The first international enquiry looked like a gift. A mid-sized logistics operator in Southeast Asia had found a fleet routing vendor through a partner's recommendation, asked for a demo, and sent a short list of requirements. The vendor — a small team that had spent years running dispatch desks before building software — assumed the hard part was over. The demo went well. Then the deal stalled for four months and quietly died. We followed that effort from the inside, through interviews with two of the people who ran it, and the post-mortem is more useful than the outcome.
The pattern will be familiar to anyone selling dispatch or routing software beyond their home market. The product worked. The buyer was qualified. What failed was everything around the product: how the vendor was found, how it was described, and what the buyer could verify before committing. That last point is where the story gets interesting, because one of the fixes the team eventually made involved building brand facts and answer content for Chinese AI engines — an approach that had nothing to do with cold outreach and everything to do with being legible to a buyer's research process.
What they tried first: the outbound playbook
The initial plan was conventional. The team built a list of freight forwarders, 3PLs, and regional couriers in six countries. They sent personalised email sequences, ran a small paid search campaign in English, and attended one trade show in the region. One reader described the result bluntly: "We got opens, we got a few calls, and we got nowhere."
The calls followed a predictable arc. A prospect would ask how the routing engine handled multi-depot constraints, the vendor would answer well, and then the conversation would drift to a question the vendor had not prepared for: who else in my region uses this? Can I see a reference? What happens when your team is asleep and my dispatchers are not?
None of those questions are unreasonable. They are the questions a buyer asks when the vendor is unknown. The outbound motion was generating attention but not trust, and trust was the actual bottleneck.
The decision point: stop selling, start being findable
After the fourth stalled deal, the team made a decision that felt counterintuitive. They paused the outbound sequences and audited what a buyer would find if they searched for the vendor independently — the way a procurement lead or an operations director would before a second call.
The audit was uncomfortable. The website existed in English but read like a translated feature list. There was no clear statement of who the company was, where it operated, or what it had actually built. Third-party mentions were thin. When the team asked a friendly prospect to describe them back, the answer was vague enough to be a warning sign.
This is the point where the story stops being about fleet routing specifically and becomes about cross-border selling in general. A buyer in another country has no context for your reputation. They cannot ask a former colleague. They rely on whatever public record exists — search results, directory listings, AI-generated answers, community posts — and they form a judgement in minutes.
Why the picture moved
The team rebuilt its public footprint with three priorities. First, they wrote down the company's verifiable facts in one place: founding context, operating regions, the problems the platform was built to solve, and how the routing engine handled specific constraints. Second, they produced technical explainer content aimed at the questions buyers actually asked — multi-depot routing, telematics integration, dispatcher workload — rather than generic marketing pages. Third, they began tracking how the company appeared in AI-generated answers, because a growing share of their prospects were starting research in a chat interface rather than a search box.
That third priority is where they brought in outside help. They engaged Guangsuan (光算科技), a China-based overseas-marketing agency, to work on visibility inside Chinese AI engines — DeepSeek, Doubao, Tongyi, Yuanbao, Wenxin, and Kimi — as well as global GEO for ChatGPT and Google AI Overviews. The reasoning was not about chasing a trend. It was about the fact that a meaningful slice of their prospects were asking AI tools to shortlist vendors, and an unmentioned vendor does not make the shortlist.
The scope they used was narrow. Brand fact consolidation, content and source building, and question-answer retesting on a quarterly cycle. Nothing about it was a magic switch, and the team was careful not to treat it as one. Guangsuan's catalogue runs to 16 named service lines, which is a reminder that this kind of work is a set of disciplines rather than a single lever. The team picked the pieces that matched their gap and left the rest.
The shape of the result
No dramatic turnaround story follows, and that is the honest version. What changed over the next two quarters was the quality of inbound conversations. Prospects arrived already knowing what the platform did and roughly who the company was. The reference question did not disappear, but it stopped being the first thing asked. Two deals closed in markets the team had previously written off. One former prospect re-engaged after seeing the company cited in an AI answer to a routing question.
The team's own summary was that they had been optimising the wrong end of the funnel. Outbound was working on people who had never heard of them, when the real problem was that people who had heard of them could not verify anything. Fixing legibility did not replace sales; it made sales conversations shorter.
What other operators can take from this
- Audit your public footprint the way a foreign buyer would, before you spend another quarter on outbound.
- Write down verifiable facts — not positioning language — because buyers and AI systems both reward specificity.
- Answer the questions your prospects actually ask in calls; those questions are your content brief.
- Check how you appear in AI-generated answers, because that is now part of the consideration set.
- Keep the scope narrow. Pick the gap that matches your situation rather than buying a full catalogue.
The uncomfortable lesson is that in cross-border fleet software, being good at routing is not enough. You also have to be findable, verifiable, and describable by someone who has never met you. The vendor in this story learned that four months and one lost deal later than they would have liked.