The first month with an AI agent on WhatsApp: week one's gaps and nerves, week two's fixes, week three's quiet, week four's numbers, and the mistakes that make the month harder than it needs to be

Switching on an AI agent is not a flip of a switch after which everything is handled. It is a month in which the agent learns your business from the questions customers actually ask, and you learn to trust it. Here is what each week usually looks like, so the first one does not scare you off.

By Ritchie, ReplyKit · Published 5 October 2026 · 6 min read

Quick answer

Day one is an afternoon of setup: pair the number, paste your knowledge, write ten lines of instructions, test, go live. Week one brings many knowledge gaps and the urge to watch every chat; that is normal. Week two is the teach loop: answer the gaps, adjust tone, tune hand-overs. Week three is when the inbox goes quiet and the "Needs you" column holds only chats worth your time. Week four is numbers: containment, hand-over rate, bookings, and a plan-size check. The mistakes that slow it: thin knowledge, no testing, over-broad hand-over rules, ignoring the gap list, and switching off at the first wrong answer. ReplyKit's 7-day trial covers the first week.

Day one Week one: gaps and nerves Week two: the teach loop Week three: quiet Week four: numbers Five mistakes that drag it out FAQ

Day one

Pair the WhatsApp number by QR (a dedicated number, ideally; see a dedicated number).
Paste the knowledge: prices, hours, services, policies, FAQ, and let it crawl the website; the list is in what to put in the knowledge base.
Ten lines of instructions: who it is, tone, what to ask, what to hand over.
Test with twenty real questions in the sandbox; fix what is wrong. The routine is in testing before going live.
Go live. Two to three hours in total.

Week one: gaps and nerves

Expect a high gap count; customers ask things you did not think to write down. That is the system working.
You will read every conversation. Fine, for a week.
The first time a customer completes a booking without noticing the agent, the nerves ease.
A wrong answer or two is likely, usually from an outdated line in the knowledge; fix the line, not the agent. The causes are in why the agent answered wrongly.

Week two: the teach loop

Answer the gap list in one sitting: the recurring questions become knowledge lines.
Adjust tone from real chats: too formal, too many emoji, too long. A line each.
Tune hand-overs: too many, add knowledge; too few, add trigger words. The method is in the teach loop.
Stop reading every conversation; read the hand-overs and a sample.

Week three: quiet

The gap list is a trickle. The inbox's "Needs you" column has the complaints, the custom quotes and the people who asked for you, and nothing else. The evening backlog is gone. This is the week owners describe as the first one in years without a phone at dinner. Resist the urge to add complexity; let it run.

Week four: numbers

Containment: the share of conversations the agent finished alone. Fifty to eighty percent is typical by now.
Hand-over rate and whether each hand-over was necessary.
Bookings made by the agent, and no-show rate with reminders on.
Reply usage against the plan; resize if needed, as in reply allowances.
Decide the next automation: the website assistant, follow-up nudges, a second number. The metrics are in metrics that matter.

Five mistakes that drag it out

Thin knowledge on day one, then blaming the agent for not knowing.
Skipping the sandbox and letting customers find the errors.
Hand-over rules so broad that everything goes to you, so nothing changes.
Ignoring the gap list, so week two never happens.
Switching it off at the first wrong answer instead of fixing the line.

Frequently asked questions

How long does setup take?

Two to three hours on day one: pairing, knowledge, instructions, testing.

Is it normal to see many gaps in week one?

Yes. Customers ask things you did not write down; the gap list is how the agent learns them.

When does it start saving time?

Week one for quick questions; by week three most owners describe the inbox as quiet.

What should I measure at the end of the month?

Containment, hand-over rate, bookings, no-shows and reply usage.

What is the most common mistake?

Thin knowledge on day one followed by giving up at the first wrong answer.

A quieter inbox by week three

Setup in an afternoon, a teach loop in week two, numbers by week four. 7-day free trial, no card.

Start free

Written by Ritchie, ReplyKitPart of the small team in Malaysia that builds and runs ReplyKit. Writes about WhatsApp automation, AI customer service and small business lead response.

How we know this. This timeline is what ReplyKit sees across new accounts, from the trial fortnight through the first paid month, including the common ways it goes slower. Product figures in this article (limits, prices, per-reply costs) are taken from ReplyKit as it runs today and are re-checked when we update the page. Where we cite outside research, the source is linked below. We sell ReplyKit, so read our product claims with that in mind; we say where a different tool or no tool is the better choice. About ReplyKit.

Sources and further reading

Harvard Business Review, The Short Life of Online Sales Leads (Oldroyd, McElheran, Elkington, 2011) · firms that responded within an hour were about seven times more likely to qualify the lead
WhatsApp Business Messaging Policy · opt-in, opt-out and prohibited-use rules every business must follow

ReplyKit is an independent product and is not affiliated with or endorsed by Meta or WhatsApp. This article is general guidance for running a small business, not legal, financial or regulatory advice.