The problem is the gap, not the team
Someone messages at 9pm asking about your pricing. Your team sees it at 9am. By then they have already booked with someone else.
That is the gap a WhatsApp chatbot closes. Not by replacing your team — by making sure nothing falls through when they are not at their desks. The enquiries that come in overnight, on weekends, in the middle of a busy afternoon when nobody has time to pick up their phone. Those are the ones you are currently losing.
"We were getting enquiries on WhatsApp constantly but our team couldn't keep up. After the bot went live, seventy percent of conversations resolved without anyone touching them."
— Client, professional services firm, UK
First: the API, not the app
There are two versions of WhatsApp for business. The app you download on your phone, and the API.
The app is fine for one person managing a handful of chats. It does not support automation. You cannot connect it to a CRM, build flows, or run anything at scale. It is a messaging app, not a business tool.
The WhatsApp Business API is what actually lets you build something. Automated replies, conversation flows, CRM integration, handoff to a real person — all of it runs through the API.
To get access you need a verified Meta Business Manager account, a dedicated phone number not already on WhatsApp, and approval from Meta. You apply through Meta directly or through a Business Solution Provider. In the UK and US, Twilio is the most straightforward route. In the UAE and Gulf region, 360dialog has the better regional support.
Approval takes two to five working days. After that, the number is your bot.
Before anyone builds anything: read your own messages
Go through your last month of WhatsApp conversations. Write down every question that came in more than twice.
Most businesses end up with the same short list. Pricing. Availability. How to book. Location. Whether they can speak to someone. That list is your chatbot. Start there.
A bot that handles six things cleanly is worth ten times more than one that tries to handle everything and handles none of it well. The ambitious ones get switched off within three months.
Building the flows
Every conversation flow is a decision tree. Customer sends a message, bot matches it to an intent, replies, moves them forward. Simple in principle. The execution is where most bots go wrong.
- Map every endpoint before you build. Every path has to end somewhere useful. Not a dead end, not a generic "we will be in touch." Either the question gets answered, the person gets booked in, or they get passed to a real person with the context already captured.
- Get the handoff right. When a customer needs a human, the bot tells them clearly, gives an honest timeframe, and pings your team immediately. A broken handoff is worse than no bot at all — it tells the customer you do not care.
- Keep the options short. Button menus with two or three options perform better than open text fields. People do not want to write paragraphs to a bot. Give them something to tap.
- Test as a frustrated customer. Go through every flow as someone who does not know how it works and is slightly impatient. Type things wrong. Ask questions outside the scope. See what breaks before your customers do.
A lead capture flow we built recently: customer asks for a quote, bot asks what kind of project, customer picks from a short list, bot takes their name and a quick brief, confirms someone will follow up within a few hours, logs it in the CRM. Ninety seconds start to finish. The team gets a notification with a qualified lead already waiting.
Connecting it to the tools you already use
A bot that only chats is useful. A bot wired into your actual systems is something else.
- Google Calendar or Calendly — live availability, bookings confirmed without anyone checking a diary
- HubSpot or Pipedrive — every lead drops straight into your pipeline, tagged and ready
- Shopify or WooCommerce — order updates, dispatch notifications, abandoned cart recovery
- Airtable or Notion — if you prefer something lighter than a full CRM
What it costs to run in 2026
Meta moved from conversation-based pricing to per-message pricing on 1 July 2025. Worth understanding before you budget.
Four message types, four price points:
- Marketing — outbound promotional messages you initiate. Highest rate.
- Utility — booking confirmations, order updates, transactional messages. Significantly cheaper.
- Authentication — verification codes. Similar to utility pricing.
- Service — replies sent within 24 hours of a customer messaging you first. Free.
Meta per-message rates by region
| Region | Marketing | Utility | Service |
|---|---|---|---|
| United Kingdom | $0.0250 | $0.0158 | Free |
| UAE | $0.0369 | $0.0088 | Free |
| United States | $0.0250 | $0.0050 | Free |
The practical implication: inbound enquiries your bot responds to within 24 hours cost nothing on the Meta side. Your real cost exposure is outbound campaigns. Five hundred contacts in the UK runs about $12–$13 in Meta fees before platform or build costs.
Build costs
- Basic FAQ and lead capture bot: £800 – £2,500
- Full integration — CRM, live booking, escalation flows: £3,000 – £8,000
- Hosting and maintenance: £100 – £300/month
Most clients cover that cost in the first quarter — through leads they would have otherwise missed, or through the time their team gets back from not doing intake manually.
If you want to see what this looks like for your business
We have built these for businesses in the UK, UAE, and the US. Drop us a message and we will give you a straight answer on what makes sense — scope, cost, and whether it is worth doing at all.