Treat each keyword as one expression of a buyer question. "Expense software for mid-market," "best spend management tool" and "alternatives to [incumbent]" often express one question asked three ways, and they belong on one page. Pull the questions from sales calls, support tickets, demo recordings and People Also Ask, then group keyword variants under the question each one expresses.
Then map each question to a funnel stage:
- Top-of-funnel pages educate.
- Middle-of-funnel pages support comparison.
- Bottom-of-funnel pages supply validation, such as pricing context, proof and fit.
A question that doesn't map to any stage your buyers pass through doesn't get a page.
Organize content into Content Clusters
A topic is a subject your buyers ask about, a Content Cluster is a topic you decide to build authority in, and each opportunity is a specific page to create or rewrite inside that Content Cluster. The chain runs in that order:
- Topic
- Content Cluster
- Opportunity
Build each cluster around a pillar page that answers the broad question and links to supporting pages that each answer one narrower question. Supporting pages link back to the pillar and to each other where a reader would need the next answer. Use descriptive anchor text, since Google's link guidance advises against generic anchors like "click here" and asks for crawlable links.
"Topical authority" is an industry concept that Google's documentation never names as a ranking factor; the documentation talks about helpful, people-first content and clear structure instead. The practical case for Content Clusters comes from how AI features retrieve. Google describes AI Overviews and AI Mode issuing several related searches across subtopics for one query, so a Content Cluster that covers the subtopics gets surfaced across more of those searches.
Commit to a handful of clusters you can finish. Ten half-built clusters give engines ten thin signals.
Lift existing pages with on-page work and refreshes
The fastest SEO gains come from pages you already have, because they're indexed and carry history. Work through existing pages in this order:
- Pages ranking on page two: Put a direct answer in the first sentence under each heading, tighten the title to the query's intent, and add the subtopics the page skips.
- Internal links: Link from your strongest pages to the ones you want to lift, using descriptive anchor text.
- Cannibalizing pages: When two articles compete for one query with the same intent, keep the one with better traffic and links and 301-redirect the other into it.
- Declining pages: Diagnose before you touch them. A refresh fixes ranking decay, where position slips. It won't fix demand decay, where clicks fall while position holds, and if a drop started right after an edit, restore the earlier version.
Sort the portfolio by direction rather than level. A website-wide traffic average can hold steady while dozens of pages slide, and the slide only shows up when you group pages by trend. Change a page's published date only when you've made significant changes to it.
Fix technical foundations first
Technical problems cap everything above them, so audit them before you scale content. Check these four areas:
- Core Web Vitals: Good means LCP within 2.5 seconds, INP of 200 milliseconds or less, and CLS of 0.1 or less, measured at the 75th percentile of page loads.
- Mobile parity: Google indexes the mobile version of most websites, so the mobile page needs the same content, links and structured data as desktop.
- Crawlability: Links need to be real
<a href>elements, and primary content shouldn't load only after a click or swipe Googlebot won't perform. - Structured data: Google's AI optimization guide says AI features need no special schema, and it recommends keeping structured data for rich result eligibility.
AI crawler access is the newest check. OpenAI treats GPTBot, which collects training data, and OAI-SearchBot, which powers ChatGPT search, as independent robots.txt settings. Anthropic splits ClaudeBot from Claude-SearchBot the same way, and PerplexityBot handles Perplexity's search results. Allow the search bots if you want to appear in those answers, then decide separately on training bots. A minimal configuration that allows the named search crawlers while blocking GPTBot, ClaudeBot and Google-Extended looks like this:
User-agent: OAI-SearchBot
Allow: /
User-agent: Claude-SearchBot
Allow: /
User-agent: PerplexityBot
Allow: /
User-agent: GPTBot
Disallow: /
User-agent: ClaudeBot
Disallow: /
User-agent: Google-Extended
Disallow: /Blocking Google-Extended affects Gemini training and grounding, not inclusion in Google Search. User-triggered fetchers such as ChatGPT-User and Perplexity-User act on a person's request, and their operators say robots.txt may not apply to them.
Earn authority with links and digital PR
AI engines lean heavily on what other publications say about you. Earned media drove more than 84% of AI citations across more than 40 million AI responses analyzed in 2026. Links still matter for ranking, and coverage in trusted sources is a large share of what engines cite. A playbook for earning both covers four kinds of work:
- Linkable assets: Publish original data, benchmarks, calculators or templates that other writers need to cite.
- Digital PR: Pitch that data to the trade publications and "best of" roundups that cover your category, because those are the sources engines pull when buyers ask for recommendations.
- Partner content: Co-publish with integration partners and customers' communities, which puts your name next to your topic on domains you don't own.
- Reviews and community: Keep review profiles current and take part in community threads where buyers ask questions, while treating any single platform as volatile, since engines change how often they cite forums.
Track unlinked mentions separately from links, because they are different outcomes. Pursue a link when it fits the relationship and the publication's editorial process, and don't expect either one to guarantee a place in an AI answer.
Prioritize with the 80/20 rule and the 3 C's
The 80/20 rule says a small share of inputs produces most of the results. On most websites, a small share of pages and queries brings in most organic traffic. Ahrefs found this pattern on its own blog and recommends treating 80/20 as a heuristic rather than an exact ratio.
The 3 C's of SEO are content, code and credibility:
- Content is whether the page answers the question fully and accurately.
