How not to lose enquiries in the evenings and at weekends

Evenings and weekends are not a small leftover of enquiries

A sales department usually works nine to six, five days a week. Clients do not live that way: one looks for a contractor late in the evening, another leaves an enquiry on Saturday because that is when they got round to it. While the office is closed, the enquiry simply lies there waiting for Monday or for morning.

In the live GetGate project 27% of enquiries came in exactly in the evening and at weekends — they used to be answered only the next morning. That is not a leftover but almost a third of the flow. If such enquiries are treated as secondary, every fourth lead is lost, and often the hottest one: the person wrote exactly when the need arose.

The problem is not that the sales department works badly. The problem is that the salespeople's schedule and the clients' schedule do not coincide. Solving it by adding shifts is expensive and awkward, so it is more logical to split the tasks: part of the enquiries is closed by a person, part by automation.

An on-duty salesperson, a chatbot or a voice robot — how to choose

An on-duty salesperson makes sense if there are few evening enquiries and they are varied: today a question about delivery, tomorrow about a return. It is hard to write a scenario in advance here, and a person will cope faster than any automation. The downside is obvious: being on duty costs money and rests on a person who also gets tired and falls ill.

A chatbot with fixed buttons is suitable if clients' questions repeat and fit into three or four scenarios: price, availability, booking. It will not replace a live conversation, but it will close standard enquiries without a person. Its weak point is that as soon as a question goes beyond the scenario, the bot gets lost and the client leaves.

A sales robot is needed where there are many enquiries, they come from different channels and the client expects a conversation rather than buttons. It answers and qualifies the client in Telegram, WhatsApp, Instagram, Avito and on the website through AmoCRM, recognises voice messages and photos, and remembers the conversation history. This is no longer a night-time stopgap but a part of the sales department that works while everyone sleeps.

A missed call in the evening is not the same as a forgotten chat

A client sometimes forgives an unclosed chat: they see the message was delivered and wait for a reply in the morning. With a phone call it is different: ringing, then silence — and the person immediately dials the next company on the list. A call is either answered or lost for good; there is no third option.

Most articles about evening enquiries discuss chats, forms and messengers and say almost nothing about missed calls. Yet a call in the evening or at the weekend most often means the person is already ready to buy and is ringing to check the last details. Silence on the line at that moment costs more than an unanswered message in a chat.

The solution is not to leave the call without a move. A sales robot can take the enquiry on the voice channel, ask clarifying questions and record the enquiry in the CRM, and if a human is definitely needed, set a task for the on-duty salesperson for the morning with the full context of the conversation. The client gets an answer at once, not twelve hours later.

What the robot decides itself and what it passes to a human

Not every enquiry should be given to the robot in full. The logic is simple: standard questions, qualification and booking go to the robot; complex cases and hot deals to the salesperson. The robot answers questions about the service, clarifies order parameters, books the client at a convenient time and enters all of it into the CRM.

As soon as a sign of a hot client appears in the conversation — readiness to pay, a specific question about the contract, dissatisfaction — the robot passes the dialogue to a salesperson and attaches the conversation history. The salesperson does not start from scratch but sees at once what the client already asked and what they agreed to. That is escalation: not “the bot broke” but a transition rule described in advance.

Such rules are set up for the client's niche rather than taken from a general template: in one place a hot signal is a question about instalments, in another a request to send bank details. The more precisely these signals are described, the smaller the chance that a live client gets stuck in correspondence with a robot when they need a person.

What this looks like in amoCRM

Out-of-hours enquiry handling cannot be set up “in general” — it is done inside a specific CRM, with specific pipelines and automations. When implementing CRM pipelines are built for the client that mirror their sales process, channels and telephony are connected, and tasks that used to be done by hand are automated.

For evening and night enquiries this means: the deal is created automatically at the moment of the enquiry, whatever the time. The robot replies to the client while the system in parallel sets a task for the on-duty salesperson or moves it to the morning, depending on how hot the lead is. Nothing gets lost in email or in individual employees' personal chats.

Integrations with payments and documents are set up the same way: if a client is ready to make a prepayment in the evening, they do not have to wait until morning to get the bank details or a payment link. The team meanwhile is trained to work with the new scheme — otherwise the automation stays a pretty diagram on paper.

What the robot says when it replies out of hours

The scenario for evenings and weekends is built the same way as for a working day, only without the option to switch the client to a live person in the first minute. The robot says hello, checks what the enquiry is about, asks two or three qualifying questions — budget, timing, what exactly is needed — and records the answers in the deal card.

If the client sends a voice message or a photo, the sales robot recognises them and answers to the point rather than asking them to write in text. It also remembers what was discussed earlier, if the client already wrote during the day and is simply continuing the conversation in the evening. This does not look like being switched to a robot but like the continuation of a single dialogue.

How well this works without a person was checked separately: in a blind comparison of the assistant's replies with those of a live salesperson, in 80% of cases the assistant's reply was no worse. That is not a guarantee for any niche but the result of a check on real correspondence, and it shows that an evening conversation with a robot is not necessarily worse than a daytime one with a person.

How to know the scheme is working

Setting up evening enquiry handling is not enough — you need to see what comes of it. For that we use sales analytics: plan against actual in real time, a breakdown by channel and by time of enquiry, and alerts if the evening flow suddenly starts sagging without explanation.

Separately there is quality control: AI listens to calls and chats, scores them against a checklist and looks for the places where the conversation could have ended in a refusal when it need not have. This works the same for daytime and evening enquiries, and it is the evening ones that are worth checking more closely — there is no salesperson nearby to correct things on the fly.

In the live GetGate project, after this scheme was introduced, a salesperson frees up about an hour of working time a day and five product groups are counted automatically without manual reconciliation. These are the figures of one project, not a market average, but they show the direction: less routine, more time for complex deals.

Where to start

The first step is not buying a robot but reviewing the current sales department: where exactly enquiries are lost, how many arrive in the evening and at weekends, which channels are involved. Without this step any automation is built blind and repeats other people's mistakes.

Next the CRM is assembled or refined, the robot is trained for the client's niche and the necessary channels are connected — from messengers to telephony. In parallel the department's other processes can be built: scripts, rules, analysis of why clients say no — that is handled by to build the sales department.

After launch what remains is to watch plan against actual, notice where the scheme fails and tune the escalation rules. Evening and weekend enquiries will not disappear — but they will stop being a blind spot nobody is responsible for until Monday morning.