An overview of chatbots: What are chatbots? How do they contribute to businesses?

Let’s discuss the very basics of the huge trend of chatbots that affect almost every kind of business

Developments in the realm of artificial intelligence and the rise of messaging have unveiled a new phase in customer interaction. Now, chatbots are on the scene as the new interface that companies communicate with their customers. They are widespread, especially in the service sector – from finance to airlines, e-commerce to health. Banking chatbots, e-commerce chatbots, HR chatbots, food order chatbots are the most common ones, but of course, it is possible to extend the list.

This blog aims to discuss the broader topic of chatbots that has gained a significant momentum in 2016 when Facebook announced that it has opened its Messenger platform to chatbots. Let’s start our discussion that focuses on answering the very basic questions of the issue by understanding why we started to hear so much about chatbots recently.

What are chatbots? Why now?

In order to understand the rise of the chatbots, we need to explain two basic developments: the first is the recent developments in the realm of artificial intelligence, and the second is the rise of messaging as a type of communication.

  • The recent developments in the realm of artificial intelligence

Over the last few years, we have witnessed significant advances in artificial intelligence, especially in natural language processing and deep learning. The specific subfield of artificial intelligence that uses these two technologies is called “conversational AI”. Today, thanks to the developments in this discipline, we see softwares around, that can speak or write as a real person. Technology has become more human, and it is now much more involved in our daility acitivities in a more natural, intense and direct manner.

We can assert that the digital world is shaped by the “web revolution” in 1990s and by the “mobile revolution” in 2000s. We can again assert that the concept of conversational AI is such a powerful trend that is able to reshape and redesign our present and even future world. In this sense, we are facing a paradigm shift.

  • The rise of messaging

The rise of messaging in recent years is very important for almost every sector. Messaging was first used for human-to-human communication and then gradually it has become a tool for customers to interact with the enterprises. In 2015, we witnessed that 4 major messaging platforms surpassed the 4 major social media platforms in terms of the number of monthly active users. According to a research by Facebook, being able to messaging with an institution increases the credibility of that institution for the customers. The ones who agree on this statement reached 76% in Brazil, 79% in India, 66% in the UK and 69% in the US. The messaging platforms are considered as the main digital hubs of the near future where many different business lines operate and talk to their customers.

an overview of chatbots 1

What is a chatbot?

Researchers predict that chatbots that has gained a very visible rise thanks to these two trends mentioned above will transform the businesses very quickly. Let’s see a few examples of these predictions:

  • Oracle – 80% of the brands will use chatbots by 2020
  • IDC – Artificial intelligence plays a key role in the digital transformation agendas of the companies. According to the report, 40% of the companies’ digital transformation initiatives will be on artificial intelligence at the end of this year.
  • FB Messenger announced that it would built a chatbot platform in 2016, at the end of the year it hosted 34,000 chatbots. Today, we know that 300,000 active chatboats serve on this platform.

Chatbots are softwares to which people communicate digitally via chatting to get a specific information or to perform a specific action. Chatbots can be used for various purposes, such as ordering a pizza, buying a movie ticket, checking-in for a flight, and checking the account balance.

Graphical Interfaces vs. Conversational Interfaces

Chatbots are different from the traditional interfaces of the digital world. With the introduction of the chatbots, it is possible to talk about a new paradigm in terms of user interfaces. Websites and mobile applications use graphical interfaces (GUI), while chatbots use conversational interfaces (CUI). Chatbots can be integrated into the popular messaging platforms such as WhatsApp, Facebook Messenger or they can serve stand alone on a company’s own web site or mobile application.

You can reach the information you are looking for instantly via chatbots, instead of searching it among the menus of a website; or they can perform an action for you that you are used to perform through mobile apps, live agents or call centers. The main difference here is that the user does what he/she wants to do through a conversation – a direct communication. Today, there are many different types of chatbots that provide information, make transactions and offer advice to the user.

