Check-list for your bot. 5 mistakes that should be avoided.

Oct 3, 2017 · 5 min read

Each day new chat bots emerge in the market. 99% of them are doomed to remain unknown and bring their developers nothing but losses and wasted time. We’ve composed a short check-list, that helps to evaluate your bot and increase its success chances.
For the last few months we, WeMakeBot company, are frequently facing requests from the following kinds of customers:
• Ones who have just recently discovered the possibilities of chatbots and who are eager to implement new technologies in their business processes. They imagine bots as demigods or big red button, which says “Profit”
• Ones who have already applied chatbots to their business processes but cannot receive any profit from the technology. Such bot owners often blame the social media marketers or people who were responsible for bot promotion.

After analyzing most frequent bot problems and burying dozens of projects, we’ve composed a short list of kisses of death for new chatbots. We view the absence of this problems as required but not the definite condition for success. We would be grateful if you, dear readers, share your opinion on the conditions which define chatbots success, please write them as comments to the article. So let’s start with the list:

Chatbot tries to save troubled business

Chatbot is capable and must reduce costs of human resources. It could also attract new customers and clients when provides useful functions. Moreover, bot could be an instrument to hold clients in case they leave and abandon business web page and social media community. However, bot cannot save business facing problems with its product competiveness, high production costs production quality. We are sorry for such outcry, but we often hear from our clients phrases similar to: “I know Ivan has ordered a bot some time ago and now is selling his business”. Explaining that Ivan’s business has, probably, malfunctioned and not the bot is becoming more and more tiring.

Bot does not improve UX

Easy-to-use interface and low access barrier (compared to mobile apps) are the advantages of chatbot. Even these pros cannot beat the cons if the bot is useless. If one is offered a bot which will reply to customers’ messages in Facebook and inform them about the address of one of two shops or will send “please call number 8 (499) 391–98–88 if you have any questions” in response to every unknown message; please drive the person who has offered such bot away and do not hesitate to use sticks, bats and certain degree of appropriated violence.

Bot has to be useful. If business is able to provide useful info to customers without a bot it should do so. One can force clients to chat with bot, for instance, by deleting contact details from the webpage, but customers most definitely flee to another company. Do not create different communication channels just for the sake of Omnichannel communication. But to solve user’s actual needs with new channel instead.

Users cannot definitely say in what part of the conversation with the bot they are

To put it simple, bot without easy navigation — is a UX killer. This principle of chatbot development appears to be obvious, but many still neglect it. Here is a simple example: each bot has a “I do not understand” response. Even if developers consider Levenshtein distance and any kind of misprints, users still able to surprise them. These are two awful and easy-to-make mistakes connected to a “I do not understand” response:

  • Bot automatically sends user to the main menu
  • Bot leaves the user without a hint what to do next after response

Analyzing the dialogs between users and bots which are being blocked one can spot that often the last bot’s response was “I do not understand” and users’ exit to main menu.

Users should see the structure of the bot clearly: navigation from main menu to various parts of the bot, end points of the conversation, ways to go back to previous parts, to main menu and so on. Changing of bot’s behavior after analyzing clients’ communication with it is a white whale for now. Obsession with this goal could puzzle users and result in negative UX.

Bot fails to transfer users’ progress to webpage

Adding a chatbot which has a function of adding goods to virtual bag/cart/basket to online shop one should make sure goods stay in basket when user go to the webpage or switch to other communication channel. Payment within a messenger would be ideal (if messenger allows to conduct online payments), transferring client to a secure payment session is less comfortable yet an acceptable end of communication. Taking customers to bag/cart/basket on webpage or mobile app with saving all their progress is user’s minimum expectation and one should not go below it.

The same is with similar kind of business. For instance, if bot sells insurance, it should be able to save and transfer all the data gathered from clients.

Bot sends spam

A muted bot is a disaster. Yet as bot is a new and free channel of communication owners do not see the line and try to squeeze all out of bot’ communication possibility. This, obviously force users to mute bots or block messages from them. Playing short game is not worth it. Just don’t send spam.


If you are not still sure whether your business need or does not need a chatbot just contact us. We are responding to all inquiries. Even to requests of businesses which cannot fit a chatbot into their processes and to people who want private chatbots (once we were asked to create a chatbot for a customer who wanted bot to answer his wife, family and friends when a client is busy, true story). We would be happy to consult you whether it is rational to use chatbot in your field.

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade