Owning your AI search visibility instead of renting it.

When someone asks ChatGPT, Gemini, or Google's AI Overview about your category, the answer they read is assembled by the engine from sources it picked for that moment. Your brand might sit in that answer today and disappear from it next week, and you will not get a notice either way. Pew Research found that when an AI summary appears on a Google results page, people click through to a website in only 8 percent of visits, down from 15 percent when no summary is present, and they click a link inside the summary itself in just 1 percent of visits. That is the position most brands occupy right now, showing up inside answers they do not control and cannot count on. Borrowing that visibility and building a presence these systems keep returning to are two different positions, and only the second one gives a business competing in Indonesian search something it can hold on to.

What renting your AI search visibility really means

For most brands, visibility in AI search today looks like an occasional appearance. Google's AI Overviews lift a line from one of your pages, an assistant paraphrases your product description, a generated summary names you as one of three options. It feels like a win because your name is on the screen. The problem is that you are there on the engine's terms, for as long as it finds you the handiest source for that exact phrasing, and it swaps sources constantly as models update and competitors publish fresh material.

The surface where this happens keeps expanding. Pew Research found that about 18 percent of Google searches already returned an AI summary in March 2025, with 58 percent of users seeing at least one that month. Gartner expects traditional search engine volume to fall 25 percent by 2026 as people move their questions into chatbots and assistants. So the space you are renting keeps expanding, while the visits that used to come with being named are drying up.

The rented model is more fragile than it looks

The first thing renting costs you is traffic. Ahrefs studied 300,000 keywords using Google Search Console data and found that AI Overviews cut the click-through rate for the top organic result by about 34.5 percent. You can hold the position you always fought for and still watch a third of the clicks disappear into an answer that quotes you without sending anyone over.

The second thing it costs you is control. When an assistant describes your product, it chooses the wording, the framing, and which competitor sits next to you in the comparison. If the model learned your brand from thin or outdated material, that is what gets repeated to a buyer who never reaches your site to be corrected. You have no slot to fix the description, no way to argue the sentiment, and no guarantee you are even included the next time the same question is asked with slightly different words.

The third cost is the one that compounds. Every appearance you rent builds equity for the platform rather than for you, because the relationship with the customer stays with the assistant. Being named there builds the customer's habit around the assistant, and none of that visibility carries into the next quarter as anything you own. For a channel this important, that is a strange arrangement to accept by default.

Ranking alone no longer earns the citation

Plenty of brands assume that if they rank well on Google, the AI citations will follow, and the data breaks that assumption cleanly. Ahrefs analysed 15,000 queries across ChatGPT, Gemini, Copilot, and Perplexity and found that only about 12 percent of the URLs cited by AI assistants also rank in Google's top 10 for the same query, while roughly 80 percent of cited pages do not rank anywhere in Google for the original prompt. The reason is that assistants do not pick a single winner the way a ranking does. They fan one question out into many sub-queries, pull candidate passages for each, and stitch an answer together from across all of them, so a page can earn a citation for a narrow angle it covers well even when it never ranks for the head term, and a page that ranks first can be skipped because it does not address the sub-question the model asked. Perplexity leans on Google's top results more than the others, at close to 29 percent overlap, so ranking still helps and remains the foundation. Treating a first-page position as the whole of your AI strategy, though, leaves most of the question space uncovered, which is why winning one keyword is a narrow bet in a system that reads across your entire footprint.

What owning your presence in AI answers looks like

Owning your AI search presence means becoming a source these systems have real reasons to trust and return to, across the full range of ways your audience asks about you. It rests on three things working at once. Your own content has to answer the questions your customers type, in enough depth and breadth that the model keeps finding you relevant. Independent sources your audience already believes have to describe you consistently, so the engine reads the same story about you from more than one direction. And you have to be part of the reviews, ratings, and communities where your category gets discussed, because recency and sentiment there feed back into what gets surfaced.

We package this as Total Graph Authority, and the framing that matters most is that it is audience-first. AI answers reflect real audience relevance, trust, and presence rather than any trick played on a crawler, so the work is to be the obvious answer for a real person, and the machine mirrors that judgment back to whoever asks next. A brand that is relevant, trusted, and present at the same time holds the one position that keeps working as the models change underneath it. If you want the mechanics of how those signals connect, we wrote them up in the graphs underneath Total Graph Authority.