Chatbots, vary as a method of communication. For example, they can be reactive in a dialogue – they can provide the answer only when they are asked to; or they can be proactive as well. Proactive chabots can initiate a conversation with the customer when they are triggered by various means, for instance, when the customer’s account balance falls below a specified amount, when a suspicious situation is detected in credit card transactions, or when a product is discounted to a specific price. If this first message by the chatbot is relevant to the user, a dialogue is initiated and the desired actions are taken via the conversation.

How long have the chatbots been speaking?

The transformation of the curiosity about artificial intelligence into the scientific researches goes back to the 1950s. Alan Turing’s study “Can machines think?” (1951) is considered as one of the milestones in this discipline. In addition, ELIZA, developed in 1966 by Joseph Weizenbaum, an MIT professor, is the first generally accepted chatbot. ELIZA was designed as a psychotherapist and was intended to create humanized dialogues. However, since natural language processing and machine learning technologies were not yet developed, it was able to respond only by word matching.

In the 1990s, Turing Test was developed based on the work of Alan Turing and it was re-launched every year with a cash prize. This test is based on the user’s ability to identify if the other part is a human or a computer software. If the user can identify the other part, the test fails; if he/she cannot identify, the test is successful.

In 2016, Google announced that it has adopted an artificial intelligence-driven strategy; in addition, Facebook Messenger announced that it has opened its platform to chatbots, the result was chatbots current situation as a top-trend in the digital world. Today, as tech giants like Google, Apple, Facebook, Amazon, Microsoft, Alibaba -GAFAMA- are highly focused on the issue, especially on NLP (natural language processing) and deep learning technologies, we see chatbots and voice assistants that are based on advanced artificial intelligence technologies.

What kind of use cases prevail?

Chatbots, today, with their mostly-know functionality, save time and energy by automating the basic tasks of customer services departments. Gartner predicts that more than 85% of customer interaction will be performed without human involvement by 2020.

In a study conducted by Drift, when the participants were asked what kind of a purpose would they use a chatbot for, we see the answers like – to get a quick answer to their question, to solve a complaint, to get a detailed explanation on a topic, to reach a live agent quickly, to make a reservation. We can see the most popular use cases as – banking chatbots, ecommerce chatbots, booking chatbots, health chatboats, IT / HR support chatbots and call center chatboats.

an overview of chatbots 2

How do chatbots contribute to the businesses?

According to a study by Spiceworks, 40% of large enterprises with more than 500 employees say that they will launch one or more chatbots in 12 months. AI-based chatbots gained a high priority for companies, and of course, the expectations increased accordingly. Now let’s have a look at how chatbots contribute to the businesses. Chatbots support companies to achieve their business goals in three ways:

  1. They support their efforts for operational efficiency and cost reduction by automating basic FAQ functions and transactions
  2. They improve customer experience by responding instantly customer queries and simple transaction requests. An enhanced customer experience increases both new customer acquisition and loyalty of the existing customers
  3. They, when well-integrated into the company’s CRM / data analysis systems,  increase sales and revenue by providing the right products and services at the right time to the customers

Now let’s discuss each area of ​​benefit in more detail.

  • Reducing costs and sustaining operational efficiency:

Recently, companies focused on automating simple and repetitive tasks that are previously performed by humans. This not only reduces costs, but also enables the employees to concentrate on high quality and tasks that are more sophisticated and require human touch. Thus, employee satisfaction and loyalty increases. The fact that humans still perform very basic tasks that an algorithm can easily perform as well, is an obstacle to the development of the employees and diminishes the quality of the customer experience. For this reason, it is very important to use chatbots for certain tasks in areas such as customer service and call centers.

