The Philosopher
Let me push back gently on the framing before answering it, because the premise contains a hidden assumption that changes the right answer.
First, interrogate the premise
"Most searches end without a click" is true in aggregate but hides massive variance by intent. Navigational queries ("Delta login") and many local queries have always had low third-party click-through — that's not new, and it's not really about AI. What's actually changed is the informational query layer, where AI Overviews and chat interfaces now answer things that used to require a click to a blog or explainer page. So the real question isn't "should I chase citations because clicks are dying everywhere" — it's "which of my content's jobs are being absorbed by AI answers, and which aren't?" That's a narrower and more answerable question than the global stat suggests.
Second, the dichotomy is softer than it looks
"Classic SEO" and "getting cited in AI answers" aren't two separate disciplines competing for the same dollars nearly as much as it seems. AI answer engines — Google's AI Overviews, Perplexity, Bing Copilot, even ChatGPT's browsing — still lean heavily on the same signals that made classic SEO work: crawlable, well-structured, authoritative content, clean markup, topical depth, demonstrated expertise. If you have strong E-E-A-T and technical hygiene, you're already doing 70% of what "AEO/GEO" requires. So the honest framing isn't "shift the budget" but "add a thin, specific layer on top of what you're already doing, and change where some of the effort points."
What's genuinely different about optimizing for citation
A few things really do differ and deserve deliberate investment:
- Quotability over ranking. Answer engines want a self-contained, extractable answer — a clear definition, a stat, a numbered list — not a page optimized to hold attention for a 1,500-word scroll. That's a content-craft shift, not a budget-size shift.
- Off-site presence matters more. LLMs synthesize from the aggregate web — Reddit threads, Wikipedia, review sites, forums, other people's coverage of you — not just your own domain. This argues for shifting some spend from on-site content production toward digital PR, community presence, and earned mentions, since being talked about elsewhere increasingly matters as much as your own page's SEO.
- It's more winner-take-most. Classic SEO has ten blue links to fight over; an AI answer often cites two or three sources total. The prize is more concentrated, so the ROI curve is steeper — bigger payoff if you win, more wasted effort if you don't.
Where a real shift is and isn't justified
This is the part that depends entirely on your business model, and I'd resist a one-size-fits-all answer:
- If you're an ad- or affiliate-monetized publisher, zero-click AI answers are close to existential — you're losing the click that was the business model, and being "cited" without a click doesn't pay your bills. Here the honest move might be diversifying revenue and distribution (newsletter, community, direct relationships) more than chasing citation, because citation alone won't replace lost ad impressions.
- If you're a B2B or expertise-driven business where being known as the authoritative source is itself valuable (leads still come from brand recall, sales conversations, later direct visits), citation exposure has real value even without a click — it's brand-building, closer to PR than to performance marketing.
- If your traffic is dominated by transactional or local queries, the zero-click trend affects you far less than the stat implies, and classic SEO remains the higher-leverage spend.
The measurement problem you should be honest about
There is no mature analytics layer for "got cited in an AI answer" comparable to rank tracking or GA4 click data. A few tools are emerging (Profound, Otterly, Ahrefs' Brand Radar, etc.) but they're early, inconsistent across models, and can't be A/B tested the way SEO can. That argues against a wholesale, confident budget reallocation right now — you'd be optimizing against a target you can barely see. Treat this as a hedge, not a conviction bet.
My actual recommendation
Don't reframe this as "SEO vs. citation" budget lines. Instead:
- Keep your core content-quality and technical SEO investment — it's the shared foundation for both worlds and is not optional.
- Reallocate a modest, specific slice (I'd start around 10–20% of content budget, adjust based on what you're seeing in referral/impression data) toward: quotable/structured formatting of existing high-value pages, FAQ and definitional content, and off-site authority building (PR, forums, Wikipedia-adjacent presence).
- Instrument what you can — track brand mentions in AI tools monthly even if imperfectly — and revisit allocation quarterly, because this landscape is genuinely moving faster than your planning cycle.
- Resist the temptation to treat this as a settled strategic pivot. It's an experiment with a thesis, not a proven channel yet, and the honest position is to hedge in proportion to how exposed your specific query mix is to AI absorption — not in proportion to a headline statistic about "most searches."