A complete guide to chatbot development. Tips for designing a chatbot interface: best practices for companies creating chatbots

Gartner estimates that by 2020, up to 85% of business communications with customers will be bots. Already now they are working with job seekers, carrying out and performing many other tasks. Plus, creating a chatbot that answers simple customer questions is cheaper and easier than hiring and training a customer service agent.

Here is a list of services that will help you make your own bot even without programming skills:

ChattyPeople

The service is suitable for small companies, although it also has functionality for large businesses. With ChattyPeople you can make a bot for Facebook messenger. It is enough to know how to use a computer mouse and keyboard. The bot created with this service can answer simple customer questions, but it can also be integrated with the Shopify e-commerce platform.

Flow XO

In the Flow XO service you can create more “advanced” bots for Slack, Messenger, Telegram and other platforms. It's not as easy to use as ChattyPeople, but you don't need to know how to program either. Flow XO has many integration options: it can be built into JIRA, Buffer and GitHub services, which will facilitate communication both with clients and within the team.

MEOKAY

This multi-platform service is suitable for those who already know a little programming, but have not yet gained enough experience. In MEOKAY, you can create many dialogue scenarios so that the bot can have a more lively conversation. Beginner developers will be helped in creating a bot by the technical support service.

Morph.ai

The strength of the Morph.ai service is speed. You can create a bot in minutes. Another feature of the service is natural language processing and the use of deep learning. They help the bot have a much more lively conversation, not limited to primitive answers. This is not the most budget solution- the optimal set of functionality without restrictions costs $299 per month. The bot can be integrated with Twitter, Slack, LINE, Facebook Messenger, Shopify, Zendesk, and even the Salesforce CRM system.

Botsify

Using this service, you can create a free bot with simple functionality for communicating with a limited number of users (up to hundreds) per month and integrate it with Facebook Messenger. Bots created using this service can work with videos, images, audio and other file formats.

Botkit

Botkit makes it easy to design and create bots, but it only supports Slack. The service is suitable as the first platform for learning and experimenting with chatbots. On Botkit you can test the technology before purchasing a subscription to one of the paid tools.

Chatfuel

According to the creators of the Chatfuel service, you can use it to launch a bot on Facebook Messenger in just seven minutes. No programming skills are needed - everything is done in a simple way user interface. The bot can be integrated with other platforms, including JSON and YouTube. Chatfuel services are free.

Recast

Recast is suitable for those who want to focus on one bot rather than making many bots on different platforms. The list of possible integrations includes everything standard applications, as well as Kik and Skype.

Cortana Intelligence

Microsoft's Cortana Intelligence platform allows you to create bots to perform a variety of tasks. This service is not for beginners. To use it, you need knowledge in the field of machine learning and artificial intelligence. But the level of abilities of the bot created using Cortana Intelligence will impress anyone.

Read us in

To date, artificial intelligence has found its way into many sectors and it works wonders. Development software to automate various business processes is becoming the hottest trend. From artificial robots to modern Node Bots, artificial intelligence is the future of technology. Chatbots are also one of these innovations based on artificial intelligence.

Messaging apps are rapidly growing in popularity, and business owners can benefit greatly from this trend. As well as social media, messaging apps provide many advanced marketing and user engagement capabilities. Business owners can sell their products and also offer services to users through popular apps, such as Slack, Facebook Messenger, Hipchat, MSN and many others. Chatbots are basically applications built on artificial intelligence that can be integrated into messaging platforms, and can also be programmed to perform a range of conversational and self-promotional tasks. Additionally, you can configure them to function in a certain way based on the responses they receive and store information.

Today, chatbots are used in various fields such as customer service, e-commerce and marketing. They can be programmed to think and behave like a human and are capable of performing various tasks independently. They can be useful in notifying online users about new offers, providing information about products and services, collecting information and feedback from clients, as well as monitor consumer preferences. So, if you are a business owner or a developer who is interested in creating bots, then we offer 9 popular platforms for creating your own chatbots.

Facebook Messenger Platform is a major player in the world of bot development. It offers tools for creating chatbots and live messaging. This platform is characterized by the presence of the Wit.ai tool, which allows the integration of conversational interface and bots into applications. It also offers built-in templates and plugins that allow you to integrate personalized text messages, photos, calls to action and more. In addition, the Facebook Messenger Platform allows you to benefit from an ecosystem of diverse development tools and resources.