Certain features of chatbots provide this efficiency:

  1. Chatbots are scalable tools – businesses can serve millions of customers with a single chatbot, and as their customer base grows, the cost does not increase
  2. Chatbots are very good tools for preemptive action. A chatbot can prevent fraud that might require huge costs for the company, just because it can interact with the customer instantly at the very right moment and perform the preemptive action
  3. Chatbots are multilingual tools – it is possible to provide services in multiple languages ​​with a chatbot technology

Thanks to these features, chatbots provide considerable saving opportunities for the companies. According to a research by Juniper, in 2022 the chatbots will save $ 0.7 and 4 minutes for each call, which means a total savings of $ 8 billion.

  • Enhancing the customer experience:

Secondly, chatbots offer customers an opportunity of direct, simple and natural interaction. Today, the most important competitive advantage for companies is to provide a good customer experience. In many sectors, competition is around the concept of customer experience. It is relevant that the existing digital channels are insufficient to meet the rising expectations of the consumers, because using these channels does not provide a simple enough, fast and natural experience, and they are often weak in terms of technology and design.

an overview of chatbots 3

Chatbots have crucial advantages compared to the call centers, websites and mobile applications. The main features of a chatbot experience that differentiate it from other channels can be summarized as follows:

  1. Simple and direct: Chatbots can quickly answer customer queries and perform the transactions without making them wait in a queue of the call center for minutes, navigate through the menus or install a separate mobile application for each task. For example, just by asking the right questions to the customer on an ecommerce portal, it is possible to direct him/her to the right product among hundreds of options. For the customer it is possible to reach the most appropriate flight without switching between hundreds of different possibilities. In summary, the customer can do all of these in a seamless way without using the menus that he/she does not know anything about but as comfortable as chatting in a natural and humanlike manner.
  2. Fast and uninterrupted: According to a Forrester survey, 73% of customers say that the most important prerequisite to provide a good experience in customer service is to appreciate the time of the customer. That means the speed and instant response is always very important for the customer. Chabots are advantageous in this sense, in addition to the instant responses, they have the ability to provide 24/7 uninterrupted service.
  3. Natural and human: For many traditional sectors, mechanical channels of the digital world, namely web and mobile are perceived as artificial and lacking the human touch by the customers. In particular, in the financial services sector, having a significant pre-digital history, getting information or advice through a human conversation and human touch is very valuable. For such sectors, chatbots become very important, promising a humanized experience similar to the physical one, rather than navigating through menus or performing the tasks self-service. In this sense, chatbots have a great potential to be “the interface” of the future.
  4. Personalized: In addition to the fast and uninterrupted service, chatbots can provide a personalized service when integrated into the company’s data analytics capabilities. This means that being aware of the customers’ activities and providing the relevant service at the very exact moment when the customer needs help is crucial. This is again a very valuable feature of the chatbots from the customer’s perspective.
  • Increasing the revenues:

It is very critical for a chatbot to identify the moment of truth of the customer, understand the needs and communicate proactively. The prerequisite for this is, of course, that the company has advanced data collection and analysis capabilities. A chatbots is a good data collection tool, while it works as a sales channel using the available data. A chatbot can add each and every data into the CRM system of the company as a result the conversation that it has with the customer. Also it can directly perform sales or create sales leads.  This aspect of chatbots increases sales and company revenue.

According to a report by PWC, in 2020, the contribution of artificial intelligence-based innovations to the global economy would reach $ 15.7 trillion, of which $ 6.6 trillion would be an efficiency effect, and $ 9.1 trillion would be the effect of increasing sales.

What kind of chatbots are commonly used?

It is possible to distinguish different types of chatbots from different perspectives. For example, we can talk about different kinds of chatbots, depending on which platform they serve through, for what purpose they are created, by which technology they have been developed. The following classification categorizes chatbots into 3 groups by evaluating their interaction, intelligence and integration competencies. In this article, we will evaluate and categorize chatbots in terms of their intelligence, in other words the technology used for their creation.

