Chatbots can talk to your customers for you. In this lies their ability to handle various aspects of customer relations, substituting a number of employees with a single bot.

In this article, we look at the essence of chatbots: how and where they work, which industries can benefit from them, and where they’ve already been successful.

Finding the best platform for your chatbot

There are two types of chatbots – those built into messengers (Slack, Telegram, Discord, Kik, etc.) and standalone applications. We advise building a chatbot in a messenger first because there are a lot of people using them already, so your service will be able to receive the recognition it deserves. Just look at these statistics from April 2018 showing the number of monthly messenger users.


In the top positions are, of course, WhatsApp and Facebook Messenger, with 1.5 and 1.3 billion monthly users respectively. In third position is the Chinese messenger WeChat, which has an average of 1 billion users a month. QQ Mobile, Skype, Viber, Snapchat, LINE, and Telegram constitute the rest of the list, each having fewer than 1 billion monthly users – yet still substantial audiences.

To choose the perfect platform for your bot, research the most popular messengers in your area. Alternatively, you could go for a multi-platform chatbot, supporting a number of messenger applications at once.

Now let’s understand how chatbots work

There are two main ways in which a chatbot can be built – with and without machine learning.

Chatbots that learn

Chatbots that use machine learning are chatbots that can learn and become better over time. The technology here isn’t half as complicated as you might think, even though it does appear quite sci-fi at first glance.

What’s particular about machine learning-powered chatbots is that they can understand natural speech. Unlike their scripted counterparts (which we’re going to discuss in just a bit), machine-intelligent chatbots can understand questions and commands the way real people phrase them, as opposed to only understanding a set of predefined commands.

This is beneficial if you’re looking to offer a human-like experience.

Let’s look at Mitsuku, for example:

Source: Mitsuku on KiK

Mitsuku is a three-time winner of the Loebner Prize (the chatbot equivalent of the Turing test). Created by Steve Worswick from Artificial Intelligence Markup Language (AIML), the bot is available on Facebook Messenger, KiK, Telegram, Skype, and Twitch, featuring also an original web version and an Android application (currently in beta).

Mitsuku isn’t an assistant who performs tasks to make your life more comfortable. She’s a friend, a partner in conversation, and an interesting way to spend your time. Mitsuku can tell jokes, ask questions about religion and philosophy, and answer whatever questions you might have for her (if they aren’t “Can you do this for me?”).

Chatbots that don’t learn

Scripted chatbots tend to be a business’s go-to choice. They specialize in responding to specific commands and answering specifically phrased questions.

A scripted chatbot has a set of questions it can respond to with a corresponding set of answers. This means each conversation can only follow a number of defined paths. It’s often the case that users don’t even type anything in, instead selecting from a list of questions and commands that the bot understands.

The H&M bot on KiK is a great example of a scripted chatbot. When you first log in, it asks a series of questions to understand your style preferences. Then it offers three options:

  • Build an outfit by selecting pieces of clothing and accessories from H&M’s catalog to create a look. The bot then returns that look as an image.
  • Vote on outfits created by other bot users to choose the best.
  • Search outfits by accessory or item of clothing (e.g. “black shirt”) to view user-generated outfits that include it.
Source: H&M bot on KiK

The chatbot is limited in what it can say and do. But if you think about it, users won’t come to the H&M bot for life advice or to discuss their favorite novel. They want to look at clothes, check out new items, put together some outfits, and then come to the store in person to purchase the stuff they’ve already seen and liked.

The H&M bot does what it’s supposed to: it involves people into the H&M brand, creates an element of interactive communication, and effectively upsells customers.

The benefits of chatbots

Depending on what task you want a chatbot to complete, you’ll find different results. There are, however, some universal benefits that a chatbot can bring to any business regardless of its primary focus.

E-commerce and online marketing

There are many ways in which the e-commerce industry has benefited from chatbot technology. When your goal is to sell products and services, the ability to communicate directly with customers is crucial.

We’ve already mentioned the H&M chatbot, but you may have heard about other examples from Sephora, eBay, 1-800-Flowers, and other companies. These chatbots have managed to substantially increase company revenues over a short period of time. Not only that, but there are many other ways that chatbots can help e-commerce businesses:

  • Substituting for emails – Instead of composing hundreds of cold emails, you can simply have a chatbot talk with your customers.
  • Managing sales funnels – Through chatting, bots can determine which customers belong to which sales funnels. This helps your business choose the best approaches to convert them.
  • Adding interactivity – Just like in the H&M example, bots can offer an element of interaction to the products and services you’re selling. This can help users feel as if they already know your selection, making them more eager to buy things they’ve already seen and liked.
  • Building customer relationships on a more personal level – It’s possible to add some personality to chatbots. This can turn the process of chatting with them into a real, almost human conversation, potentially making customers enjoy your brand more.
  • Solving the abandoned cart issue – Customers often add products to the cart and never end up buying them. Before chatbots, marketers would send emails to remind users about their carts, but the process has changed since chatbot technology has been introduced. Now it’s enough for a bot to text your customers with a “Hey,you cart’sstill waiting for you!” as a friendly reminder.

Travel, hospitality, and tourism

Chatbots can do a great deal for the travel, hospitality and tourism industries. They offer 24/7 access to data, allowing customers to book trips and rooms instantly and on the go. And using chatbots is cheaper for businesses, too! Employees don’t need to answer calls and repeat the same stuff over and over; customers can just text their requirements to a chatbot.

