Your SEO Playbook Isn't Ready for AI Search. Here's How to Fix 6 Critical Gaps.

AI search didn’t just change where you rank. It changed which efforts are worth making in the first place. Three things you should be doing more of. Three you should be doing less. Let’s get into it.

Here’s the uncomfortable truth most SEO content won’t tell you: the tactics that built your current organic traffic aren’t the same ones that will build your AI search visibility. And if you’re running the same playbook from 2023 into an AI-dominated search landscape, you’re putting serious effort into the wrong places.

Adam Tanguay, Head of Growth at Jordan Digital Marketing, laid out the shift clearly in a Search Engine Land piece published this week: AI search and traditional SEO overlap significantly — but they are not the same game. The priorities that move the needle in one don’t always move the needle in the other.

So here’s the direct version: three things to do more of, three things to do less of, and the specific reasons each one matters right now.

The Three Priorities You Should Be Doing More Of

1. Build Brand Entity Signals – Not Just Content

This is the one most teams are getting completely wrong, and it’s costing them AI visibility they don’t even know they’re missing.

AI systems don’t just retrieve content, they recognize entities. Before an LLM will confidently cite your brand in an answer, it needs to “know” who you are: what your brand stands for, what category you operate in, who your key people are, and how consistently that information appears across the web. Without strong entity signals, you can publish genuinely excellent content and still get bypassed in AI answers because the model can’t confidently attribute it to a recognized brand.

What this means practically: your brand information needs to be consistent, complete, and cross-referenced across Wikipedia, LinkedIn, Crunchbase, industry directories, and any publication or platform where LLMs pull entity data. Discrepancies between these sources create ambiguity. Ambiguity reduces citation confidence. Reduced citation confidence means someone else gets mentioned instead of you.

The connection to PR is something most SEO teams haven’t fully absorbed yet. Earned media mentions are now entity-building signals, not just brand awareness plays. Every time a respected publication mentions your brand accurately and in context, that’s another data point reinforcing your entity in the training data AI models use. Your SEO team and PR team need to be working from the same brief, not in separate silos.

Author entities matter too. If your content is bylined by genuine subject-matter experts with their own credible web presence, real LinkedIn profiles, conference speaking history, published work elsewhere, that adds E-E-A-T signals that AI models weigh. Anonymous content or content attributed to generic “team” bylines is structurally weaker in an AI citation context.

The move: Audit your entity footprint this week. Search your brand, your key authors, and your core product names across Google’s Knowledge Graph, Wikipedia, LinkedIn, and Crunchbase. Every gap or inconsistency is a citation risk.

2. Stop Thinking in Keywords. Start Thinking in Topic Ownership.

AI systems don’t favor pages that rank for individual keywords. They favor sources that demonstrate comprehensive authority across a topic. The distinction sounds subtle. The strategic implications are enormous.

A thin content footprint, one or two pages on a subject, surrounded by gaps was survivable in traditional search if those pages were well-optimized and well-linked. In AI search, that same footprint reads as superficial. The model assesses the depth of your coverage across a topic, not just the quality of a single page.

The question that should be driving your content strategy has changed. It’s no longer “what do we rank for?” it’s “what topics do we want AI systems to associate us with?” Those are different questions, and they produce different content plans.

Topic clusters, interconnected content that covers a subject from multiple angles with strong internal linking, are now table stakes. Not because Google said so in a blog post, but because that’s how AI models assess topical authority when deciding which sources to draw from. Internal linking matters more than it used to because it signals the topical relationships between pieces of content for LLM ingestion, not just for PageRank.

The payoff is also different. A strong content cluster can generate broad AI visibility across multiple related queries, not just traffic to individual pages. When a buyer is researching a category in ChatGPT or AI Mode, a brand with comprehensive topic coverage appears across that entire research journey. A brand with scattered individual pages appears occasionally, if at all.

The move: Pick the two or three topic areas most critical to your pipeline. Map your current content coverage against every major subtopic and question a buyer in that space would have. The gaps in that map are your content investment priorities, not your keyword rank gaps.

3. Earn Mentions in the Communities AI Actually Trusts

This is the one that makes traditional SEO teams most uncomfortable, because it doesn’t produce a clean metric or a trackable link. Do it anyway.

LLMs are trained on the broader web, including forums, communities, and platforms that traditional SEO largely ignored because they don’t pass PageRank. Reddit, Quora, niche industry forums, professional communities: these surfaces carry weight in AI training data in ways that directly influence whether AI models treat your brand as credible and well-regarded.

The mechanism is pattern-matching. AI models assess what the web says about your brand across many sources, not only what ranks in Google. Owned content alone cannot manufacture the signal that comes from third parties talking about you positively and authentically in the communities where real users have conversations. And as Tanguay notes, Reddit carries particular weight because LLMs have been heavily trained on it and treat its content as authentic user sentiment.

This is not an invitation to spam Reddit with brand mentions. That strategy has been documented extensively and it ends in penalty, ban, and public embarrassment. The play is genuine participation showing up in the communities where your customers and prospects are asking questions, contributing real expertise, and building a reputation that AI models can pattern-match as positive organic sentiment.

Monitoring unlinked brand mentions is now as strategically important as tracking backlinks. Where is your brand being discussed? Are those discussions accurate? Are they positive? The answers to those questions are shaping your AI visibility whether you’re tracking them or not.

The move: Identify the three to five online communities where your target buyers have real conversations. Find where your brand is mentioned currently. Assign someone to monitor and participate in those communities consistently not to promote, but to build genuine presence.

