Chatbot Software in 2020: Choosing the Best Bot to Attract Customers to Your Business

By:  Tim Schreyer

Chatbots are the wave of the future! Or, at least they were … in 2016.

Well, guess what? It’s the future. Now that some of the hype has worn off, let’s see how chatbots are doing.

Let’s see what kinds of chatbots are available today and what kinds of conversations customers can have with them.

More to the point, let’s see what problems today’s chatbot software solves to benefit your business, and how you can select the type of chatbot that attracts customers to your business.

To get the most value from today’s chatbot software, you will need to understand what chatbots can do for your business, what skills you need to build a chatbot, and what chatbot platforms are available today.


Designed by Freepik

How the Successful Businesses use Bots, and Why

Chatbots automate communication and many of the tasks that are necessary when working with people. Chatbots are good at building relationships. They are conversational, so they can engage customers in a way that is warm and friendly.

Businesses have found A LOT of uses for chatbots!

Chatbots can answer FAQs. Users can make purchases using chatbots, and they can ask chatbots about order status. Chatbots can retrieve interesting articles from a knowledgebase. Chatbots can make reservations for restaurants, flights, hotels and cars. Some companies have even used chatbots to pre-screen job applicants by reviewing resumes and conducting initial interviews. Finally, chatbots can engage in small talk, keeping the conversation light, fun and friendly!

There are several benefits to using chatbots in your business.

Chatbots can talk to many people at the same time, so your customers don’t have to waste time waiting in a queue. Chatbots work on the messaging apps people are already using so your customers don’t have to clutter their smartphones with more apps. Chatbots relieve your employees from repetitive and mundane tasks, and chatbots are available 24/7.

You can make a chatbot using any of the several chatbot builders or chatbot platforms that are readily available today. Most of these do not require any programming skills, but you will need an understanding of your internet server, website, and databases in order to connect your chatbot properly to your business.


Where Chatbot Software fits in your Customer Service System

Think of a chatbot as a person, chatting away with your customers. It sits at the front-end of your customer service system, between the user and the rest of the system.

This is the configuration used for most chatbot implementations.

Human avatars designed by Freepik, robot avatar designed by Top Png, and computers designed by SmashIcon and Eucalyp at FlatIcon


The chatbot speaks to the user through a messaging app. Common messengers are Facebook Messenger, Slack, or a WordPress plug-in on your website.

The chatbot can connect users to a human when necessary. It does this using another messenger, or a Customer Relationship Manager (CRM).

The chatbot can also find information or handle tasks for the user. It uses your system’s Application Program Interface (API) to do this. The API may search through a Content Management System (CMS) to find information.




Designed by kjpargeter at Freepik


The Two Types of Chatbots that are Available Today

There are two types of chatbots: Rule-Based Chatbots and AI/NLP Chatbots.

Rule-based chatbots follow a strict set of rules that define their conversations.

AI/NLP chatbots use Artificial Intelligence (AI) and Natural Language Processing (NLP) to understand and respond using normal human language.

AI/NLP chatbots are fun to talk to! They let users control the conversation and jump to any desired topic.

So what have we learned in the last four years? Which one is best? …


Designed by iconicbestiary at Freepik


Users’ Response to AI: Too Smart, Too Soon?

In 2016, developers were predicting that rule-based chatbots would soon be obsolete.

They were wrong.

It makes sense. AI chatbots could finally talk with humans! We were one step closer to the ubiquitous talking computer from Star Trek! Who needs boring, rule-based chatbots when we can talk to an AI chatbot?

But …

We live in a world of menus and buttons. People like menus and buttons. They don’t have to memorize commands. They don’t have to guess!

Rule-based chatbots can ask multiple-choice questions, giving the user a kind of menu. They can offer buttons, so users don’t have to type their answers.

So which one is best? Either one! They are both great! Your choice depends on how familiar users are with your business.

AI/NLP chatbots work better with users that know what they want, and Rule-based chatbots work better with users who want more guidance.


Designed by Freepik


Do You Have the Skills to Build a Chatbot?

No programming skills? No problem! You don’t need programming skills to design a chatbot. You need people skills. You need to understand what your customers want to talk about.

You need to be your company’s expert at Customer Service (CS) and Customer Experience (CX).

The challenge is to predict all the conversations your bot will ever have. That may sound difficult, but it isn’t impossible. You can do it by understanding the two types of conversations chatbots can handle.

I’d like to show you how these conversations work, so you can see how you would design either type of chatbot. So let’s look at two chatbot examples.