Chatbots are already working for Marriott, KLM Royal Dutch Airlines, Wynn Las Vegas, and Waylo. With chatbot technology, companies can benefit in many ways:

  • Engage audiences – Once customers ask a chatbot about something, it can analyze what they’ve written to produce personalized content. When asking for a plane ticket to Los Angeles, for instance, users can receive information about room availability in partnering hotels, learn about the best nearby restaurants, and so on. With the ability to learn so many useful things at once, customers are likely to continue returning to the chatbot again and again.
  • Anticipate user needs – After learning enough about a particular customer, an intelligent chatbot can offer services based on their previous requests. If a user has been travelling to Chicago once a month for a year, the bot might offer them information on room availability a couple days in advance of their usual travel date.
  • Give recommendations on nearby locations – Let your chatbot know about cafes and restaurants near your hotel or airport, and users will very much enjoy asking it where to have coffee or brunch. For airports, users could ask about services and facilities.
  • Offer automated services – A bot inside a hotel could let users order meals or room service without having to call anyone.
  • 24/7 customer service – Whenever users have a question or concern about a hotel or transportation, a customer service chatbot can answer them. If the technology is unable to present relevant information, a customer can simply be redirected to a real person.

Case study: Social Media Platform for Sharing Your Favorite Travel Experiences


When it comes to healthcare, nothing substitutes a real professional. However, there are some cases in which chatbot technology could be a real lifesaver by promoting healthy living and helping patients figure out a number of important questions. Chatbots could guide users through emergencies, giving them step-by-step CPR instructions or explaining how to help someone with diabetes, and perform many other tasks:

  • Support self-care and self-monitoring – A chatbot doesn’t necessarily have to answer questions and share information. It can help patients track their health and fitness. For instance, a patient could take their body measurements (blood pressure, weight, pulse, blood sugar level, etc.), give them to a chatbot, and then see a comprehensive analysis of their data over time. If some measurement is far off, the chatbot could show concern and offer to schedule a doctor’s appointment.

    What chatbots can also be great for is setting reminders to take medication, add health data, exercise, drink water, and so on.
  • Offer reliable medical information – Googling symptoms of a hypothetical disease is something of a joke now. However, if a chatbot is connected to a number of reliable medical databases, it could be able to give patients relevant medical advice and offer ways of understanding their conditions.
  • Get important information from new patients – You know those long questionnaires that you need to fill out whenever you come to a new doctor’s office? Answering these questions would be many times simpler with a chatbot that could request, record, and then analyze important patient information.

Read more: How to Build an Effective Medical Mobile App

  • Perform automated appointment follow-ups – It’s important to check in with a patient sometime after an appointment. A chatbot can do that by asking people about how they feel and figuring out whether they need another appointment.

    This also works well for postoperative care. A chatbot could then serve as a pocket nurse, reminding about medications, explaining some things patients might be experiencing, and scheduling doctor’s appointments if necessary.
  • Show electronic healthcare records. Being more of an internal hospital tool, this feature could allow doctors to quickly receive information about patients by simply typing in a patient ID number.

Case study: Developing a Mobile Healthcare App for Doctor-Patient Consultations

On-demand services

Imagine describing your perfect pizza in a text message and then having it appear on your doorstep. This is exactly the sort of feature a chatbot can provide for on-demand services. Just look at how Pizza Hut did it with their Facebook Messenger and Twitter chatbot.

Source: Pizza Hut bot on Facebook Messenger

Pizza Hut’s technology allows customers to place orders with a single touch of a finger. It can answer frequently asked questions and present information about the latest promotions. Because the bot is so easy to access and doesn’t need to be installed separately, users don’t feel any pressure talking to it, which results in higher retention rates and easier ordering.

Read more: How to Build a Food Delivery App for a Restaurant

Banking and finance

A great number of global banks have already integrated chatbots with their services: American Express, PayPal, Bank of America, Mastercard, Visa, and others. You might have used some of their financial assistants yourself – such as Eno, HiCharlie, or Trim. Among the features that financial facilities could implement with chatbots are:

  • Account alerts and notifications – A chatbot could let you know whenever unusual activity happens on your account to determine whether it’s you. It could also remind about fees, upcoming charges, and so on.
  • Tips and suggestions on financial management – A chatbot could help users figure out the most efficient ways of spending money based on their past expenses. For instance, it could monitor subscriptions and then point out ones that they no longer use so that they can stop paying for them.
  • Customer service – Chatbots could answer customers’ burning questions around the clock, always being friendly and informative no matter what they’re asked or when they’re asked it.
  • Help with enterprise resource management – Chatbots could also help automate repetitive internal bank processes.

Read more: How To Maximize the Success of Business Process Automation

Customer service

Last but not least, we come to customer service – the area where chatbots have probably done the most good. It’s just so convenient to have a digital assistant that can answer customer questions in as much detail as necessary. Chatbot-based customer service can be applied to any industry, accomplishing two main tasks:

  • Automating frequently asked questions – It’s often the case that customers want answers to the same questions but don’t want to read the FAQ page. Answering common questions isn’t an efficient way for your employees to spend their time, however, and having a chatbot answer them could be a much better solution.
  • Differentiating between questions the chatbot can answer and questions that should be referred to a real person – However smart they might be, chatbots can’t always fully substitute for real people. A chatbot could look at the question it’s being asked, grade it based on its competence to answer, and then refer the more complex issues to a human assistant.

Develop a chatbot with SteelKiwi

Do you feel inspired by the numerous benefits of chatbot technology? If so, we strongly encourage you to contact our sales representatives to start discussing your product today.

You’re also welcome to look through our SteelKiwi projects page to learn about the awesome products we’ve created so far.