The central development is a tension: Google says AI Mode is creating longer, more involved search behavior and that query growth reached its fastest pace in five years, while search-ad revenue growth slowed from 15% to 12% year over year and YouTube slowed from 15% to 9%. Google also said that fast AI Mode query growth was accompanied by lower monetization at that stage.

For Taiwanese businesses, the issue is not simply whether a new ad placement will eventually be visible. When consumers split “what should I buy?” into use case, comparison, risk and confirmation questions, a brand may appear earlier in the decision process without producing an immediate click or order. Google announced expansion and format changes; that does not mean every market or account already has access, nor that existing search metrics will automatically carry over.

Millennium’s editorial judgment is that Taiwanese brands should not treat AI Mode as an isolated new media placement. They should first build a “question–evidence–next step” content structure. The reason is straightforward: when search moves from a single keyword to a sequence of questions, competition involves more than bids. It also involves whether a brand can provide clear, verifiable and situation-relevant information at each stage. This is Millennium’s independent analysis, not a promise of results from Google in Taiwan.

For a hypothetical Taiwanese home-appliance brand targeting renters, simply increasing bids on “small-space appliances” would be insufficient. It could first organize likely follow-up questions about noise, dimensions, cleaning and after-sales support, with each answer leading back to the same product decision. This is a hypothetical example, not a reported client result; its value is making the path from exposure to decision observable rather than relying on one keyword hit.

Millennium's view：Millennium’s editorial view: the real change in AI search is not simply another ad placement. Brands must treat search as a continuous decision journey, prove that content can handle follow-up questions, and only then assess whether to increase media investment. Google reported growing AI Mode queries and longer interactions alongside lower monetization and slower overall search-ad growth. That shows that more activity is not synonymous with more commercial value. If businesses chase a new format alone, they may not know whether the problem lies in exposure, answer quality, the landing experience or conversion. Staged content and measurement make that uncertainty separable.

The first question is whether content covers genuine follow-up concerns rather than merely accumulating high-volume terms. The second is whether media data can separate awareness, consideration and conversion instead of treating every interaction without an immediate click as waste. The third is whether the brand can keep product, price and service claims consistent as queries become longer. Because the source describes Google’s direction and revenue gap at a broad level, Taiwanese companies cannot infer that any particular industry must rise or fall.

A sensible approach is a small, stoppable test: select one product and one high-intent situation, build follow-up content, record exposure, interaction, visit quality and later actions by stage, then compare them with an existing search campaign. If the new format is not yet available where the company operates, the same measurement can first be simulated through existing content and search journeys. That builds readiness; it does not assume access to AI Mode ads.

Suggested action 1：List five to eight genuine follow-up questions for one core product and create verifiable specifications, use cases and next-step content for each; start with existing search content without claiming access to AI Mode ads. Within four weeks, complete the question-to-content map and use a manual review to confirm that every question has consistent product evidence and a next-step link.

Suggested action 2：Split one search activity into awareness, consideration and conversion stages, recording exposure, interaction, visit quality and later actions separately instead of judging multiturn value only by click-through rate. During the single-product test, produce a weekly stage-based record and identify at least one traceable path from exposure to a later action.

Suggested action 3：Run a stoppable comparison for one high-intent situation: maintain existing search content on one side and add follow-up information and consistent product claims on the other, then inspect quality differences by stage. After the test, compare qualified visits, product-information interactions and subsequent inquiry rates between the two groups, recording the sample period, differences and whether to extend it without presuming revenue growth.