The correct way to build a Chatbot that gathers critical information

In a pragmatic conversation there are a number of issues that are considered "Critical information". Without these critical pieces of information, it won’t be possible to perform the task the conversation was made for. A smart Chatbot knows how to collect those pieces and how to ensure that it has all the critical information. Before choosing the Chatbot platform, it is important to understand whether your business needs a "Critical information gathering" mechanism, and if so, which Chatbot platform will perform the task most effectively. In this article you will learn about the ways in which the Chatbot collects vital information, and how to help it achieve that goal.

For demonstration purposes, let's imagine that we own a local pizzeria and we want to insert a Chatbot into our website. What are the essential questions we must formulate to make it possible for the Chatbot to take the order from the customer most effectively?

"What type of pizza would you like to order?"

"Where would you like to receive the invitation?"

Apparently, without these two details, it will be impossible to make the order. But are these the only details we would need for the order? Probably not. We will also want to present the customer with additional offers, such as:

"Would you like to order a drink?"

"Would you like pizza extras?"

As opposed to the first pieces of information, this additional information serves for the purpose of enriching the order, even though the order can be made without it. When designing a Chatbot, it is important to distinguish between questions that are of utmost importance, (without them there will be no deal,) clarification questions, and extra info that should probably be included, but one can make the order without it.

The customer can leave at any moment

The reason for this is that when chatting with a Chatbot, you can not tell if and when the customer will leave the conversation. Unlike a frontal conversation or a phone call, where a company representative can always say "Wait, I forgot to ask where you want us to send the order?"

When chatting with a Chatbot, the customer can leave the screen at any moment, and the Chatbot owner has no way of knowing whether the customer has lost interest, her battery died suddenly, or perhaps his girlfriend suggested to drop it and go out for pasta in a romantic restaurant instead.


Talk business

One of the best ways to prevent the customer from abandoning the conversation is by creating a businesslike conversation. The customer, even if he/she will not admit it, quickly identifies whether the conversation is progressing in the right direction, or it is diverting into small and less important details. If the first questions are important and practical, they create the sense of rapid progress in the discourse, and succeed to capture the customer’s attention. The customer can tell that "something practical and effective is happening." If the questions lead the customer into small details right at the beginning of the conversation, the customer is likely to anticipate a long and tedious conversation, which poses the danger that they will despair and leave the Chatbot before completing the order.

For that reason, when formulating a conversation for the purpose of ordering a pizza, we will prefer to plan it in the following order:

"What kind of pizza would you like to order?" (top priority)

"Where should we send the pizza?" (top priority)

"Would you like some extras on the pizza?" (Secondary importance)

Instead of the following:

"What type pizza would you like to order?" (top priority)

"Would you like some extras on the pizza?" (Secondary importance)

Customer: "Olives".

Chatbot: "We have black olives and green olives, what would you prefer?"

"Where should we send the pizza?" (top priority)

planning a Chatbot conversation process:

Stage 1 - Collecting the critical questions

Consider what details are the most critical and necessary to enable the customer to make the order. Write down these questions, then try to choose a few that you can do without. Our goal is to require minimum effort from the customer when they fill in the details. If a question is only borderline essential, it is usually best to let it go.

Stage 2 - Collecting secondary questions

Think of all the details you would have liked to collect from the customer had they had all the patience in the world. Do not be conservative, write down all the questions that come to mind, there are clients that will be happy to provide us with very useful information.

For example:

"By the way, only if you have time and would like to answer, I see that you booked a room in our hotel last summer, were you satisfied with our service?" After the customer already filled out all the essential details, the client may agree to summarize their service experience from last summer.

Stage 3 - Optimal conservation planning

Arrange the questions in the order that seems most logical to you. Then ask your acquaintances: "Which conversation sounds more businesslike?" These answers will be your guide to figuring out the most optimal arrangement for your Chatbot conversation.


The order in which the Chatbot collects the information is of great importance. Understanding the correct order is crucial upon constructing a Chatbot, since it can severely influence the Chatbot’s capacity to gather the essential data. It is best to avoid wasting the end user’s time with questions that are not truly necessary.


עוד מאמרים

שיכפול צ'טבוט

התראה על פעילות בצ'ט

כיצד לבחון "הבנה" של שפה טבעית בצ'אטבוט

יכולות הלימוד של הצ'אטבוט של חברת Liberty - Chatbot

מידול גרפי לבניית צ'טבוט, האם זה אפשרי?

הבעייתיות במידול שיח לכדי תוכנת צ'אטבוט

צ'אטבוט במקום טופס באתר אינטרנט

צ'אטבוט כדרך לשמור על ידע בארגון

שיחות "איפה" ושיחות "משימה"

תהליך הקמה של צ'אטבוט

צ'אטבוט לקהל הרחב לעומת צ'אטבוט פנים ארגוני

למה הצ'אטבוט הוא באמת מהפכה

כיצד צ'אטבוט יכול לקדם שיחה עניינית ויעילה לביצוע משימה

שיטות מקובלות להבנת שפה טבעית בצ'אטבוט

כיצד לבנות נכון צ'אטבוט לאיסוף נתונים

חזרה אל: כל המאמרים