How we accidentally turned one CS agent into four (without hiring anyone)

Over the last year, we kept seeing the same pattern with companies expanding into Southeast Asia. They'd tell us they needed to hire more customer support. But after digging a little deeper, the problem usually wasn't ticket volume.

It was fragmentation.

• One customer messages on WhatsApp.

• Another uses LINE.

• Vietnamese customers only want to use Zalo.

• Website visitors use live chat.

• Every market speaks a different language.

So instead of one experienced support person helping everyone, companies end up hiring separate agents because of tooling limitations rather than actual workload. That got us wondering:

❓ What if the limitation wasn't the agent? What if it was the inbox?

So we started building around that idea.

1️⃣ The first challenge: Language

Most translation tools only solve half the problem. They'll translate the customer's message so the agent can read it. But when the agent replies, they're back to copying text into another translator. That completely breaks the conversation.

We wanted the translation layer to disappear entirely.

• If a customer writes in Vietnamese, the agent simply reads English.

• If the agent replies in English, the customer receives Vietnamese.

Neither side has to think about translation at all.

2️⃣ The second challenge: Messaging platforms

People outside Southeast Asia often underestimate how fragmented messaging really is.

• Vietnam is dominated by Zalo.

• Japan and Thailand rely heavily on LINE.

• Indonesia and Malaysia primarily use WhatsApp.

• Many international communities prefer Telegram.

You can't realistically tell customers:

"Please contact us on our preferred platform."

They'll simply contact someone else. So instead of treating every messaging app as a separate support queue, we merged everything into a single inbox. The interesting result wasn't convenience. It was utilization. The same experienced support person who previously handled one language and one channel could suddenly support customers across multiple countries without changing applications or constantly switching mental context.

3️⃣ The AI part came later

Initially, we assumed companies would spend weeks building knowledge bases. Most didn't. Especially startups. They just wanted to start answering customers. So we tried something different. Instead of requiring documentation before AI becomes useful, we let it learn from real conversations as agents work. Companies can still connect structured knowledge later if they want, but they don't have to spend weeks preparing content before seeing value.

4️⃣ One unexpected challenge

One thing that surprised us during development was how difficult some integrations were. Zalo, for example, turned out to be far more challenging than WhatsApp or Telegram. Documentation was limited. Examples were scarce. There weren't many existing platforms to learn from. It ended up taking far longer than we originally estimated.

5️⃣ The biggest takeaway

The biggest lesson wasn't actually about AI. It was realizing how many businesses were solving a software problem with hiring.

Sometimes you genuinely need more support staff. But sometimes your best support agent is only operating at 25% of their potential because they're trapped inside disconnected systems. I'm curious whether other founders have run into the same issue.

If you operate across multiple countries, how are you handling:

• Multiple messaging platforms?

• Multiple languages?

• AI-assisted support without spending months building a knowledge base?

I'd genuinely love to hear what's worked (or hasn't worked) for you.