Chatfuel is one of the most popular chatbot building tools for Facebook Messenger. The tool has powerful artificial intelligence capabilities that allow you to create chatbots without any programming knowledge. Chatfuel is used by more than 7 million users worldwide. Chatfuel is used by famous companies such as Uber, National Geographic and Tech Crunch. More than 230,000 chatbots have been created using Chatfuel for various industries - sports, hospitality, publishing and e-commerce.

Another useful tool for creating and integrating chatbots into messaging applications. It supports multiple platforms such as Android and iOS and also integrates with Stripe, which allows you to proactively engage customers online, providing an enhanced online shopping experience, accepting payments and hassle-free order fulfillment. The platform allows you to use a wide range of other messaging platforms such as Twilio, WeChat, Shopify and Telegram. Smooch integrates with Facebook Messenger, using most of its features - images, gifs, emojis and buttons. It also syncs with other business tools - Front, Slack, Hipchat and Zendesk.

Botsify is another bot creation tool focused on Facebook Messenger. Botsify offers an easy-to-use interface that allows business owners to create bots without having to write any code. The platform provides the Facebook Messenger API, which provides capabilities that allow developers to set up personalized replies, message scheduling, integration with WordPress and Medium, and pull data from analytics tools. Botsify supports multimedia content, including videos, images, and audio files.

A modern platform for creating and hosting bot projects for Facebook Messenger and Slack. Source projects are hosted on GitHub and after running the bots as Docker containers, they can be written in any programming language - Ruby, Python, Go and JavaScript. Beep Boop offers some great features such as creating notifications, hosting public bots and auto-grading support, as well as continuous integration via GitHub.

Pandorabots is a set of web-based tools and services that allow developers to create and host chatbots. There is also a tool called AIaaS (Artificial Intelligence as a Service), which is a full-fledged API for accessing the bot hosting platform and integrating the chatbot into web and mobile applications. Pandorabots allows you to create artificial intelligence based chatbots for customer service, IoT interfaces, advertising, games and much more. The Pandorabots community consists of approximately 225 thousand developers and to date more than 285 thousand chatbots have been created using the tool.

Microsoft offers its own platform for developing bots that provide advanced conversation creation. Microsoft Bot Framework allows you to create bots for Skype, Facebook Messenger, Slack, Office 365, E-mail and many others. The tool offers a bot builder with SDK for Node.js and .Net. Developers can also visit the Bot Directory, which has an impressive collection of ready-made chatbots.

This is an excellent bot creation platform for several popular messaging platforms, providing an open API for Viber, Telegram, Slack, Facebook Messenger, Kik, Slack and Telegram. Dexter offers over 250 integrations for analytics, hosting, and other enterprise solutions. It also provides good documentation and technical help for installing the SDK, creating custom bots, mailing services, and more.

In fact, there is nothing complicated here. The main thing is to think through the actions (commands) for the bot.

For mine, I made a small menu of 3 items - where you can pick up a useful gift, get access to my free course through affiliate programs and contact me in PM.

This menu looks like this:

I created the bot itself through the Manybot platform. This can be done for free, but in the dialogue with your bot, users will see a link to the developer.

What features does the bot have?

1) Creating custom commands and responses to them.

2) Creating a menu of commands. This is convenient because users do not need to enter commands, but can use a ready-made menu for navigation.

3) The ability to send messages to subscribers (including deferred ones), view the number of subscribers in the bot.

4) Several interface languages, including Russian.

5) You can set up auto-posting to a bot from an RSS blog or website, as well as from a YouTube channel, VK, Twitter.

How to create your own Telegram bot?

I recorded a step-by-step video instruction - an example of creating a simple bot with a menu:

You can see my bot in action Here. In the dialogue with him, click “Start”, he will tell you what to do next.

So, to start creating your Telegram bot, go to the website manybot.io, there click on the “Create a bot” button. You will be redirected to the initial instructions.

You need to go to Manybot in Telegram and click “Start” in the chat with the bot to receive further instructions.

You will see a description of the bot's capabilities in English, and you will also be asked to select a language.

After choosing the Russian language, you can start creating a bot; all instructions will be given in a dialogue (chat).

First of all, enter the command /addbot or simply click on it in the message received from the bot.

The next step is to type the /newbot command in the dialogue with BotFather and follow further prompts. Namely:

1) In English, we are asked to indicate the name (name) of our future bot. I named it ViktoriaHelps and sent this name in the BotFather dialogue.

3) After that, I received a message congratulating me that my bot was created, a link to it in Telegram, as well as a token that I need to copy.

In response, I receive a message that I need to add a description (welcome message) for my bot. The user will see this description in the chat with the bot when he switches to it.

