Money6 min read
Is SEO worth it for a small business when AI answers everything?
AI Overviews cut clicks on the exact queries small businesses used to win. That does not mean the traffic disappeared — it moved, and some of it still pays. Here is what we watched on one live account, and the test to run before you spend.
Yes, for most small businesses SEO is still worth the spend. The reason has shifted since five years ago: clicks on informational queries are getting rarer as AI Overviews answer them directly, while the buying moment on transactional queries still runs through a search page, and that's where the spend earns its keep.
The click didn't vanish, it moved to a different query
"Is AI killing SEO" is the wrong question, because it treats search as one thing. It isn't. When someone types "what is a fungus on grass," they used to click three articles, compare them, and pick a fungicide. Now they read the AI Overview once and buy. That query lost its click, and the page ranking first for it lost most of its traffic — the drop measured across AI Overview queries averages 58% off the top position's click-through ratemarket.
When someone types "seo agency dallas" or "book airport transfer da nang," the intent is different: they are close to a decision and want a page to act on. Those queries still send clicks, because an AI summary can't book the transfer for them. The businesses losing traffic are mostly the ones that built their whole SEO plan on the informational half of the funnel (buying guides, "what is X" explainers, top-of-funnel blog posts) and skipped the transactional half.
What we watched on one live account
We didn't want to answer this question from a report. We pulled the actual numbers from a client we run SEO for — four travel-services sites in Vietnam, no paid search budget in the last 90 daysmeasured.
| Signal | Last thirty days | What it proves |
|---|---|---|
| Organic clicks on the "message us" button | 813measured | People still act on a search result, not just read it |
| Door-coded contacts matched to a CRM record | 34, all 34 matchedmeasured | The click-to-conversation link is provable, not assumed |
| Of those, matched to a paid or booked trip | 9 of 34measured | Roughly one in four conversations becomes revenue |
| Revenue tied to those 9 contacts | $1,112, of which $565 already collectedmeasured | Small, real, and traceable line by line |
The tracking is eleven days old at the time of that snapshot: treat it as an early signal, not a monthly run rate. What it does settle is the doubt behind the question. A search click, with zero ad spend, still turns into a paid booking on this account, including one contact that arrived through a ChatGPT linkmeasured.
Three questions that actually decide it
Skip the traffic charts. A small business owner can answer this in five minutes.
- Does someone search before they hire or buy from you? If your customers find you through a referral, a storefront, or a sales call, no amount of AI disruption changes your answer: you were never depending on search in the first place.
- Does your site convert the visitors it already gets? SEO spend on a page that doesn't turn a visit into a call or a form is money spent moving people to a place that wastes them. Fix that first; it's cheaper than more traffic.
- Can you fund six months without a payoff? Search rankings, AI citations, and directory trust all build on the same slow signal: consistent, relevant content over time. Neither the old search game nor the AI-answer version pays out in week three.
Three yeses means SEO is worth it. Two means it might be, once the weak one is fixed. Zero or one means the money is better spent on ads this quarter, and SEO can wait.
What buying SEO looks like differently now
- Ask what happens to a page after it's read; ranking position alone tells you little now. A page written to be summarized by an AI Overview and a page written to close a $2,000 saleestimate are different documents, and an agency that writes one generic version for both is cutting a corner you'll pay for.
- Ask for the click-to-revenue join. A pile of clicks means nothing without knowing how many became a conversation and how many of those became money; see the table above for what that actually looks like when someone bothers to build it.
- Expect zero-click queries to keep losing clicks. Don't fight that trend with more blog posts aimed at "what is" questions; put the budget on the pages closest to a decision, where AI still can't finish the transaction for the visitor.
- Track AI-referred visits separately, even if the volume looks tiny today: the ChatGPT-linked booking above is a single contact so farmeasured. The businesses that show up first in AI answers tend to keep showing up, because the models favor what's already established.
The honest version of "is SEO worth it": worth it for the query where your customer is ready to act, shrinking fast for the query where they're just reading, and worth checking your own numbers against before anyone tells you which one you're in.
Sources
- Ahrefs, "AI Overviews reduce clicks" analysis, December 2025 — cited via Neil Patel's summary of the CTR-decline figure, retrieved 2026-09-29.
- Main Street ROI, "AI Is Changing Search: Is SEO Still Worth It for Small Businesses?" — background on how one agency frames the small-business decision, retrieved 2026-09-29.
- AEO Reach internal data — Rocket SQL snapshot of a client account's search-to-CRM attribution, 20.09.2026.