People still search. They just increasingly read the answer instead of clicking it, or ask an assistant and never open a results page at all. The job has shifted from winning the click to being the source the answer is built from. This page is about that work, and about which parts of it are genuinely new.
SparkToro's analysis of Similarweb's US clickstream panel found that in the first four months of 2026, 68.01% of Google searches ended without a click. That is US data and India will differ, but the direction is not in dispute: the results page increasingly answers the question itself.
This is usually reported as the death of SEO. It is not. Being ranked and indexed is precisely what makes you eligible to be in the answer. What dies is the assumption that a ranking converts into a visit, and the reporting built on that assumption.
So the goal changes. Not traffic for its own sake, but being the business that gets named when someone asks a question in your category, and being findable by the smaller number of people who do click and are far closer to deciding.
There is a lot being sold right now as GEO, AEO or AI optimisation, usually as a separate and expensive service. Before you buy any of it, including from us, read what Google itself publishes:
"There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary."
"You don't need to create new machine readable files, AI text files, or markup to appear in these features. There's also no special schema.org structured data that you need to add."Google Search Central, AI features and your website. Google also states that the best practices for SEO remain relevant for AI features in Search.
That is Google describing Google. It covers AI Overviews and AI Mode, which is where most Indian buyers will meet AI answers first.
Outside Google it is not quite the whole story. ChatGPT, Perplexity and Claude run their own crawlers, and there is real work in making sure they can reach you, that your site states plainly what you are, and that you are referenced by sources they already trust. That work exists. It is just far smaller than the packages being sold for it, and it does not replace the fundamentals.
Our position, which costs us money to hold: if your site is not properly indexed and does not answer your buyers' questions, no amount of AI optimisation will help. That is the work, and it is the same work it always was, done for a search page that now talks back.
A crawl and an audit before anything is promised: what is indexed and what is not, which pages compete with each other, what the site says it is, and whether the technical foundation is sound. Most sites have two or three specific problems rather than a general one, and naming them is cheaper than a retainer.
Titles and descriptions that describe the page, headings that mean something, canonical tags, a sitemap that is accurate, structured data that matches what is visible, and pages that work on a phone. Unglamorous, frequently skipped, and the reason a lot of good content never ranks.
What you are, where you are, what you sell and who vouches for you, stated consistently on the site, in the structured data and across the profiles that carry your name. This is the part that matters most for assistants, because a model cannot recommend a business it cannot confidently identify.
Content built around what buyers actually ask, written so a direct answer can be lifted out of it, with figures sourced so they can be checked. Finite, not an endless calendar. The page you are reading is an example of the format, including the part where we quote a source that undercuts our own sales pitch.
Search Console for impressions, positions and queries. Analytics for assistant referrals. A fixed monthly set of questions asked of ChatGPT, Gemini and Perplexity to see whether you get named. The last one is a sample, not a metric, and we report it as such.
One number we deliberately do not quote: the widely repeated claim that AI referral traffic converts several times better than organic search. Published figures range from "considerably worse" to "23 times better" depending on who is publishing and what they sell. Until that settles, we would rather say we do not know.
If you want to understand the subject before you hire anyone for it, these two go into far more detail than a service page should, and neither is a pitch.
SEO, AEO and GEO: what is actually different separates the three terms and says which work matters.
Does SEO still matter if your customers are asking ChatGPT? covers what stopped working, what survived, and why your traffic report is about to mislead you.
If yours is not here, ask us directly.
Mostly not, and it is worth being straight about that. Google's own documentation says there are no additional requirements to appear in AI Overviews or AI Mode, no special schema and no AI text file needed, and that the best practices for SEO remain relevant. For assistants outside Google, such as ChatGPT and Perplexity, there is some genuinely separate work around crawler access and being cited by sources they trust. It is real, but it is a smaller job than the people selling GEO packages suggest.
No. Nobody can, and an agency that offers it is either guessing or planning to rank you for terms nobody searches. Rankings depend on competitors who are also working, and assistant answers vary between users and sessions. What we commit to is the work being done properly and reported honestly, including when it is not moving.
Technical and content fixes can change how you appear within weeks. Ranking for anything competitive takes three to six months, and for a new domain longer, because it depends on indexing, competition and being referenced elsewhere. One good month is weather. Three is a pattern. If nothing is moving by month three, you will hear that from us rather than a reassuring report.
Not with the precision of Search Console, and we will not pretend otherwise. What can be measured directly: impressions, clicks, positions and queries in Search Console, and referral traffic from assistants in analytics. What has to be checked manually: whether an assistant names you when asked a question in your category. We run a fixed set of those questions monthly and track the direction, which is honest about being a sample rather than a metric.
No, and the assumption that you do is why a lot of marketing budgets go nowhere. A brand needs enough content to answer the questions its buyers actually have, kept accurate. That is a finite job followed by maintenance, not an endless calendar. If the honest read is that you have enough content and a positioning problem, we would rather tell you.
Ask an assistant what your business does and see what it says back. If the answer is wrong, vague or empty, that is a fixable problem and a better starting point than a proposal. Tell us what it said.
Tell us what it said →This is one of five things we do. See the rest, or how engagements are priced and structured.