Building content the models have to reference

The query fan-out finding has a direct consequence for how you build content. Chasing single keywords leaves most of the question space uncovered, so we build around whole topics instead. One pillar page carries the complete overview of a subject, and a set of cluster pages answers the narrower questions around it, covering what the thing is, why it matters, how it compares, what the benefits are, what reviews say, and how to buy. Together they cover the query space an assistant fans out into, which is the same surface those citations are drawn from.

This is the heart of our Topic Ownership Strategy, and the idea is to put authority in the structure rather than in a lucky phrase. When a topic is covered end to end and the pages link sensibly to each other, you stop depending on one URL ranking for one term and become the source the model reaches for whichever way the question is framed. As an illustration of what the approach tends to produce, we have seen it lift organic sessions by one and a half to three times and raise assisted conversions by 15 to 30 percent, though every business starts from a different base and those figures are not a forecast.

Take a lender writing about business loans against a BPKB vehicle document, a common query in Indonesia. The pillar page explains the product in full, and the clusters answer the real questions around it, how the process works, what documents are needed, how it compares to a bank loan, what the risks are, and how existing customers describe the experience. An assistant answering any one of those questions has a genuine source to pull from, and the brand shows up across the whole decision rather than in one lucky spot. The discipline that makes this work is restraint, because owning a handful of topics completely beats spreading thin across dozens, and depth is what the models reward.

Getting the entity and trust signals right

For an engine to cite you with confidence, it has to know exactly what you are, which is an entity problem, and structured data is how you answer it. Marking up your organisation, your people, your products, and the relationships between them with schema gives assistants an unambiguous description of your brand rather than something they have to infer from scattered mentions. Google has confirmed structured data as an input for its AI features, extending well beyond classic rich results, so it is one of the more dependable levers available to you.

In practice this means an Organization entry that ties your brand to the same identity everywhere it appears, with sameAs links to your verified profiles, Product markup on what you sell, and article markup on the content that carries your expertise. Done consistently, it stops the engine from confusing you with a similarly named business and gives it a stable node to attach every mention back to. This is the entity graph doing its work in the background, and it is often the fastest fix for a brand that gets described under the wrong name or the wrong details in AI answers.

Trust is the other half, and it is not something you can assert about yourself. When Gartner published its forecast on search volume, its own advice to companies was to produce unique content that is useful to customers and that demonstrates experience, expertise, authoritativeness, and trustworthiness. In real terms that means depth from people who do the work, and it means the sources your audience already believes describing you consistently, whether that is analyst coverage and G2 for a B2B service, Tripadvisor and OTA scores for a hotel, or marketplace ratings for a retailer. For most Indonesian brands that comes down to consistent, current listings and reviews on the platforms local buyers already trust, from Google Business Profile to the marketplaces and travel sites where your category lives. The engine reads that agreement across independent sources as a sign that you are a safe brand to name.

Turning AI mentions into an audience you own

Even a well-optimised presence in AI answers is still a presence on someone else's surface, so the strongest position also builds channels the assistant sits outside of. Direct relationships are what turn a borrowed mention into an owned asset, whether that is an email list, a community, a returning readership, or customers who search for you by name because they already know you. Similarweb reported that web visits across generative AI platforms grew 70 percent year over year to reach 9.5 billion between June 2025 and May 2026, so more of your future audience will meet you first inside an assistant, and what you do with that introduction decides whether the growth works for you or only for the platform.

Presence in the wider conversation feeds both sides of this. Reviews, ratings, testimonials, and mentions from creators and communities keep your brand active where your audience already spends time, and that activity is part of what assistants weigh when they decide who is worth surfacing. Being talked about recently and warmly is a form of ownership too, because it is hard for a competitor to rent their way past a brand that people are already discussing on their own.

How this comes together for an Indonesian business

Across the enterprise brands we work with in Indonesia, the pattern that holds up is the one described here, owning the query space with structured topic coverage, making the entity unambiguous, and earning trust and presence from the sources the audience already uses. One serviced-residence brand we work with grew its AI-surfaced audience from 7.4 million to 34.6 million in a single quarter by building that foundation instead of chasing individual placements. The work is slower than buying a burst of visibility, and it holds up far better every time the models shift.

A sensible first step is to find out how you are being described and cited today, which is what our free AI visibility audit is for, checking your brand and three competitors across the major assistants so you can see where you are rented and where you are owned. From there our Topic Ownership Strategy gives the content structure, and our published measurement methodology tracks the four things that move the number, brand mention frequency, citation share of voice, sentiment, and AI-attributed traffic. The brands that start building this now are the ones AI systems will keep returning to while everyone else is still hoping to be quoted.