Automation and bots 7 min

WhatsApp and website bots: saving time or wasting it

A WhatsApp or website bot saves time when it handles frequent, simple requests whose answer already exists: an order status, opening hours, booking an appointment. It wastes time as soon as it holds back a customer who needs a person, or answers off the mark without handing over to someone who can help.

Contents7
  1. 01Where a bot really saves time
  2. 02Where a bot wastes time
  3. 03WhatsApp or website chat: what changes
  4. 04Does a bot need AI?
  5. 05The rule that decides everything: the handover
  6. 06How to tell whether a bot will save you time
  7. 07Frequently asked questions

A bot is neither a good nor a bad idea in itself. Placed in front of a question your customers ask twenty times a day, it frees up your team. Placed in front of a customer with a real problem, it becomes a wall, and the customer ends up calling, annoyed, the person they wanted to reach in the first place. This guide from our section on process automation and business bots helps you tell the difference before investing, on WhatsApp as on your website.

In short

  • A bot saves time on frequent, simple requests whose answer is already in your tools.

  • It wastes time when it offers no way out to a person, or when it answers without knowing.

  • Many useful bots need no AI at all: buttons and a database are enough.

  • On WhatsApp, Meta sets the rules, and its billing changed again on 1 October 2026.

  • A bot is judged by its handover: the person who takes over should not have to ask anything again.

Where a bot really saves time

Three conditions, and you need all of them: the request comes up often, it is settled in one or two steps, and the answer already exists somewhere in your systems. A bot creates no information: it fetches what you have, faster than a person opening three applications.

What a bot handles well

  • Where is my order, with the tracking number

  • What are your hours, your address, your lead times

  • Booking, moving or cancelling an appointment

  • Sending a missing document for a file

What it should hand to a person

  • A complaint, an unhappy customer

  • A request outside the planned scope

  • A commercial gesture, a refund, an exception

  • Anything that commits the company

There is also a use where the bot wins almost every time: the team bot. A salesperson logging a meeting from their phone, a technician sending a photo of a job sheet, an alert landing in the team channel when an order is out of the ordinary. Here nobody expects a conversation: the bot replaces a form or a round of re-keying.

Where a bot wastes time

When there is no way out

This is customers’ first complaint. In a Gartner survey published in July 2024, covering 5,728 customers, the top concern about AI in customer service was that it would become harder to reach a person. A bot that loops over three options without ever offering “talk to someone” turns a simple question into an angry phone call, and the time “saved” is lost again at the switchboard.

When it answers without knowing

A bot connected to a language model can answer confidently a question it has no answer to. The company remains responsible: Air Canada was ordered to honour what its chatbot had wrongly stated, a case we cover in our article on AI agents in production. A bot that does not know should say so and hand over, not improvise.

When it knows nothing about your customer

A bot that asks for the order number, then passes the conversation to an agent who asks for the order number again, has wasted the time of two people. Without a link to your CRM or order system, the bot is just a slower form.

Watch out

The hidden cost of a misplaced bot

A wrong answer costs twice: the time of the customer who believed it, then the time of the agent who has to correct it, often facing someone more annoyed than at the start.

WhatsApp or website chat: what changes

On the website, the visitor is often anonymous. If they close the tab, the conversation ends, unless you asked for a way to get back to them. Website chat is mostly for guiding: finding the right page, the right offer, the right form.

On WhatsApp, the customer is already there, the conversation stays on their phone and they can come back the next day. In return, Meta sets the rules: a dedicated number, approved message templates to write first, a 24-hour window opened by each customer message, and billing that has changed several times. According to the WhatsApp Business Platform pricing page and a TechCrunch article republished by Yahoo:

  1. 1 July 2025

    Meta moves to per-message billing, by category (marketing, utility, authentication, service).

  2. 15 January 2026

    General-purpose AI assistants can no longer use the WhatsApp Business API; a bot that serves a company’s customers is still allowed.

  3. 1 October 2026

    Service messages are charged beyond 1,000 per month and per number.

A WhatsApp bot that answers everything, including what a page of the website would explain better, therefore now has a per-message cost on top of its build cost. For your team, Telegram is often simpler: no platform fees, and no message templates to get approved.

Does a bot need AI?

Not necessarily. A button bot (“1. Track my order”, “2. Book an appointment”) only needs rules and access to your data. It always gives the same answer to the same question, costs nothing extra per message and cannot make anything up. Our guide to automation without AI, when a rule is enough covers this choice.

AI is justified at one precise step: understanding a request written freely (“the parcel arrived open”) to file it under the right reason, or preparing a draft reply that a person reviews. It sorts; it does not decide.

