{"id":7996,"date":"2026-08-01T00:45:00","date_gmt":"2026-08-01T00:45:00","guid":{"rendered":"https:\/\/dextora.agency\/insights\/telegram-bot-instead-of-manager-when-automation-works\/"},"modified":"2026-08-01T00:45:00","modified_gmt":"2026-08-01T00:45:00","slug":"telegram-bot-instead-of-manager-when-automation-works","status":"publish","type":"insight","link":"https:\/\/dextora.agency\/en\/insights\/telegram-bot-instead-of-manager-when-automation-works\/","title":{"rendered":"A Telegram Bot Instead of a Sales Rep: When Automation Saves Money and When It Just Annoys Customers"},"content":{"rendered":"

Replacing a sales rep with a bot sounds appealing: it works around the clock, never gets sick, never asks for a raise. In practice half of the bots that get deployed are switched off within six months, and some of them managed to damage customer relationships first.<\/p>\n

The technology is not the problem. The problem is that a bot and a person solve different tasks, and confusing the two costs more than having no bot at all.<\/p>\n

What a bot does well<\/h2>\n

Automation wins where a question repeats and has exactly one correct answer.<\/p>\n\n\n\n\n\n\n\n\n\n
Scenario<\/th>\nWhy it suits a bot<\/th>\nWhat it saves<\/th>\n<\/tr>\n<\/thead>\n
Order status<\/td>\nThe answer comes from a database, no variants<\/td>\nUp to a third of all enquiries<\/td>\n<\/tr>\n
Booking an appointment<\/td>\nSelection from a schedule, no judgement needed<\/td>\nCalls outside working hours<\/td>\n<\/tr>\n
Initial qualification<\/td>\nThree or four scripted questions<\/td>\nRep time spent on poor-fit leads<\/td>\n<\/tr>\n
Standard questions<\/td>\nHours, address, delivery terms<\/td>\nThe same answers typed daily<\/td>\n<\/tr>\n
Reminders and notifications<\/td>\nOne-way flow<\/td>\nNo-shows and missed deadlines<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

Where a bot becomes irritating<\/h2>\n

Irritation comes not from automation itself but from four specific things.<\/p>\n

No route to a human<\/h3>\n

The crudest mistake. If a customer with a non-standard situation circles a menu with no “talk to a person” button, they will go to a competitor and explain why.<\/p>\n

The bot pretends to be human<\/h3>\n

When someone realises they spent half an hour writing to a script calling itself Elena, it reads as deception. An honest “I am a bot and can help with the following” performs better.<\/p>\n

A menu instead of an answer<\/h3>\n

A four-level tree that does not contain the option needed. The person wanted to ask one question and received a quest.<\/p>\n

A bot in complex sales<\/h3>\n

Where the decision is expensive and depends on detail, automation at the first step destroys trust. The client wants to know whether you understood their situation, not to be handed four buttons.<\/p>\n

The simple rule that divides them<\/h2>\n

The bot takes questions whose answer does not depend on context or on who is answering. The person takes everything else.<\/p>\n

One test question: could two employees give different but equally correct answers to this? If yes, it is not a bot task. “When will order 4512 arrive” has one answer, so automate it. “Would this kind of engagement work for us” has many, so keep it human.<\/p>\n

How to calculate the saving<\/h2>\n

The argument “a bot replaces a rep” is almost always wrong. The correct calculation is different: how many enquiries the bot closes completely, without escalation.<\/p>\n

The sequence:<\/p>\n

    \n
  1. Take a month of enquiries and break them down by topic.<\/li>\n
  2. Mark the topics where the answer is unambiguous. Usually 40 to 60% of the whole flow.<\/li>\n
  3. Estimate how many of those the bot closes without handing over. Realistically about two thirds of the marked set.<\/li>\n
  4. Multiply by average handling time.<\/li>\n<\/ol>\n

    The resulting figure is not a headcount reduction. It is time freed up for the complex enquiries that actually generate revenue. That is how it should be sold internally.<\/p>\n

    The mandatory minimum for a decent bot<\/h2>\n
      \n
    • A route to a human at every step.<\/strong> Not at the end of a branch, always.<\/li>\n
    • Honest introduction.<\/strong> It should be immediately clear this is automated.<\/li>\n
    • Handling of unclear requests.<\/strong> Instead of “I do not understand”, hand over to a person with the conversation history.<\/li>\n
    • Memory of context.<\/strong> The customer should not type their order number twice.<\/li>\n
    • A link to the CRM.<\/strong> Otherwise conversations live apart from customer history and the rep starts from zero.<\/li>\n
    • Explicit hours.<\/strong> If nobody can pick up at two in the morning, the bot should say honestly when a reply will come.<\/li>\n<\/ul>\n

      Deployment mistakes<\/h2>\n
        \n
      • Starting with the hard part.<\/strong> Trying to cover every scenario at once produces a year of development and a bot nobody uses.<\/li>\n
      • Nobody reads the logs.<\/strong> The transcripts contain the real questions customers ask. It is the cheapest source of improvement and it is almost always ignored.<\/li>\n
      • The bot as the only channel.<\/strong> Removing phone and email in the name of automation is a reliable way to lose part of your audience.<\/li>\n
      • No owner.<\/strong> A bot needs monthly refinement. Without someone responsible it freezes and ages alongside the price list.<\/li>\n<\/ul>\n

        Where to start<\/h2>\n

        Take the single topic that consumes most of your team’s time and build a bot for that alone. Two weeks of development, a month of observation, reading every conversation. Then add one scenario at a time.<\/p>\n

        Automation saves money exactly up to the point where the customer notices things got worse for them. That line is drawn not by technology but by whether you left a person available at the moment one was needed.<\/p>\n","protected":false},"excerpt":{"rendered":"

        A bot and a person solve different tasks. We break down which scenarios are genuinely worth automating, why menus with no route to a human drive customers away, how to calculate the real saving, and where to start.<\/p>\n","protected":false},"author":5,"featured_media":7992,"template":"","insight_category":[156],"insight_tag":[180,168,166],"class_list":["post-7996","insight","type-insight","status-publish","has-post-thumbnail","hentry","insight_category-trends","insight_tag-automation","insight_tag-business-process","insight_tag-crm"],"acf":[],"_links":{"self":[{"href":"https:\/\/dextora.agency\/en\/wp-json\/wp\/v2\/insight\/7996","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/dextora.agency\/en\/wp-json\/wp\/v2\/insight"}],"about":[{"href":"https:\/\/dextora.agency\/en\/wp-json\/wp\/v2\/types\/insight"}],"author":[{"embeddable":true,"href":"https:\/\/dextora.agency\/en\/wp-json\/wp\/v2\/users\/5"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/dextora.agency\/en\/wp-json\/wp\/v2\/media\/7992"}],"wp:attachment":[{"href":"https:\/\/dextora.agency\/en\/wp-json\/wp\/v2\/media?parent=7996"}],"wp:term":[{"taxonomy":"insight_category","embeddable":true,"href":"https:\/\/dextora.agency\/en\/wp-json\/wp\/v2\/insight_category?post=7996"},{"taxonomy":"insight_tag","embeddable":true,"href":"https:\/\/dextora.agency\/en\/wp-json\/wp\/v2\/insight_tag?post=7996"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}