FREE BUILD CHANCE FOR WINNIPEG BUSINESSES · CLAIM YOURS →
Togo Automation
All industry blogs

Your inbox logs who asked. Nobody replies tonight — in the language the parent actually speaks.

Parents research schools in the evenings and on weekends, and the first thoughtful reply usually starts the first real conversation. By the time you open your inbox the next morning, you're no longer the school that answered — you're one of the three tabs they still have open.

01 The pain

A parent finds you at 9pm, asks whether you have a class for their child, and waits.

The interest is real and specific — and it cools fast. Harvard Business Review's analysis of over two million sales inquiries found that companies responding within an hour were seven times more likely to have a meaningful conversation with the lead than those who waited longer. Choosing a school is a slow decision; parents will compare a few options no matter what. But the school that answers first, in their own language, is the one that frames that comparison.

And inquiries about children's classes almost never arrive during office hours.

02 Where your software stops

Most schools your size run on some mix of a class-management tool, a website form, an email inbox, and a group chat. None of those reply to anyone. The inquiry sits until a human sees it.

If you're on a bigger platform — say HubSpot, which markets its CRM heavily to schools — the honest picture in 2026 is different, and worth knowing. HubSpot's Breeze Customer Agent can auto-reply around the clock, across nine channels, and HubSpot says it works in the languages HubSpot supports. But look at what it takes and how it behaves:

  • It requires a Service Hub Professional or Enterprise subscription, plus usage-based fees on top ($0.50 per conversation it resolves).
  • Its answers are only as bilingual as the knowledge base you build and maintain — feed it English-only content and its Mandarin is guesswork.
  • It's designed to resolve conversations autonomously. It sends without asking you. For a bookstore's return policy, fine. For a message going to a parent about their child, that's a different risk.
  • And it doesn't make the small first judgment that actually books the assessment — which of your programs fits this particular child — because it doesn't know your programs the way you do.
The pattern

Your current tools either don't reply at all, or reply on their own terms — generic, English-first, and unsupervised. Neither one answers a Mandarin-speaking parent at 9pm the way your best front-desk person would.

03 Why not just any multilingual chatbot?

Multilingual chatbots are everywhere now — that part is genuinely commodity. What the off-the-shelf ones don't carry:

Your programs. A generic bot answers from generic training. The layer I build answers only from your actual class list, schedules, and policies — and if it can't find the answer, it says "not found" instead of guessing.

Your control. Rented chat widgets send whatever they generate. Here, nothing personalized goes out on its own: your front desk approves it first. Whoever taps approve owns that answer — and every answer has a traceable source.

Your context. I'm a native Mandarin speaker who's been the parent-side of these conversations. The difference between a translated reply and a reply that sounds like it came from someone's 阿姨 at the front desk — Chinese-speaking families feel that in the first two sentences.

Your ownership. This isn't a subscription you rent forever. The finished system is yours, the data stays in Canada (AWS Canada region), and I'm in Winnipeg — not a dashboard in another time zone.

04 The gap I fill

I build a reply layer that works in two speeds:

Instant, for facts. Schedule, tuition, age ranges, location, holidays — answered the moment a parent asks, in English or Mandarin, straight from your own documents. No approval needed because there's no judgment involved, and no guessing because "not found" is an allowed answer.

Approved, for judgment. When a parent describes their child and asks what fits, the system drafts a warm, personalized reply — suggesting the right program and a trial or assessment slot — and sends it to your phone. One tap to approve, and it goes out. The parent gets a thoughtful answer the same evening; you stay the one who said it.

Either path ends the same way: with a booked trial class or assessment on your calendar — because a reply without a next step is just a faster version of "we'll get back to you."

And it meets parents where they actually ask — website, email, Facebook or Instagram messages — not just a chat bubble no one opens. (If most of your inquiries come through WeChat groups, tell me; that's exactly the kind of thing we map in the first conversation.)

I also don't hand you a login and disappear. I stay through implementation until your front desk is actually using it — because a system your team ignores is worse than no system.

In one line Your tools log who asked. I build the part that answers tonight, in the language the parent actually speaks — instantly for facts, with your approval for anything that touches their child.