AI SEO vs Traditional SEO: Which to Fund First

Two sales pitches are landing in Malaysian business owners’ inboxes this year, and they contradict each other. The first says your existing SEO is obsolete and you need to buy AI SEO before your competitors do. The second says AI SEO is a repackaging exercise and you should ignore it. Both are positions taken by people who want your money to move in a particular direction.

There is a duller way to settle it, which is to read what the search engine itself published. Google updated its official guidance on optimising for generative AI search in July 2026, and that document answers the funding question more directly than most agency proposals do. It is worth reading before you sign anything.

This article settles AI SEO vs traditional SEO the practical way. It works through what Google’s guidance actually says, where AI answers pull their sources from, and what an AI answer costs you in clicks. It ends with a twelve-month budget split for three stages of business. If you want the full scope of what this category covers, our AI SEO, GEO and AEO page sets it out.

One disclosure first: we sell both. The recommendation below tells most readers to buy less of the newer, more expensive thing for now. It is still the right advice.

PART 1 · DIAGNOSE

What Are You Actually Buying When You Buy AI SEO?

IN BRIEFMost AI SEO scopes contain three different things at once: normal SEO relabelled, tactics Google says it ignores, and a small amount of genuinely new work. Separating them settles AI SEO vs traditional SEO before you compare a single price. Our guide to answer engine optimisation covers the underlying mechanics.

Google’s guide to optimising for generative AI search takes the terminology head-on. It accepts that AEO and GEO describe work aimed at AI search experiences, then adds that from Google’s side, optimising for generative AI search is optimising for the search experience — and is therefore still SEO. The AI features sit on top of the core ranking systems, not beside them.

That does not make every AI SEO proposal dishonest. It does mean any scope you receive should be sortable into three buckets before you price it:

• Work you were already buying — crawlability, indexing, page speed, internal links, useful content. Google’s AI features documentation states plainly that there are no additional requirements to appear in AI Overviews or AI Mode.
• Work Google explicitly says it ignores — llms.txt files, chunking content into fragments, rewriting copy in an “AI-friendly” voice, chasing manufactured brand mentions.
• Work that is genuinely new — visibility inside assistants Google does not operate, and preparing a site for browser agents that read pages on a customer’s behalf.

The third bucket is real and it is small. The commercial problem is that the first bucket is usually the largest, so an AI SEO retainer often costs more than an SEO retainer while delivering much the same list of tasks.

BENCHMARK BRIEFING 1 OF 4

Which AI SEO Line Items Are Genuinely New Work?

IN BRIEFMapped against Google’s two published AI documents, most items on a typical AI SEO scope resolve to traditional SEO, and several resolve to nothing at all. Only two categories are new spending. Use this table as a checklist when you read a search marketing proposal.

The table below takes the line items that appear most often in AI SEO scopes and places each one against what Google has actually published about it.

PART 2 · DIAGNOSE

Does AI Visibility Depend on Rankings You Do Not Have Yet?

IN BRIEFLargely, yes. AI answers are assembled from pages the search index already rates highly, so a site with no rankings has little for an AI system to pull from. That makes ranking work a prerequisite, not an alternative — the same logic behind earning AI Overview citations.

Google describes the mechanism openly. Its AI features use retrieval-augmented generation, which relies on the core Search ranking systems to retrieve relevant pages from the index, then generates an answer from what those pages say. The AI layer does not go looking for good pages on its own. It reads what the index hands it.

This changes the sequencing question completely. If your pages are not being retrieved for anything, buying AI-specific work is like paying for shelf placement in a shop that does not stock your product. The constraint sits one step earlier.

BENCHMARK BRIEFING 2 OF 4

Where Do AI Answers Actually Get Their Sources?

IN BRIEFA study of 1.9 million citations found roughly three-quarters of AI Overview sources already rank in the top ten for the query. Around one in seven come from outside the top 100. That ratio is the strongest argument for funding rankings first, ahead of any dedicated GEO engagement.

The bars below show where cited pages sat in ordinary search results at the time they were cited.

The 14.4% that rank nowhere is the honest counterpoint, and it is why AI-specific work is not pure theatre. But it is the exception, and chasing an exception is a poor use of a first budget.

PART 3 · DESIGN

So Which One Should Get the Money First?

