The Hardest Border to Cross Is No Longer Language

Most strategy shifts I’ve made as a founder, I can trace to a market report or a competitor doing something clever. This one I trace to a single question a client asked me last fall — and to the fact that I didn’t have a clean answer for her.

The CMO wanted to know why her procurement team kept arriving at vendor calls already half-decided. Not fully — half. And the half they’d decided didn’t reliably trace back to anything her marketing team had published. Her search rankings were healthy. Her content calendar was full. Something upstream of all of it had already happened to her buyers.

So we went and looked. Her buyers weren’t starting in Google. They were starting in ChatGPT and Perplexity, and the answers those tools handed back were assembled from sources her team didn’t own, and in several cases didn’t know existed. By the time a buyer reached her website, the website’s job had quietly changed — from forming an opinion to confirming one.

That conversation is the reason westOeast spent the last six months moving a real share of our work — call it a third — from SEO toward what the industry now calls GEO, generative engine optimization. I want to be honest in this piece about what that has produced and what it hasn’t. But first I want to explain why it landed harder for an agency like ours than it might for others.

Language used to be the border

For most of a decade, our work has lived on a specific seam: helping brands cross between East and West markets. Global companies landing in China. Chinese companies trying to earn trust abroad. Different buyers, mirror-image problem — getting a stranger in an unfamiliar market to take you seriously.

For years, the hardest part of that crossing was language, and language in the deep sense: not translation, but idiom, proof, the shape of a credible claim in a culture that isn’t yours. We got good at that. It’s most of what the agency was built on.

What changed is that a new gatekeeper showed up in front of the human one. Before a buyer in Chicago reads your carefully localized page, an AI engine has already decided whether you’re a real entity worth describing — and, if so, how. For a domestic brand, that machine usually has enough ambient context to get the basics right. For a brand operating across markets and scripts, it often doesn’t. We have watched AI engines confidently merge a client with a similarly named company, describe them in the wrong industry, or simply hedge — “this company appears in some discussions but is not well established” — which to a buyer reads as a quiet warning.

That is the part that reorganized our thinking. The border didn’t disappear. It moved earlier, and it stopped being about language. It became about whether a machine can identify you at all.

The decision, and what it cost

Deciding to shift a third of the agency’s work was not a clean strategic flourish. It was a slow, slightly uncomfortable few months.

The cost wasn’t tooling. It was instinct. Our team has a decade of SEO reflexes — write the content, earn the links, watch the ranking, count the click. Three of those four steps don’t carry over cleanly. Unlearning them, on live client work, while still hitting the targets clients signed up for, is harder than any software migration.

There was a client cost too. Not every client wanted this. A few were paying for a number to go up on a familiar dashboard, and the honest version of GEO reporting makes some numbers look worse before they look better. A couple of those relationships ended. We made our peace with that, but I won’t pretend it was comfortable.

And we got things wrong. The one I’d flag: we assumed that the authority sources behind a brand — the third-party profiles and references an AI engine leans on — could be strengthened once and would carry across markets. They don’t. What an engine treats as a credible source for a brand in its home market often does nothing for the same brand in a different language market. We rebuilt that part of our approach after the first quarter, and I’d expect to rebuild it again.

Six months in: what’s proven, what hasn’t

I’m wary of agencies that narrate a transition as a clean success, so here is the unglamorous version.

On a couple of engagements, we can now see a brand turning up in AI answers where it previously didn’t — described accurately, occasionally recommended. That shift took months, not weeks, and even now I’d call the trend directional rather than nailed down. On other engagements, nothing has moved yet, and we don’t fully know why. One pattern we suspect but haven’t proven: the engines update their picture of a brand on their own schedule, not ours, and a quiet stretch can simply mean the next refresh hasn’t happened.

What I’m fairly confident of after six months is narrower, and it’s this: AI search visibility is an operating problem, not a tooling one. Buying a monitoring dashboard tells you the gap exists. It does not close the gap. Closing it is slow, specific work — correcting how a brand is described in the places engines actually draw from — and someone has to do that work by hand. There is no version of this you purchase and switch on.

What I’d tell a CMO

If you’re a CMO weighing where this sits on your roadmap, three things, briefly.

First, before you measure visibility, check identity. Ask the major engines a plain question — who is this company, what do they do — and see whether the answer is even correct. If it isn’t, no amount of content investment lands, because the engine doesn’t yet know who it’s describing. For cross-border brands this is the most common blocker, and almost nobody checks it first.

Second, don’t run this as a side project. It needs its own workstream and its own success measures. Folded into the SEO team’s existing targets, it loses every prioritization argument, because its payoff arrives one or two quarters later than a ranking does.

Third, plan in quarters. The engines move slowly and so does this work. If someone promises you a fast result, that is the part I’d be skeptical of.

Closing

We didn’t make this shift because of a headline about AI replacing search. We made it because a client asked a question we couldn’t answer, and the honest follow-up took us somewhere we hadn’t planned to go.

The award-winning campaign, the clean dashboard, the decade of SEO instinct — those were the easy part. The rebuild is harder, and we’re still in it. I’d rather write about it while it’s unfinished than narrate it cleanly once it’s safe to.

If you’re working through the same question for a brand that crosses markets, I’m glad to compare notes.

Suya Wang

Suya Wang is the founding partner of westOeast, a B Corp certified marketing agency working at the intersection of SEO and Generative Engine Optimization for cross-cultural B2B and SaaS brands.