After sending Manybot a description, I see a message that my bot has been successfully created. For it to appear in your contacts on the left, follow the link to it and click “Start” in the chat.

By the way, the description of the bot can be edited at any time. This is done through the /setdescription command in the dialogue with Manybot.

Setting up commands and menus in the bot

In response, we receive a message with a description of our bot (what we indicated in the previous step). Your bot’s subscribers will also see it.

The message below will display a menu for managing your bot (only you, as the bot owner, can see it). If this message does not appear, you can call it yourself by entering the /help command into the chat or selecting the “Help” section in the bot menu.

This message displays the commands you can use to control the bot. For example, to send a new message to subscribers, use the /newpost command. Just click on it, or send it to the bot in dialogue and follow further prompts. It's the same with other teams.

As an example, let's create a few of our commands in the bot and then embed them in the menu so that new subscribers can interact with your bot.

To do this, in the menu under the message input field, click on the “Custom commands” item.

Select “Create a team”.

We indicate the name of our team - it must be in Latin and without spaces. The title may contain numbers and underscores (to separate words). Before the command we use the slash (slash) /.

Enter the command in the message field and press Enter.

The next step is to set the bot's response to the command. That is, the subscriber enters a command into the chat and receives a prepared response to it. That's how it works.

We receive a message that the command has been successfully created, we see it in the bot menu below. We create other teams in the same way. In particular, a command, when clicked on, the user will receive a message with a link to download the gift.

By the way, when creating a team, we can add a question - a regular, detailed one, with answer options. To add a question, follow the bot's prompts.

If you want to edit or delete a command, click on it in the menu.

A new menu will open from which you can select the desired action.

For example, choose editing. Here the steps will be as follows:

1) First we need to delete the previous response to the command and confirm the action.

2) Then click “Add messages to team”.

3) Enter a new answer, send it to the bot, save it.

4) We receive a message that the command has been successfully edited.

To return to the previous menu, click “Back”.

How to add a command to the bot menu?

Now I'll show you how to add a command to the menu. This menu is made for the convenience of the user in dialogue with the bot, so that he does not have to enter commands manually.

Click on the “Customize main menu” button.

As well as a return button to the previous menu.

How to remove a command from a menu?

If you want to remove a menu item or change its name, just click on it.

You will see another menu with available actions. There are many settings here, in particular, you can change the location of items and create multi-level menus. But that’s not about that now...

Again, by scrolling down the menu, you will see the “Rename” and “Remove item from menu” buttons. Select the desired action and follow the bot's prompts.

For example, I'll simply remove the "Go to Blog" option from the menu.

Once the benefits and relevance of chatbots become clear, a logical question arises: where do I start if I want to build a profitable chatbot development business? Let's figure it out.

1. Define target audience. This will help you concentrate on the group of the most promising clients and not waste time communicating with those who are unlikely to order such a service. But how to identify those who want to buy?
   First, you need to clearly understand who will benefit from the chatbot. The need for a chatbot at a wholesaler or steel mill is not obvious, so selling the service to such potential customers will be difficult. But a restaurant, media outlet, concert venue, taxi, airline, car rental or travel agency will definitely find a chatbot useful. It will be much easier to close a deal in these niches.
   Secondly, it’s worth considering who can afford the amount for which you want to sell the chatbot. For example, if you offer your service to a private teacher who earns 30 thousand rubles a month, he is unlikely to be willing to pay that much for a chatbot. And for a company with an income of 150-200 thousand rubles per month, developing a chatbot for that kind of money will be feasible and profitable. Look around: there are so many micro-businesses around us with such a turnover!
2. Understand the needs of customers and their clients. Think about what business problems of the client this tool will help solve? For example, the ability to automatically book tables in a restaurant will relieve the administrator, and the ability to select and buy goods from an online store using a chatbot will save the business owner from the need to hire several consultants.
   The next question that needs to be answered is what user needs will the chatbot help solve? Let’s say a person needs to buy plane tickets, but he only has a smartphone at his disposal. Imagine how much more convenient it will be to talk to a bot than to try to find and book a suitable flight through a website.
3. Create a demo of a chatbot for a specific niche. When you have a visual version of the bot at your fingertips, communication with the client will become more effective. You will be able to show customers all the features and benefits of this tool, and as a result, you will be more likely to close the deal.
4. Show potential clients the power of chatbots. Explaining how a chatbot works at a meeting is very simple. Show the process on your smartphone, and most of your interlocutor’s doubts will disappear by themselves. Customers often have questions about the ability to implement a feature that is not in the demo version. Careful familiarization with the functionality of chatbots will give you freedom in such situations. In any case, it is worth remembering that the plugin system allows you to connect a maximum of solutions and make chatbots unique.
5. Make a deal. If the client is satisfied with everything, sign a service agreement. If the customer asks for time to think, agree on control point and don't forget to remind yourself. In the meantime, continue to look for clients in this niche, go to meetings and negotiate with representatives of other companies. This way, you can kill two birds with one stone, or even more, by selling the development of a chatbot to several customers in one area at once. This means that from the first sales you will recoup the costs invested in the business and make a profit.