IN BRIEFAI SEO vs traditional SEO comes down to three possible sequences, and only one suits most Malaysian businesses reading this. The deciding question is whether buyers already find you in ordinary search results, which is also the first thing any competent SEO partner should check.

BENCHMARK BRIEFING 3 OF 4

What Does an AI Answer Cost You in Clicks?

IN BRIEFA 300,000-keyword study measured position-one click-through rates before and after AI Overviews rolled out. The gap works out to roughly a third of the clicks a top ranking used to earn — which changes what you should count as a return on SEO, not whether to buy it.

The figures below track the same measure a year apart, for keywords that later triggered an AI Overview and for comparable keywords that did not.

Note what this does and does not say. Ranking first still earns more clicks than ranking fifth. What has changed is the size of the prize, and Google reports that clicks arriving from AI-assisted results tend to be higher quality — visitors stay longer. Fewer, warmer visits is a different business than more, colder ones.

PART 4 · DEPLOY

What Actually Changes in the Work Itself?

IN BRIEFFour things genuinely change, and none of them replace the standard technical fixes. They are additions at the margin, best funded once the base is producing.

1. Report on assistants separately. Search Console’s generative AI report covers Google. Visibility inside ChatGPT or Perplexity needs its own check, even if that check is manual at first.
2. Raise the originality bar on content. Google’s guidance singles out commodity content — the generic listicle — as the thing least likely to be picked up. First-hand specifics travel; summaries do not.
3. Watch what third parties say about you. Assistants read forums, reviews and directories. Earning genuine coverage matters more than it did; buying fake coverage still does not work.
4. Check the site renders for agents. Browser agents read the DOM and the accessibility tree. Clean, semantic markup helps them and helps screen-reader users at the same time.

Each of these is measured in hours per month, not thousands of ringgit. That is the practical test of whether an AI SEO premium is justified: ask what the extra fee buys beyond this list.

BENCHMARK BRIEFING 4 OF 4

How Should the Split Look Over Twelve Months?

IN BRIEFThe split should move with your ranking base, not with the calendar. This illustrative ladder shows how the AI-specific share might grow across three stages of a Malaysian business, from a standing start to an established position.

THE VERDICT

The Order That Wins for Most Malaysian Businesses

IN BRIEFFund traditional SEO to working strength, add AI-specific work as the rankings arrive, and review the split every quarter against one number. Our AI search visibility service is built around that order for exactly this reason.

AI SEO vs traditional SEO invites you to pick a side, and the honest answer refuses the framing. For Google, there is one search experience and one set of foundations under it. Outside Google, there is a genuinely separate and still-small opportunity in the assistants people are starting to ask instead.

So the practical order is: get retrievable, get ranked, then get cited. Skipping to the third step is the mistake, and it is an expensive one because the work is priced as a premium.

Set one review trigger before you sign anything. If your top-ten positions have not grown at all by month six, the problem is the traditional work, not the absence of AI work — and no amount of AI-specific spend will cover for it.

FAQ

Frequently Asked Questions

1. Is AI SEO just traditional SEO with a new name?

Mostly, but not entirely. It depends on which search surface you mean: for Google, the company’s own documentation says optimising for its AI features is still SEO, so the overlap is nearly total. For ChatGPT and Perplexity, which Google does not operate, there is separate work worth doing once your base is solid.

2. Will my rankings still matter if people stop clicking?

Yes, because rankings now do two jobs instead of one. What changes is the payoff: the same position earns fewer clicks than it did, while also feeding the AI answers that shape whether anyone considers you. Judge the position by both, not by traffic alone.

3. How much should a Malaysian SME spend on AI-specific work?

Very little in the first year, and nothing at all before you rank. The right figure depends on how many top-ten positions you already hold — under five, the answer is zero; above ten, a tenth to a quarter of the search budget is reasonable while you learn what it returns.

4. Do I need an llms.txt file or special AI markup?

Not for Google. Its guidance states that Search ignores llms.txt and similar files, so they neither help nor harm your visibility there. If a proposal lists creating one as a deliverable, treat that as a signal about the proposal rather than about your website.

5. How do I check whether AI answers are affecting my traffic?

Compare impressions against clicks in Search Console over twelve months. If impressions hold steady or rise while clicks fall, AI answers are the likely cause rather than a ranking loss — a distinction worth confirming before anyone rebuilds your content on the strength of a traffic dip.