The Simplest Conversation is the Hardest to Program

The earliest rule-based chatbots used a spreadsheet to list all possible queries and responses. These chatbots are still useful for answering FAQs.

The next two pictures show a query-response design. The first picture shows the two-column spreadsheet and the second shows a possible conversation.

Human avatar designed by Freepik, robot avatar designed by Top Png


For the query-response conversation, you must predict everything the user might ask in the first column. The second column tells the chatbot how to respond. You can have multiple responses for each query, and the response might include functions that that chatbot will need to execute.

This conversation can be the hardest to program. It isn’t just that you have to predict every single query a user might ask or state. You also have to predict all the different ways the user will say it.

But there is some good news. You don’t always have to predict every variation of a query. Some rule-based chatbots actually use a small amount of artificial intelligence to interpret the user query. For example, a chatbot may realize that “Hi there” is the same as “Hello”, but it might not understand “Hiya”.


Trees Give Your Bot’s Conversation a Purpose

Today’s rule-based chatbots use a conversation-tree, which is a kind of flowchart. It has branch points where user input breaks the conversation into separate paths.

The next two pictures show a conversation-tree design. The first picture shows the flowchart and the second shows a possible conversation.

Human avatar designed by Freepik, robot avatar designed by Top Png


This type of conversation is driven by the chatbot.

The conversation-tree gently leads the user into the conversation. It gives you a way to show the user everything that the chatbot can do for them, similar to a menu in a computer program. This makes it easier for the user to understand the chatbot’s purpose.


But What’s Really Cool is a Bot that Understands Your User’s Intent

Now let’s have some fun!

AI/NLP chatbots understand what the user intends to do.

That isn’t just an expression. When you design an AI/NLP chatbot, you specify a type of function called an “intent.” Each intent includes parameters called “entities.”

The next pair of pictures illustrates this.

Human avatar designed by Freepik, robot avatar designed by Top Png


Let’s take a good look at the first box. Designing an intent means that you will fill-in the values for all of the items listed. This example shows the values you have filled in for the “Book A Hotel Room” intent.

The intent needs “Example Queries” that might be used to request this intent. In most cases, you would provide at least ten ways to request a hotel room.

When the user makes a request or asks a question, the chatbot compares that request to all of the example queries it has in memory.

So in the second picture, the chatbot searches through all of its intents, trying to find an example query that matches with “Can you get me a room?”. It doesn’t need a perfect match. Instead, it uses artificial intelligence algorithms to find the closest match.

(Internally, the chatbot decides that “Book A Hotel Room” is a match to within, perhaps, 75% certainty).

Pretty clever, eh?

The chatbot then realizes that it needs values for the $location, $date, and $room_size parameters. We call these parameters “entities”.

It asks for these entities using the prompts that you have provided.

When it has all the information it needs, the chatbot executes the API call for reserving a hotel room, and it returns either the Success Response or the Failed Response to the user.

By the way, with some chatbot design platforms, you can add variety by including multiple responses. The chatbot will select one of the responses at random.


Context Ties It All Together

Most users will want to do more than just one thing. In other words, they will request multiple intents. But users HATE to repeat themselves.

AI Chatbots can use something called “context” to borrow information that your user has already given for a previous intent.

Let’s see how this is done…

Human avatar designed by Freepik, robot avatar designed by Top Png


The chatbot begins by comparing the user’s first question to ALL of the example queries you have provided. It decides that this is the “Book A Hotel Room” intent.

The chatbot also realizes that “queen sized”, “Colorado Springs” and “next Friday” are appropriate values for the $room_size, $location, and $date entities. It has all the information it needs, so the chatbot immediately executes the API call and reserves a room.

But now the user makes a second request.

In this second request, the chatbot weaves through a long, complicated query, and determines that “need an SUV” is a match for the “Book A Car After A Hotel Room” intent. It also realizes that “SUV” is an appropriate value for the $car_type entity.

This intent also needs a date and a location. But in this case, it borrows the values from the “Book A Hotel” intent, using the $hotel context, referring to these entities as $hotel.location and $

You can specify which intents share their contexts with each other. You can also specify that certain intents are only available after other intents have been requested.

A more complete example would have two intents for booking a car: one that is only available if the hotel intent has NOT been called, and this intent that is only available if the hotel intent HAS been called.

Your user is happier, because she didn’t have to give the location and date twice.

Now let’s see what’s available for designing today’s chatbots!


What Chatbot Platforms are Available Today

This list is separated into three sections, showing Rule-Based chatbots that use conversation trees, AI/NLP-Based chatbots that use Intents and Contexts, and Chatbot examples that include an integrated CRM and/or CMS.

