← Blog
Bots & Scripts
September 22, 2026 · 4 min read

WhatsApp and Telegram bots: when a bot beats an app

If your customers already live in a messaging app, going to them is easier than calling them somewhere else.

Bots & Scripts

Getting someone to install a new app is hard. They have to open the store, download, sign up and allow notifications. A bot starts with a single message inside an app they already open every day. Still, not every job fits in a chat window. Here is where a bot is enough, where it isn't, and how to decide.

Where bots are strong

  • Short, repeated tasks: booking an appointment, asking for order status, sending a document.
  • Notification-heavy work: shipping updates, payment reminders, stock alerts.
  • Communities: announcements, polls, tasks and reward mechanics.

These jobs share one thing: the user knows what they want and gets there in a few steps. There is nothing to download, learn or remember. Your notification also lands in a chat list they already check.

Which platform suits which job?

  • WhatsApp: suits customer service, orders and bookings, because most of your customers are already there. The business account must be approved; conversations you start use pre-approved message templates, and the platform charges for some of them.
  • Telegram: suits communities and mini apps. Setting up a bot is easy, and buttons, menus and in-chat web screens give you plenty of room.
  • Discord: suits gaming and interest-based communities. On a server, bots assign roles, enforce rules and manage events and announcements.

Choose the platform by where your customers are today, not by technology. The same logic can run on several platforms at once, but starting with one helps you learn faster.

Build and maintenance costs

A bot usually launches sooner and on a smaller budget than an app. There are few screens to design, no separate iOS and Android versions, and no store review to wait for. In our work a bot is mostly measured in weeks, a mobile app in months.

Maintenance differs too, but a bot isn't maintenance-free. It must run around the clock on a server, alert you when something breaks and keep logs. When platform rules change, or your products and prices do, the conversation flows need revisiting. An app adds store updates and yearly operating-system changes.

Where you need an app

Work that needs a rich interface, such as browsing a map, long forms or multi-step editing screens, doesn't fit in a chat window. If people will spend twenty minutes a day in your product, give them a space of their own.

Chat has other limits too. Users see one thing at a time; comparing options side by side or scanning a long list is tiring. You are also on someone else's platform: its rules decide what you can send, when and in what format. Working offline, or heavy use of the camera and location, also calls for an app.

The option in between

Telegram mini apps sit between the two: they open from inside the chat but offer a full web interface. We built the Schmeckle Token game this way, so players never leave the conversation.

How to design a good bot conversation

  • Keep steps short: one question per message.
  • Offer buttons instead of free text: options such as “Book an appointment”, “Order status” and “Talk to an agent” remove typos and misunderstandings.
  • Say what the bot can do in the first message, so users needn't find its limits by trial and error.
  • Summarise and confirm before completing anything: date, time, amount.
  • Leave a way to reach a person at every step.

The hand-off moment

When the bot hands a conversation it can't answer to an agent, it should include a short summary of what has been said. If customers must explain their problem twice, the time the bot saved is lost.

What to avoid matters as much as what to do. The mistakes we see most often:

  • Trying to move the whole website into the bot. A good bot does a few jobs very well.
  • Passing the bot off as a person. Users should know they're talking to a bot.
  • Sending notifications too often. Once a user blocks the bot, you lose that channel entirely.
  • Not connecting the bot to existing systems. If orders or bookings are retyped into another screen, the work hasn't shrunk; it has only moved.
  • Keeping no logs. You can't improve a flow without seeing where conversations break off.

A short checklist for the decision

  • Which channel do your customers write to you on today?
  • Does the job finish in a few short steps, or need long work at a screen?
  • Will people use the product a few minutes a week, or at length every day?
  • Is the real value in the notification or the interface?
  • Is the order, stock or booking system the bot will connect to ready?

If the answers point to short steps, occasional use and notifications, a bot is enough. If long sessions and rich screens dominate, consider a mini app or a full app.

A bot goes to where your customers already are; an app invites them into a space of your own.

A practical rule: start with a bot. Once users begin hitting its limits, they will tell you which screens need to move into an app.

Work with us

Have a project in mind?

Describe your idea in a few sentences. We'll review it and come back shortly with scope, timeline and budget.