The Three Priorities You Should Be Doing Less Of

4. Stop Chasing High-Volume Keywords With Thin Content

This one stings because volume has been the north star of SEO planning for a decade. But in an AI Overviews world, high-volume informational queries are increasingly the queries where clicks disappear entirely.

AI Overviews absorb the click for generic informational queries. Ranking number one for a broad head term now frequently means you’ve put significant effort into attracting traffic that never arrives because the user got their answer before they ever saw your blue link.

Volume alone is no longer a proxy for opportunity. A query with 50,000 monthly searches that reliably triggers an AI Overview may deliver less actual traffic than a query with 2,000 monthly searches where users still need to click through to accomplish something. The metric that matters is no longer search volume, it’s whether a user will still need to visit a site after AI has answered the query.

The content that survives this shift is specific, authoritative, and answers a narrower question better than anything else available. Content that helps users take action, make a comparison, or access something only your site provides these are the queries where clicks still happen because AI can’t fully resolve them. Generic informational content is increasingly a zero-click investment.

The move: Before you greenlight any content piece based on keyword volume, ask one question: if AI produces a solid answer to this query in the search results, does the user still need to click anywhere? If the answer is no, reconsider the investment.

5. Stop Pursuing Exact-Match and Manipulative Link Building

Low-quality link volume is not an AI citation signal. This cannot be stated plainly enough.

LLMs weight the authority and editorial credibility of sources not raw link counts. A page cited by a hundred low-authority links is not more trustworthy to an AI model than a page cited by five links from publications with genuine editorial standards. The link-building tactics that moved rankings through sheer volume, private blog networks, exact-match anchor text manipulation, mass outreach designed to extract a link rather than earn it, have almost zero transfer value to AI visibility.

The publications that matter for AI citation are those with real editorial standards: trade publications in your industry, respected blogs with genuine audiences, academic-adjacent sources, and established media. These are not surfaces you can game with a link-building campaign. They require content genuinely worth covering, relationships built over time, and a brand that has earned the right to be mentioned.

The framing Tanguay offers is precise and worth using with your leadership team: a hundred low-quality links won’t get you cited in ChatGPT. Five links from publications your target audience actually reads might. The unit of measurement has shifted from quantity to source authority.

The move: Reorient your link-building brief around one question: which outlets does our target buyer actually read and trust? Build toward coverage in those outlets specifically. Stop measuring success in link volume.

6. Stop Micro-Optimizing CTR for Queries AI Is Answering

Title tag and meta description optimization for click-through rate assumes there’s a user choosing between blue links. For a growing and measurable share of queries, that choice is being made before the blue links ever appear.

The zero-click query is not a future trend. It’s the present reality for informational searches, and it’s expanding. Spending significant time A/B testing meta descriptions and title tag formulations for queries that AI Overviews consistently dominate is optimizing for an interaction that isn’t happening or happening far less than your keyword tool suggests.

This doesn’t mean CTR optimization is dead. For transactional and navigational queries, where the user needs to get somewhere specific, compare options, or complete a purchase, clicks still happen and blue link optimization still matters. Those are the queries more resistant to full AI resolution. Those are where CTR effort has real ROI.

For informational head terms where AI Overviews consistently appear, the strategic objective has changed. The goal is to become the cited source inside the AI answer, not the link below it. That requires a fundamentally different kind of content investment than CTR optimization does.

The move: Segment your keyword portfolio by AI Overview frequency. Pull the queries where AI Overviews appear consistently and pull them out of your CTR optimization queue. Redirect that attention to earning citations within those AI answers instead.

The Measurement Reframe Nobody Wants to Have

Here’s the conversation that’s coming whether you initiate it or not: traditional SEO metrics, impressions, clicks, organic traffic are going to look worse for some brands even as their pipeline performance holds or improves.

Google makes this explicit and it deserves amplification: you may lose volume in traditional SEO metrics while the metrics that matter most ; conversions, pipeline, revenue remain stable or grow. AI search increasingly rewards brands that get cited in answers over brands that rank in blue links. Citations drive qualified, intent-rich visitors who arrive having already narrowed their consideration set. Those visitors convert differently than someone who clicked a blue link on page one for a generic query.

The implication for how you report SEO performance is significant. If your CMO or CFO is watching organic traffic decline and asking questions, the answer isn’t to defend the old metrics. It’s to reframe the measurement model around the outcomes AI search actually produces and to build the tracking infrastructure to show those outcomes before the conversation becomes about budget cuts.

The brands that make this transition clearly and early will have the narrative advantage when organic traffic numbers shift. The ones that don’t will spend the next 18 months explaining why traffic fell without being able to show why it didn’t matter.

The Six-Point Summary

To make this immediately actionable, here’s the full framework in one place:

Do more:

Build brand entity signals across every surface AI models use to verify who you are
Develop topical depth through content clusters, not keyword-by-keyword content production
Earn unlinked brand mentions and genuine community presence in spaces AI models trust

Do less:

Chase high-volume informational keywords with thin content that AI Overviews will absorb
Pursue link volume through manipulative or low-authority tactics that don’t translate to AI citation authority
Micro-optimize CTR for queries where AI answers precede the blue link entirely

None of these six points requires a budget increase. They require a reallocation of where your existing SEO investment goes and a willingness to stop measuring success the way you did when the game was different.

The game is different now. The six-point problem is solvable. The only question is whether you solve it before or after your competitors do.

Source: Adam Tanguay, “6 SEO priorities to rethink for AI search,” Search Engine Land, July 6, 2026.,