(All of the logo images in this section are the property of each product owner).


The Landbot conversation generator


Rule-Based Conversation-Tree Chatbot Builders


ManyChat is a great way to begin your exploration into chatbots. Its Facebook chatbot software is powerful, easy-to-learn, and include features from more expensive chatbots.

ManyChat integrates with Google Sheets, Shopify, ActiveCampaign, HubSpot, MailChimp, PayPal, Stripe and several other popular services.

Experimenting with ManyChat is a good way to learn how to build rule-based chatbots. Their website offers a free video tutorial on building bots. It is also featured in the Udemy course, “How to Automate Sales and Support Using Chatbots”, at

(The author does not receive any benefit if you take these classes).




Another good rule-based chatbot builder for beginners is LandBot.

LandBot features an easy-to-use drag-and-drop interface. You can experiment with it by opening a free account on their website.

LandBot’s conversations can include text, media, and buttons to answer multiple-choice questions.

LandBot chatbots integrate with Google Sheets, Stripe, MailChimp, Salesforce and other tools.




Drift chatbot is an advanced chatbot builder that uses conversational marketing to help businesses connect with customers that are ready to buy.

A Drift chatbot can ask visitors a few qualifying questions to determine if they are ready to buy, and then schedule a meeting with a sales representative.

Drift serves the entire range of business sizes ranging from a free version for individuals all the way to enterprise level companies.



MobileMonkey (Was previously named ChattyPeople),

MobileMonkey is a platform designed for marketing on Facebook messenger. It can broadcast announcements to your Facebook fans, make appointments, track purchases, and answer FAQs.

It collects lead information including name and email address, and includes analytic tools to measure customer responsiveness. MobileMonkey uses Zapier Integrations to share lead information with more than 1,000 applications.




Botsify is a paid subscription service that works with Facebook messenger.

Botsify uses a simple drag-and-drop platform builder to create conversation-tree based chatbots. Botsify learns by keeping a log of failed conversations which you can review and add to its conversation-tree.

Botsify can be integrated with several services including WordPress, Shopify, Slack, Alexa, Google Sheets, RSS feed, JSON APIs, and ZenDesk.



Designed by pressfoto at Freepik


AI/NLP-Based, Intent/Context Chatbot Builders

DialogFlow (Was previously named,

DialogFlow, which is Google’s chatbot platform, is a very good choice for an AI/NLP based chatbot. It offers both free and usage-based versions.

Google DialogFlow features intent and context based conversations. It integrates easily with many of the most popular messaging apps, including Facebook Messenger, Slack, Twitter, Twilio, Skype, and Telegram. DialogFlow supports 14 languages, and includes nice-to-have features like built-in small talk, which makes your chatbots friendlier and more fun to chat with.

Experimenting with DialogFlow is a good way to learn how to build AI/NLP chatbots. DialogFlow is featured in the Udemy course, “Hands On Chatbots with Google Dialogflow”, at

(The author does not receive any benefit if you take these classes).



Quick Facebook Chat App,

Quick Facebook Chat App is a free Facebook chatbot software platform that is especially useful for selling products via eCommerce. This chatbot comes from a company offering other e-commerce modules, including plug-ins for WordPress, Shopify, WooCommerce, and other eComm systems.

Here is an example showing its intelligent conversation capabilities. Notice how this conversation retains the “Buy Watch” intent from line to line, so the user can say “it” instead of “watch”:

User: “I have a question about the watch you are selling.”

Bot: “I’d be happy to help!”

User: “Do you have it available in gray?”

Bot: “Yes, let me check our stock for you.”

User: “Thanks!”




Authors note: I think you will enjoy chatting with the chatbot on ADA’s website. When you ask questions, it will answer and also suggest other questions you might like to ask.

ADA specializes in customer service and in customer experience. Their chatbot builder requires no programming and is designed for people who specialize in Customer Service.

In fact, ADA recommends that chatbots should be built and owned by your Customer Experience department, because Customer Representatives are the people who understand your customers best. They know what questions your customers will want to ask.



Amazon Lex,

If you’re familiar with virtual assistants like Amazon Alexa, then you have an idea of what Lex can do.

Amazon Lex uses Intent-Chaining, which suggests follow-up intents, as in this example conversation:


User: “Can you show my sales from last month?”

Lex: “Sure. Which regions?”

User: “Southern.”

Lex: “Here are your sales for the southern region.”

User: “Can you email me that report?”

Lex: “Sure. Would you also like to see the top five customers?”

