After AI Steals the Clicks: What Still Makes SEO Work

AI Answers Change Clicks, But Don't Kill SEO

When generative AI delivers answers directly on the search results page and users never click through to any website, the "zero-click" trend understandably worries many practitioners: is SEO finished? The reality is more nuanced. Author Sunny Guha's take is measured — AI answers have changed click behavior, but they haven't erased the need for clear entities, credible pages, and genuinely useful in-depth content. What's actually shifting is the center of gravity: from chasing clicks to becoming the source that AI cites and users trust.

In other words, when your content is sufficiently complete and your entity markup sufficiently clear, even if AI summarizes the answer, you're still the authority being excerpted. That's a far more durable position than ranking first with thin content.

The "zero-click" phenomenon originated from Google's large-scale rollout of Featured Snippets and Knowledge Panels around 2016, allowing users to get direct answers on the search results page without clicking through. Generative AI (such as Google's AI Overview and Bing Copilot) has amplified this trend further — AI aggregates information from multiple sources and reorganizes it into a complete answer, causing organic click-through rates (CTR) to drop noticeably for certain query types. Research from firms like Semrush and SparkToro shows that more than half of mobile searches now generate no clicks. But this doesn't mean content has lost its value — AI needs to crawl and cite credible sources when generating answers. Pages with clear entity markup and authoritative in-depth content are actually better positioned to become AI's reference material and cited sources, gaining exposure and brand trust through a different channel.

Start from the Job to Be Done, Not the Tools

The first principle Guha emphasizes: write out the job to be done in a single sentence before listing the evidence you need before trusting automation. The order matters. Many teams' problems stem from deploying a stack of SEO tools and automation rules first, then working backward to find what those tools can solve — letting means drive ends.

If a task depends on a specific channel (like local SEO, which small businesses often neglect), make that dependency explicit in onboarding documentation and internal runbooks. The reason is practical: when time pressure hits, new colleagues tend to skip steps that aren't clearly documented. Making dependencies visible is the first line of defense against systems degrading under pressure.

Install a Minimum Viable Loop

Rather than pursuing comprehensive coverage from the start, Guha recommends getting one minimum viable loop running first. The core actions include:

• Strengthen entity consistency on key pages: Help search engines and AI accurately understand who you are and what you offer — this is a prerequisite for being cited.
• Write comparative content that answers real needs: Comparison content naturally fits users' decision-making context and makes it easier to shorten the conversion path on landing pages.
• Optimize internal links pointing to low-traffic commercial URLs: Pages that get little visibility but carry conversion responsibility are often overlooked by internal linking structures.
• Measure assisted conversions, not just last-click: In an environment where AI is capturing clicks, last-click attribution severely underestimates content's true contribution — assisted conversion is a much more accurate picture.

The value of this loop lies in being small, measurable, and able to quickly validate whether a given action actually improves decisions — rather than piling up a stack of "optimizations" with no clear attribution.

Assisted Conversion is a core concept in multi-touch attribution models, as opposed to Last-Click Attribution. The last-click model assigns all conversion credit to the final channel a user visited before converting, while assisted conversion measures the role a channel played in the middle stages of the conversion path. Both Google Analytics 4 (GA4) and major ad platforms support viewing assisted conversion data. In an era of widespread AI summaries, users may first learn about a brand through an AI answer, then search the brand name directly or visit the site to complete a purchase. In that scenario, the content page's contribution is "assisted" rather than "last-touch." If you only look at last-click, this value disappears entirely from reports, leading teams to incorrectly cut content assets that are still doing meaningful work.

Make Risk Visible Early

Guha spends considerable space on risk visibility — precisely the part most SEO content ignores. You need to decide in advance how to detect several types of problems: sync failures, stale recommended content, policy compliance issues, and authentication failures. When health signals exist, pipe them into the system; when the loop can't make a determination, keep a clear human escalation path.

The underlying logic here: treat SEO as a system you can explain clearly to finance and clients, not a layer of mysterious optimization magic. Systems that can be explained can be trusted, budgeted for, and reviewed.

One practice Guha returns to repeatedly is writing changelogs — logging automation rules and manual interventions. When results fluctuate, you need to be able to distinguish whether the cause is market noise, creative fatigue, or a setting someone changed last Tuesday. Without logs, every fluctuation becomes an unsolvable guessing game.

Changelogs are a foundational practice in software engineering, recording the time, content, operator, and reason for every system modification. Bringing this habit into SEO operations means that every change to robots.txt, every redirect rule adjustment, every Schema markup update, and every automation policy enabled or disabled should have a written record. The practical value is in after-the-fact attribution: when traffic shows an anomalous spike or drop at a particular point in time, logs let you quickly investigate whether a platform algorithm update, a technical change, or a content adjustment caused it — rather than having the team guess repeatedly in a fog of data. For SEO teams relying on multiple third-party tools (crawlers, backlink monitors, content publishing platforms), changelogs also enable cross-functional collaboration — keeping the actions of creative, technical, and data teams visible to each other, reducing "I thought you hadn't touched that" miscommunications.

Align Access, Packaging, and Next Steps

Beyond content itself, Guha also advises reviewing the principle of least-privilege access and incident response habits, checking the clarity of business boundaries, and continuously conducting empirical assessments with real case examples. These items might seem to belong to "operations" rather than "SEO," but they're precisely what keeps the entire system usable under real budget pressure.

A few operational habits worth keeping:

• Write decisions down: Teams that document decisions move faster later, because they don't need to rediscover the same arguments every month. A brief internal note beats a long undocumented debate that resurrects itself after every new hire.
• Test small before scaling: When something is uncertain, test it on a small portion of budget or traffic first. Gradual rollouts protect learning systems and protect cash when assumptions prove wrong.
• Document exceptions you've deliberately allowed: Undocumented exceptions quietly become default policy and create inconsistent customer experiences.
• Revisit regularly: Reexamine assumptions after major site launches, seasonal shifts, or account restructuring — what worked last quarter may now need lighter or heavier treatment.

The Enduring Pattern: Small Loops, Clear Ownership, Reversible Changes

The core pattern running through the entire framework is three things: small loops, clear ownership, reversible changes. Only when the minimum viable loop demonstrates that it can improve decisions without hiding risk should you expand the scope of automation. This is a pragmatic philosophy of growth — not getting swept up by promises of "full automation," but ensuring every expansion has verifiable justification.

Guha recommends a quarterly review of these practices, because platforms, policies, and your own product all evolve. Small improvements compound into clearer reporting and fewer emergencies. Sharing working notes with the people responsible for creative, measurement, and sales handoff keeps the whole system aligned and prevents silent contradictions between ads, pages, and follow-up.

Ultimately, SEO in the age of AI answers isn't a new set of secrets — it's about executing the fundamentals well: becoming the complete, credible, citable source, and operating it consistently through a system you can explain clearly.