Maturity model: How can chatbots meet expectations?

an overview of chatbots 4

  • Menu / word based chatbots:

For this type, it is not possible to talk about an artificial intelligence technology. The user proceeds by selecting from existing menus and buttons or by writing some expressions. The menus and buttons are used in the customer interface and the decision trees, that we are familiar in call center IVR systems, exist at the back end. The system tries to recognize some words in the statements and match them with a response. Such chatbots can answer a limited kind and number of questions.

Menu / word based chatbots are insufficient to make sense of the written expression as a whole. In this kind of chatbots, in order to be able to direct the user correctly, all possible misspellings must be introduced to the system. They can be used for very simple Q&A functionality, but cannot meet the needs in complex situations.

As an example, let’s assume that the € / TL ratio is defined in the software. When the user asks a menu / word based chatbot “How many euros are 100 dollars?”, the chatbot recognizes the word “euro” and replies by showings the euro equivalent of 100 TL,  instead of understanding the whole statement and identifying what the user aims to learn in this question.

  • Statement-based chatbots:

It is possible to talk about NLP (Natural language processing) and machine-learning technologies in statement-based chatbots. The chatbot understands and processes the user’s expression as a whole. However, when the dialogue continues, it can lack the capability of following-up of the dialogue from the beginning and providing an appropriate answer.

Let’s take our previous example of currency conversion, a statement-based chatbot can understand the question of “How many euros are 100 dollars? But it fails to understand the question “Oh, how many pounds then?” Because answering that kind of a follow-up question requires the ability to follow the dialogue from the beginning and maintain it in a context-based manner.

  • Dialogue-based chatbots:

Such chatboats are also AI-based chatbots that are developed by using NLP (Natural language processing) and machine learning / deep learning technologies. Dialogue-based chatbots are not limited to a word or statement-based approach. But they can continue the flow of the dialogue and provide relevant responses aligned with this flow and the context as well as understand the natural human phrases as a whole.  If we test the same example with a dialogue-based chatbot, it responds to the question, “How many euros are 100 dollars?” and also responds to the subsequent question “How many pounds then?” as it can follow the conversation from the beginning and it is aware of the context.

Which kind of a chatbot is the best?

Actually, it is more accurate to ask, “What kind of chatbots are more suitable for the intended purpose?“. For example, for a B2C textile company, as the interactions are limited, a statement-based chatbot would be sufficient. It is also important to consider the value offered to the customer. If the value to be created for the user is provided via a menu / word-based or statement-based chatbot, it does not make sense to develop a dialogue-based one. Another parameter is the preferences of the target audience. It is crucial to know the target audience very well and develop a chatbot that is aligned with their expectations. Some users may prefer to proceed with menus, some may prefer to continue a dialogue by typing free text. It is essential to take into consideration these parameters when determining the type of chatbot.

Cbot can develop these 3 different types of chatbots for different needs. The company that gained the patents for its own technology, has successful use cases in terms of natural language processing and deep learning technologies. It has participated in European Union supported ITEA projects. Its own natural language processing infrastructure enables the company to provide advanced artificial intelligence solutions to the enterprises. Especially in the sectors such as banking / financial services, e-commerce and telecommunications, where customer expectations are considerably high and the scope of intents is very wide, statement-based or dialogue- based chatbots enhance the customer experience apparently. Cbot develops solutions based on its advanced artificial intelligence technologies for these sectors.

Conclusion

Chatbots –one of the most obvious trends that affected almost all sectors since 2016– provide companies the opportunity to manage their costs better, to offer a better experience to their customers and to increase their revenues. In order to design the chatbot that will serve these 3 purposes in the best way, it is required first to consider how chatbots are aligned with the broader strategy of the company, which problem the chatbots will address and accordingly what kind of a chatbot should be developed.

There are many questions about chatbots to be answered on companies’ agendas. So it seems that this topic will preserve its position as a priority to be discussed in depth for a while.

For any questions or comments, you can contact us via hi@cbot.ai.

Chatbots: The new UI for a better CX in banking