At the end of the conversation, Lex not only sent the requested report, but also suggested related customer information, which comes from a different intent.




Rulai is a chatbot builder that is rooted in research. The company is a research organization that develops artificial intelligence techniques for use with chatbots.

The Rulai interaction design console is a no-code platform for business users. Rulai chatbots include pre-trained intents for banking, hospitality, and insurance.

Rulai is on the Forbes 2019 AI 50 list. The Rulai research team has published over 400 papers. You can find many of them on their website., isn’t a chatbot, but it is unique and useful enough to include in this list. is for computer programmers who would like to develop AI/NLP based apps. It serves as a front-end to the app, converting your user’s natural language into the functions that control your app.

For example, you could write a thermostat controller that accepts commands like, “Please set the temperature to 68”, or a scheduling app that accepts commands like, “Please remind me to feed the baby tomorrow at 7 AM.”


Designed by Freepik


Chatbot Examples Integrated with a CRM and/or a CMS

Zendesk Chatbot, and

Zendesk Answerbot,

Zendesk Chatbot and Zendesk Answerbot are two chatbot builders that are integrated with the Zendesk Sunshine CRM and the Zendesk Guide CMS. Chatbot is for conversations with customers, and Answerbot automatically answers emails.

The Sunshine CRM is a Customer Relationship Management tool that helps your customer service representatives track and interact with customers. It shows them customer contact information, purchase history, order status, and the conversation.

The Guide CMS is a Content Management System that stores articles that may help your customers. Unlike databases which require computer programming skills, Guide is easy to use by your customer service representatives.

Reps are encouraged to write articles and submit them to the Guide knowledgebase. Zendesk Chatbot and Zendesk Answerbot will read these articles and send any useful ones to the customer.

Chatbot, Answerbot, Sunshine, and Guide help you offer an outstanding customer experience and to become recognized as a knowledge-centered organization.




HubSpot provides a free CRM that includes many free options. Their free chatbot builder creates conversation-tree chatbots tightly integrated with the CRM. Hubspot is also a great choice if you’re looking for a free chatbot for your website.

Hubspot chatbots can book meetings, provide answers to common customer support questions, qualify leads, and more.

The CRM comes with plug-ins and options that can be purchased based on subscription. These options include a marketing hub, a sales hub, a service hub and a CMS knowledgebase. Option prices range from $50-$3200/month for the hubs, and $300/month for the CMS., is an enterprise level chatbot building and management platform. Their visual bot builder makes it easy to build bots that enable high traffic business to consumer communication.

The dashboard has a built in CRM, machine learning, conversation logging and real-time insights, for chatbots that continue to learn and grow.



Qualified, for Salesforce Pardot,

Qualified for Salesforce Pardot is a chatbot designed specifically for the Salesforce Pardot CRM. This chatbot qualifies website visitors by evaluating their questions and responses. If their questions and responses indicate they are ready to buy, the system calls them back via telephone and connects them to a sales representative.

Salesforce Pardot is a CRM with an open interface that allows it to connect with Qualified (and other) chatbots. The Qualified chatbot uses the Pardot CRM to show sales representatives a real-time view of your visitor’s journey through your website, including the conversation so far, the history that brought this user to your landing page, the pages the user viewed, and any actions the user has taken.

Qualified grades the visitor with a score that indicates where she or he is in the prospect – lead – customer funnel.





Designed by rawpixel at Freepik


Use Chatbots for the Best Customer Experience

The future is here!

Chatbots have improved, and we have learned more about what customers want. Customers love chatting, and they also like multiple-choice questions.

Sophisticated AI/NLP chatbots “chat” using artificial intelligence.

Simpler rule-based chatbots answer questions and ask multiple-choice questions using heavily-templated spreadsheets or conversation-trees.

Chatbots automate customer support functions, saving you time and money. They work with the messaging tools your customers are already using, like websites, Facebook Messenger, and Slack. Customers are more likely to use them, because they don’t have to install “yet another app” onto their smartphones.

Chatbots help your business succeed by building relationships with your customers. But they are only as good as their programming.

This doesn’t mean you have to hire a programmer. Rather, it means you should include sales and service reps on your chatbot design team. These are the people who understand your customers best, and they will help you design a chatbot that gives your customers the best experience possible.

Chatbots are all about providing the best customer experience possible.

So take a look at the websites for the chatbot software examples in this article. Look at the features they offer, look at other chatbot platforms, and find the chatbot that will work best for your customers!


Important This site makes use of cookies which may contain tracking information about visitors. By continuing to browse this site you agree to our use of cookies.