- Code is whether search engines and AI crawlers can fetch, render and parse it.
- Credibility is whether the page and brand show experience and earn trust from other sources.
Search Engine Land's Periodic Table of SEO uses all three as element groups.
Put the two together as a quick-win lens. Take the top pages and queries by traffic or pipeline, ask which of the three C's is the weakest on each, and fix the cheapest gap first.
A slow template on a high-converting page is a code fix you can ship in days. A missing comparison section is a content fix you can ship in a week. Credibility work takes longest, so start it early and don't wait on it.
Plan a realistic budget and timeline
Your budget has three lines, and each one buys something different:
- In-house team: A senior SEO or content lead's salary plus writers or freelancers, usually the largest line and the one that keeps judgment inside the company.
- Agency retainer: Monthly SEO agency fees typically run from $2,000 to $20,000. The pricing guide was updated September 2026.
- Software: An SEO suite, rank tracking and an AI visibility tracker, with Ahrefs plans listed from $129 to $1,499 per month in November 2026, and vendors change plans often.
Budget share should follow stage. Before product-market fit, keep SEO lean and limited to bottom-of-funnel pages buyers need to decide, because the payback is too slow to carry the company. At the growth stage, fund one or two full Content Clusters and the technical work under them. At scale, assess whether refreshing and consolidating the existing portfolio deserves more budget.
A realistic SEO ramp runs like this:
- Month 1: Fix technical blockers, set up AI referral tracking, and baseline rankings, impressions and AI answer presence.
- Months 2–3: Refresh page-two pages and add internal links, then judge early results after a few weeks, as Google's starter guide suggests.
- Months 4–6: New cluster pages may begin earning impressions and early rankings.
- Months 6–12: Pipeline from organic starts to become measurable. CXL's B2B framework puts significant ROI at 6–9 months for established businesses with domain authority above 50, and at 9–12 months or longer for new businesses or competitive industries.
Paid search, by contrast, produces clicks in days and stable bidding within weeks. Present both curves to finance together, so the slower one doesn't get cut before it pays.
Measure SEO and AI visibility ROI
Clicks alone undercount SEO's effect now. 68% of US Google searches in January through April 2026 ended without a click, based on desktop and mobile panel data. Buyers form opinions inside results pages and AI answers before they ever visit, so your measurement needs a layer that sees those answers.
Build the framework from leading to lagging:
- Impressions and rankings on buyer-intent queries, including AI Overview and AI Mode appearances.
- Mention and citation presence in AI answers on the prompts you care about.
- Clicks and AI referral sessions to the pages that answer your target questions.
- Conversions from those sessions, plus self-reported attribution.
- Pipeline and CAC.
Guidance on AEO measurement uses a similar set, covering share of answer, citation presence, brand mention frequency and sentiment, and AEO-influenced conversions.
Expect real attribution limits and work around them. In GA4, create a custom channel group with a regex that matches chatgpt.com, perplexity, gemini.google.com, claude.ai and copilot, and place it above Referral so the grouping catches those sessions.
Some AI traffic still lands in Direct when referrers are stripped. Search Console counts AI Overview and AI Mode clicks inside the Web search type with no separate filter, and its newer generative AI reports show impressions only, rolling out to a subset of websites. Add a "How did you hear about us?" field to the demo form, because buyers who shortlisted you inside ChatGPT often show up in analytics as branded search or direct.
For the AI answer layer, sample prompts daily and report seven-day ranges, because answers to the same prompt change from run to run. A single weekly reading overstates precision.
Our own measurement product, CheckThat, measures AI visibility across brands, B2B software categories and AI responses. Its measurement model, which is still being built out, uses four dimensions as shared vocabulary.
Presence covers whether AI recommends you, and Reputation covers what the outside world says about you. Perception covers what story AI tells and whether it's accurate, and Influence covers how much you control the sources behind the answer.
Read the signals together. If pages that answer target questions gain impressions but not clicks, rework the answer and the page's next step.
If AI referral sessions rise but demos don't, look at fit and the conversion path. Each published page also adds data, which shows you the next set of directly relevant questions to cover, along with questions adjacent to your market and buyers.
Put the strategy into practice
Run the steps in this order, because each one depends on the one before it:
- Set revenue, pipeline and CAC targets for organic, and assign a leading indicator to each.
- Decide what SEO does relative to paid, social and email, and use paid search to test keyword value.
- Pick priority page types and KPIs for your business model.
- Turn keyword research into a map of buyer questions by funnel stage.
- Choose a handful of Content Clusters and plan pillar and supporting pages.
- Fix Core Web Vitals, mobile parity, crawlability and AI crawler access.
- Refresh and interlink existing pages, consolidate cannibalizing ones and triage declines.
- Start earning mentions and links in the sources engines cite.
- Prioritize weekly with the 80/20 rule and the 3 C's.
- Measure from impressions to pipeline, sample AI answers as ranges, and feed what you learn back into the question map.
The GrowthX platform runs this loop in one workspace, from Context, which holds who you sell to and how you sound, through Portfolio, Opportunities and Creation to Insights, with your team setting direction, approving every draft and deciding which findings should change Context. If maintaining this workflow across separate tools is the bottleneck, book a demo. See how GrowthX connects SEO and AEO around your buyer questions
