Is SEO dead, or still worth it in the AI era?

The parts that determine whether machines can reach and read you at all, because every AI engine inherits them:

• Crawlability. AI crawlers arrive through the same door search crawlers always used. A site that blocks or confuses them is invisible to answers and rankings alike.
• Clean structure. Headings, semantic HTML, and structured data are how any machine, ranking or generating, parses what your pages actually say.
• Server-rendered content. Most AI crawlers don't execute JavaScript, so the SEO-era discipline of putting real content in the initial HTML matters more now, not less.
• Site health. Speed, working links, and a coherent architecture still gate everything above them.
• Local fundamentals, for location-bound businesses: accurate listings and consistent identity data feed both map results and AI answers.

Call this layer the plumbing, and notice its best property: it can be finished. Unlike the rankings arms race, technical foundation is work you complete and maintain, not a monthly war. If your SEO history bought you clean plumbing, that investment didn't die; it's the part that transferred.

The parts whose entire payoff depended on the click, because the click is what left. Three investments age worst:

1. Rankings as the scoreboard. Position three means little when the page above every position answers the question. SparkToro's analysis found fewer than one in three Google searches now sends a click to any website, and Search Engine Land's zero-click research confirms the majority of journeys ending on the results page itself.
2. Keyword-volume content. The blog written to blanket keyword variations earned traffic that no longer arrives, and generative engines actively prefer evidence-dense pages over keyword-shaped ones. Volume without substance now loses twice.
3. The traffic-first funnel math. Strategies premised on impressions converting to visits converting to leads inherit the collapse at step two: Pew measured clicks on traditional results at 8% of visits when an AI summary appears, with the summary's own source links clicked just 1% of the time.

The pattern across all three: SEO's value used to be delivered on your website, after the click. Increasingly the value is delivered inside the answer, before any click, which is a different game with different winning moves.

Audit what the retainer actually buys, because the label 'SEO' now covers both the work that compounds and the work that expired. The judgment-level split:

Worth paying for:

• Technical foundation work: crawlability, structure, speed, schema, and fixing what blocks machine reading.
• Content built around real buyer questions with genuine answers, the material AI engines cite.
• Measurement that includes AI visibility: what engines say and cite, not just where you rank.

Worth interrogating:

• Monthly deliverables denominated purely in rankings and keyword positions.
• Content volume aimed at query variations rather than questions a buyer would recognize.
• Reports celebrating impressions while traffic, and inquiries, stay flat, which is the zero-click era showing up in your invoices.

The conversation that settles it takes one meeting: ask your provider how their work makes you citable by AI engines, and what they measure beyond position. Providers who have adapted answer specifically and are worth keeping. Providers who respond that AI is a fad haven't adapted, and their deliverables show it.

As one workstream with two scoreboards, not two rival programs. The relationship is layered, and getting the layers straight saves both budgets:

1. The plumbing layer is shared. Crawlable, structured, healthy sites feed ranking algorithms and generative engines alike. One investment, both scoreboards.
2. The content layer diverges in aim, not in kind. Ranking-era content chased keywords; citation-era content answers real questions with quotable evidence. The second happens to perform respectably at the first anyway, because search engines also reward substance.
3. The authority layer converged. Mentions, reviews, and third-party confirmation now drive both traditional trust signals and AI citation behavior, with off-site mentions tracking AI visibility more closely than backlinks.

The practical upshot for an owner: you don't need an SEO program plus a GEO program (what GEO is, in full). You need one honest visibility program, foundation, real answers, third-party evidence, freshness, measured against both scoreboards, with the weight shifting toward the citation side as your buyers' behavior does. Businesses running two separate programs are usually paying twice for the shared layers and optimizing the two scoreboards against each other.

Test it against the scoreboard that pays now, in an afternoon:

1. Run your money questions. The five queries that historically brought clients. Check your classic rankings, then ask the same questions to ChatGPT, Perplexity, and Google's AI answer, and note who gets named and cited.
2. Grade the divergence. Ranking well while absent from answers means your SEO bought plumbing and position but not citability, common, and fixable, since the foundation is the hard part. Absent from both means a deeper problem than strategy drift.
3. Check what the engines can extract. Open your top pages and ask whether a machine could lift a direct answer from the opening lines, or whether the content circles its keyword the old way.
4. Follow the leads, not the traffic. Zero-click behavior means inquiries can hold steady while visits fall; conversely, ranking-era traffic can flow while the buyers who matter never see you in answers.

That audit, run systematically across engines and queries, with gaps ranked by revenue impact, is exactly what our free AI Visibility Scan does. Your SEO history is not wasted either way; the audit tells you which parts to build on.