What’s the biggest advantage the chatbot has given you?

“Not having to manually answer every single question my customers ask. And not walking into the office on Monday with more than 70 contacts waiting for a reply.”

Natalia Rodríguez — Director, Alier

In four months

What changed, in numbers.

  • 88% of the conversations the assistant handled began outside office hours (37 of 42).
  • 28 seconds median first reply outside office hours.
  • MXN 69,338 bought by customers who came through the chat, across 53 orders. Of that, MXN 61,727 was bought after talking to the AI.
  • 22 sales worth MXN 23,138.16 that the assistant took part in directly.
  • Eight sales carried from the first question to the payment receipt, all outside office hours and seven of them on weekends.
  • 2.5 to 3 hours the team spent every Monday catching up on the weekend’s messages.

The problem

One phone, one person, and Mondays.

Every order came through a single line — the same one used for calls and for WhatsApp — operated by one staff member whose day started at 8:30. After 6:00 in the evening on weekdays, from 2:00 on Saturday and all day Sunday, customers got nothing but the WhatsApp Business auto-reply saying the office was closed. Nobody read their questions until the next working day.

On weekdays, catching up on the previous night’s messages took 40 minutes to an hour. On Mondays, with everything piled up since Saturday noon, it took between 2.5 and 3 hours. Two or three more people had to step in on Monday morning to answer messages and assemble orders before the 1:30 pm carrier cutoff.

The cost was not only time. When a reply landed a day or two late, some of those customers had already lost interest. And the questions were always the same: product, price, coverage by postal code, shipping cost, and payment methods.

Why it had not been solved

She had already gone looking. Nothing fit.

Natalia got quotes from about three companies and ran into the same thing every time: oversized packages — 5,000 messages a month when she needed 500 to 700 — high up-front costs, and the sense that they would leave her alone with the tool. She also tried building block-by-block flows on another platform and found it overwhelming.

The underlying worry was technical: migrating the number to the WhatsApp API sounded complicated, and she was afraid of losing her contacts and her history. In practice the migration turned out to be simple and fast.

Implementation

One day of training, and the rest was ours.

Training the assistant took Natalia a single day: about 3 hours to load product, pricing and shipping information, 2 to 3 hours of initial training, and an hour of testing.

Those tests used real customer messages taken from her other number, typos and abbreviations included. The assistant understood the context and the variations in how people write. For later adjustments, a thumbs-up and thumbs-down system lets her correct a single answer without touching the whole script.

Fernvay handled the rest: migrating the number, the configuration, support when something broke, and the adjustments — without charging for each change separately.

You can see the interface and everything the assistant does on our AI chatbots page.

Four-month results

Every figure here has a source.

Result Figure Source
Conversations that began outside office hours 88% (37 of 42) Platform records, 24 Jun – 21 Oct 2026
Median first reply outside office hours 28 seconds Platform records
Customers who came through the chat and bought 30 customers, 53 orders, MXN 69,338 Alier’s sales records
Of those, what they bought after talking to the AI 28 customers, 45 orders, MXN 61,727 Alier’s sales records
Sales the assistant took part in directly 22 orders, MXN 23,138.16 Sales records cross-checked against the conversations
Sales carried from the first question to the payment receipt 8 WhatsApp conversations
Sales that started at night or on a weekend 13 of the 14 closed in chat; four of those customers have already reordered Conversations and sales records
Website failures caught at night 9 cases: coupons, discounts, cancelled orders, sizes and declined payments WhatsApp conversations
Wholesale leads captured outside office hours 4: a nursing home, two resellers and a veterinary hospital WhatsApp conversations
Team time on Monday mornings Before: 2.5 to 3 hours Natalia Rodríguez

The team also gained a shared inbox. Everything used to live on one phone; today they work from desktop and mobile, and the platform records who answered each message. Every morning a summary of the previous day arrives by email with the open items and the alerts.

The weekend both numbers were on

Everything above came from the smaller of two lines.

Alier runs two WhatsApp lines. The original is the one its marketing agency uses in every campaign, and the only one that takes phone calls. The second is the one migrated to the WhatsApp API for the assistant.

For four months the assistant covered only the second. The original kept being answered by hand, for fear WhatsApp would block the account if it were connected another way.

On the weekend of 24–25 October, Natalia connected the original line as well, because the staff member would be away and she had events both days. It was the first time the assistant handled both numbers at once.

  Weekend of 24–25 October
New contacts 68
Messages sent 419
Written by the assistant 368 (88%)
Conversations resolved with no human stepping in 53 of 71
Escalations to the team 3, all closed
Conversations with negative sentiment 0

And the figure that reframes everything above it: of the 71 contacts handled that weekend, 60 came in on the original line. The assistant had spent four months covering the line that carries roughly one in six conversations. The account was not blocked.

Three moments

What a WhatsApp chatbot looks like on an ordinary Sunday.

A Sunday at midday. A customer asked about prices. The assistant replied in 28 seconds, explained sizes and absorbency, quoted with the first-purchase discount and the shipping cost to her postal code, shared the bank details, received the receipt, and took the delivery address. The whole order closed in 25 minutes, on a Sunday, with nobody from the team working. And that is the point: whoever comes in on a 30-piece pack is the one who later buys 100 and 200.

A veterinary hospital. It came in outside office hours asking about wholesale: 700 units a month. The assistant took the details and the volume and passed them to the commercial email. The customer closed, and has now placed three orders.

The Golden Sale. In the week of Grandparents’ Day, Alier sent its promotion as a mass broadcast from the platform. The year before, a similar attempt from WhatsApp Business on Mother’s Day ended with the account blocked. Here the system controls the sending pace and pauses it if response rates drop, so the number is not put at risk. The number withstood the broadcast without being blocked.

In her words

Natalia Rodríguez, Alier.

“Sales went up. We started selling through WhatsApp at hours when the office was closed and on weekends — hours that simply went unanswered before.”

“If the chatbot were taken away today, we would lose around 3% of our monthly revenue. And that percentage grows every month.”

“Our customers are older adults and their families. They ask a lot of questions, and often they repeat the same one. An AI chatbot has patience that does not run out: it answers every question with the same calm, and it looks after how the customer feels, depending on how you train it. That has shown up in our service.”

“The support was first-rate. There were a few snags along the way, mostly with the WhatsApp API connection, and in under 24 hours we had a solution — for that and for everything else.”

Translated from Spanish.

How this was measured

No figure here comes from us alone.

  • The period runs from 24 June to 21 October 2026.
  • Conversation counts, response times and message volumes come from the platform’s own records.
  • Peso amounts come from Alier’s sales records, cross-checked conversation by conversation against the WhatsApp chats.

The full report, with the detail of every conversation and the classification methodology, is available on request.

Who this works for

If your business answers messages the next day, there are sales waiting there.

A WhatsApp chatbot suits a business that sells by message and whose information needs explaining: product, sizes, coverage, shipping, payment methods. It works best in B2C, where the volume of people is high and the questions repeat, whether on WhatsApp or Messenger.

Book a Free Workflow ReviewSee how the assistant works

Privacy Preference Center