The use of smart assistants is relevant in a variety of business areas where there is a need to automate work processes. Bots can relieve the support service by recognizing common problems and offering hints (in other words, the “ FAQ" in real time).

Let's consider another example - you are the owner of a food delivery business, a chatbot may well take on the function of a waiter to receive (in chat) and confirm the order (by phone). In fact, there are a huge number of examples; a bot can provide up-to-date information about the status of a parcel (order), advise on company products, book tickets and search for information and pictures on the Internet.

Bots are able not only to solve problems, but also to aggregate information about interactions with clients, thereby “learning.” This simplifies subsequent iterations - for example, when contacting the bot again, the client can tell the bot something like: “like last time” (when ordering food in a restaurant), the bot will bring up information about the previous order and send it for processing - fantasy? No, these are very real cases in coffee shops and restaurants in the USA.

In the near future, this will reduce personnel costs - sounds scary doesn't it?

• Framework
• Hosting

Supported platforms:

• FB Messenger
• Slack
• Telegram

XO flow- a powerful cross-platform solution for creating bots without programming skills. The intuitive interface of the editor allows you to customize the logic of the future bot in a matter of minutes. Key advantages also include a huge number of built-in integrations (more than 90) with all kinds of external services.

The platform offers several tariff plans(free is not provided), costing from $19/month. Tariffs include a limited number of iterations (interactions with bots). It is possible to customize the plan by adding new bots and the number of interactions.

• Framework
• Hosting

Supported platforms:

• FB Messenger
• Slack
• Telegram
• Skype

Recast.ai- a platform based on its own natural language recognition technology (Natural language processing), according to the developers, the core of the platform easily recognizes user input queries, highlighting key phrases:

The service has excellent multi-platform functionality: Facebook messenger, Slack, KIK, Weechat. Bots can be embedded into the web interface and process incoming mail. The user-friendly interface allows you to create your own bot in 8 minutes. (information from the official manual).

The platform offers a wide range of development tools (SDK) for Python, NodeJS, PHP, Android, as well as a functional API. It is worth noting that there is a large community of experienced developers who will be happy to suggest a solution to beginners.

Pricing - from a free tariff for developers, which includes an unlimited number of bots (which must be posted publicly), and 3 basic plan to create private assistants. Tariffs also limit requests; it is also possible to place a bot outside the site or on it (in this case there is an SLA guarantee).

Supported platforms:

• FB Messenger
• Telegram

ChatFuel- one of the first public online services for creating chatbots. At the time of writing the review, it supports several platforms - Facebook and Telegram, and soon the developers promise to add Viber, Slack, KiK and WhatsApp.

The platform offers a convenient and simple interface for designing and launching a bot in a product. According to the developers, the built-in natural speech recognition (NLP) tool “understands” a wide range of queries and is capable of responding to them situationally (passing a query parsed by keywords for processing to a user-specified algorithm).

ChatFuel already uses several well-known brands to interact with customers - National Geographic, TechCrunch, UBER.

The service offers a free plan (up to 100,000 requests), which significantly exceeds the offers of competitors (on average, 2 times the number of iterations in the free plan).

In our opinion, now Chatfuel is the real optimal solution, at least in terms of functionality/cost ratio.

API.ai- not exactly a platform for creating chatbots, it is an American startup founded in 2010 by immigrants from Russia - Ilya Gelfenbein, Pavel Sirotin and Artem Goncharuk. At the beginning of the journey, the team was developing a voice interface for managing applications (similar to Siri).

On September 19, 2016, the platform was acquired by Google. According to vice president of development Scott Huffman, the acquisition of Api.ai is one of the planned investments in machine learning technologies.

Today, the capabilities of Api.Ai are widely used by services for creating chatbots as an external NLP tool. To properly process incoming requests, the bot needs to convert the input text into a language it understands - highlighting key commands. This is exactly what happens “under the hood” of the service.

The “parsed” request is returned to the bot via the external programming interface, which then responds to keywords, dates and actions based on the given operating logic. The logic of work, in turn, is configured in the interfaces of services for creating chatbots.