And if the bot uses AI, the person must know. Since 2 August 2026, Article 50 of the AI Act requires that people be informed they are interacting with an AI system, unless this is obvious from the context. Our guide to the AI Act and GDPR for a small business explains who carries this obligation.

The rule that decides everything: the handover

A good bot is recognised by the way it stops.

A bot that saves everyone time

  1. Recognise

    The bot identifies the reason for the request and the customer.

  2. Answer

    Only if the answer is in your data.

  3. Hand over

    Otherwise, to a person, with all the context.

  4. Take over

    The agent sees the history and asks nothing again.

The person who takes over, agent, receptionist or salesperson, remains the one who settles the cases that matter. The bot takes the repetitive questions off their hands and prepares the file; it does not replace them.

From the field

Our expense-entry agent has run on Telegram in production since July 2026: you send it a photo of a receipt, it proposes the entry, and a person approves it with one button before anything is written. On an online plush-toy store, the control centre that sends the store’s events to four team channels was delivered and tested in a single day. On Palmora Property, a real-estate portal, alerts arrive in two dedicated team channels. In all three cases, the bot serves the team, and a person decides.

How to tell whether a bot will save you time

  1. Count.

    Sort a month of incoming requests by reason: you will see which ones really come back.

  2. Sort.

    Keep the frequent reasons whose answer is already in your tools.

  3. Plan the way out.

    Decide who takes over, on which channel, and how fast.

  4. Read.

    Read conversations every week during the first month: they show where the bot gets stuck.

Our business bot on Telegram or WhatsApp, connected to your CRM and in service within a week starts from this sorting. When the bot triggers a longer process (an order to create, a file to complete), our n8n automation, with or without AI takes over.

The other articles in the section cover neighbouring topics: customer support where AI sorts and people review, AI voice agents on the phone in 2026, automated supplier invoices, employee onboarding, reports without copy-and-paste, choosing between n8n, Make and Zapier, monitoring n8n workflows, automating software without an API, the n8n licence and AI and young people’s first jobs.

Frequently asked questions

Does a WhatsApp Business bot cost money?

Yes, per message, since 1 July 2025. Meta bills the messages a business sends by category, and since 1 October 2026 service messages, those that reply to a customer within the 24-hour window, are charged beyond 1,000 per month and per number. Add to that building and hosting the bot.

Does a chatbot need AI?

No, not for simple requests: a button bot connected to your data always answers correctly and costs nothing extra per message. AI becomes useful to understand a freely written request and file it under the right reason, or to prepare a draft reply that a person reviews.

Do you have to tell customers they are talking to a bot?

If the bot uses AI, yes: since 2 August 2026, the AI Act requires that people be informed they are interacting with an AI system, unless this is obvious. It is also a matter of trust: a customer who finds out afterwards that they were talking to a machine takes it badly.

Is a bot better on WhatsApp or on the website?

It depends on who is writing. Website chat guides an often anonymous visitor; WhatsApp suits a known customer who wants to follow their request over time. For your team, Telegram is often the simplest. In every case, the bot must be able to hand over to a person.

Sources

  1. A Gartner survey published in July 2024 (new tab)gartner.com
  2. The WhatsApp Business Platform pricing page (new tab)developers.facebook.com
  3. A TechCrunch article republished by Yahoo (new tab)tech.yahoo.com
  4. Article 50 of the AI Act (new tab)ai-act-service-desk.ec.europa.eu
Source 1

A Gartner survey published in July 2024

Read on gartner.com(new tab)
Source 2

The WhatsApp Business Platform pricing page

Read on developers.facebook.com(new tab)
Source 3

A TechCrunch article republished by Yahoo

Read on tech.yahoo.com(new tab)

Jim · founder of Brain Plus AI

Also in automation and bots

  • Practical guide12 min

    Automate customer support: AI sorts, humans review

    Incoming tickets and emails read, sorted and prepared by AI, then reviewed by your team when the reply commits the company: how to split the work, what can go out on its own, and why sorting costs a few cents when each step has the right model.

  • Understand7 min

    AI voice agents on the phone: what they can do in 2026

    A voice agent picks up, understands a spoken request and answers with a synthetic voice, connected to your phone system. What it does well in 2026, what it does badly, the rules to follow, and the place of the people who answer the phone.

  • Choose and price9 min

    Automation without AI: when a rule is enough

    Not every automation needs AI. A fixed rule costs less and always gives the same result: here is how to spot the steps where AI really pays off, and those where it only adds cost.

WhatsApp