Indeed, chatbots are nothing new and have been known for many years. Since the first artificial intelligence developments, scholars have tried to reproduce typical human interactions and automated communication processes artificially.
However, with the recent development of information technology and computing capacity and the access to an immense amount of information encoded in the textual language (thanks to the use we make of social and web platforms), the understanding of natural language NLP ( natural language processing / natural language understanding) has had a huge acceleration coming to produce the first significant bit applications in the business world. But let’s go step by step. Let’s first see what a chatbot is and what a human language simulator consists of.
What Are ChatBots
ChatBots are artificial intelligence software platforms programmed to interact with the user in a conversational mode. Chatbots are software that exploits the potential of NP (natural language processing) and artificial intelligence to simulate conversations and advanced interactions. From this point of view, the evolution is epochal because it marks the end of one-way mass communication (a single platform that communicates to a vast audience) to two-way communication (one user), with the creation of marketing technologies capable of establishing unique interactions with each user.
More than 2.5 billion people today use at least one mobile app to communicate, and 90% of interactions between users take place on mobile devices (smartphones and tablets). Most of these interactions occur with textual communication systems (such as chat, Messenger, whatsapp, etc.). As we see them today, chatbots can be integrated into the most common communication systems such as chat, Skype, Whatsapp, Facebook Messenger, Slack, Telegram, Cortana and many others.
Companies now understand how to interact with users. They must be present where they spend the most time, using the tools most suited to establish contact as naturally and spontaneously as possible. Therefore, the potential of chatbots as a marketing tool to create and manage two-way contacts (one-to-one) seems enormous, as this type of contact (different from the one-to-many mass communication we are used to) presents a strong criticality in costs and management.
The potential of chatbots Let’s think, for example, of corporate customer care (assistance centers). We have been witnessing a confusing process for a long time: users demand direct contact with companies and increasingly personalized services. On the other hand, companies find it difficult to manage the growing costs and standards of the assistance points and keep up with the evolving market and product.
We are therefore witnessing the growth of increasingly inefficient (but increasingly critical) call centers, often outsourced and with obvious management difficulties (for cost containment), an evolution and increasingly rapid market dynamics (with obvious training of the assistance centers and the rotation of personnel – moreover underpaid). However, let’s examine the call center processing data. We find the classic Pareto 80/20 distribution, i.e. 80% of the requests are generally attributable to 20% of the cases, thus having (expensive and inefficient) resources that respond for the majority part of the time to the same questions.
In this phase, one of the most interesting opportunities a chatbot offers is providing a 24-hour operational contact: if properly built, a chatbot could manage and filter the most repetitive and documented cases, thus leaving the management of cases to human operators. They require real problem-solving skills. The real revolution, therefore, lies in programming advanced conversational interfaces, which, by interfacing with corporate infrastructures (ERP, CRM, etc.), can build advanced interactions that are better than the current ones.
Examples Of Using ChatBots
Here are some of the most common use cases of chatbots, which represent an excellent marketing and cost containment opportunity.
Customer Care / Customer Assistance
As we said before, support and customer care activities, especially in simple and limited contexts, are the ideal fields of application for ChatBots. Providing simple, timely information or answering frequently asked questions (FAQ), explaining how a device works, and giving indications on the point of sale or how to use it are perfect scenarios for a ChatBot, which can provide unlimited support 24/7.
As we said, the assistance center always responds to the same requests. A ChatBot could provide a great first filter to determine the type of request and, if capable, provide a correct first response, leaving the opportunity to switch to a human if there is an element of uncertainty or if requested.
The bot has excellent potential in simple and inexpensive products that do not require a very long process and are perhaps easily integrated into conversational contexts. Let’s imagine, for example, the purchase of an offer coupon with a very rapid expiry. This product could easily be purchased in the context of a social chat (like Messenger), perhaps after a click on an Ad.
If the product is known and there are few variants, the user could easily rely on a bot to complete a purchase of this type. Below we see, for example, the integration of a clothing sales process via chatbot on Messenger. The optimal integration between bots and CRM here could have the maximum potential, providing offers and cross-selling by drawing on complex data from different business sources.
Both B2B and B2C can involve cases of complex information requests that require calculation and processing skills, such as, for example, knowing the weather forecast in a given location and finding out if there is an offer for a trip to the same location. In the B2B world, this could be, for example, the case of a product configurator and supply of a customized quotation: by collecting data with a question and answer process and crossing the data with price lists and internal management systems, complex bills of sale could be generated.
Outbound activities are those started automatically by ChatBots, perhaps exploiting the ability of social platforms to connect directly with users. Let’s imagine, for example, a dentist’s ChatBot that can interface with customers via the Facebook page. For example, it could remind users of the right time for a dental cleaning, cross-referencing data from old visits and the study agenda. When choosing a good use case for ChatBots, considering the current stage of evolution, it is advisable to start from simple interactions and conversations in particularly defined and controllable contexts.
Limitations Of ChatBots
The evolution stage of ChatBots is still quite primitive, and perhaps this is an interesting element for those who want to take advantage of this opportunity on the market and start using this tool for marketing or support purposes. The major limitation concerns the ability to generate intelligent deductions and understand complex conversations. Simple conversations taking place within a simple context are easily managed, but as the context of the conversation expands, the chatbot’s capability becomes apparent.
Real Examples Of ChatBots
Some of the real examples of the use of bots that have been launched these days are, for example, the chatbot launched by BMW to configure the new 5 Series. In this case, the BMW Messenger chatbot is programmed to collect personal information (preferences, etc.), which will enrich the corporate CRM, providing, on the other hand, information on the car and commercial network.
Platforms To develop ChatBot
The big market players have been moving for some time to develop platforms to facilitate the development and integration of bots. We all know Microsoft, Google, Facebook and Amazon to discuss the most advanced. Below we examine the most used ones in the above areas.
Microsoft Bot Framework
Perhaps the oldest history is that of Microsoft (https://dev.botframework.com/). It offers advanced integration with Cortana (Cortana is a speech recognition and assistance software developed by Microsoft for the Windows Phone, Windows 10 Mobile, Windows 10, Android operating systems and for the Xbox One console OS.
Cortana allows users to interact in natural language to perform operations such as conversations, sending SMS, setting reminders, turning phone or PC features on and off and obtaining online information). The Microsoft SDK is open source and allows you to produce bots that can be easily integrated throughout the Microsoft world (Skype, Office 365, Team, Bing, etc.) and beyond (Facebook Messenger, Slack, etc.).
This web platform (http://api.ai/) was recently, in September 2016, purchased by Google, which is investing heavily in this sector, even if a little late compared to other competitors. The API.AI platform is free, and recently, the usage limits (number of calls) have also been removed.
More than a platform for building bots, we can better define it as a Natural Language Understanding platform that allows you to build bots that are simple in flow but effective in understanding and learning ability. API.Ai offers excellent multi-language support (currently 14 languages) and the possibility of integration with websites and platforms (Twitter, Slack, Skype, Telegram, Messenger, etc.).
Acquired by Facebook in 2015, this platform (https://wit.ai/ ) is also totally free and open source. Wit has excellent integration with Facebook Messenger, allowing you to build bots for this platform quickly and effectively. However, it has a very limited integration set with third-party applications while featuring open-source libraries and SDKs for iOS, Ruby, Node.js and Python.
ChatBots are rapidly evolving, and all the major players in the sector are moving quickly and investing heavily. The clear intent is to provide marketers with advanced tools to increase engagement with the public and users within the platform and where users spend most of their time (on the move and in social contexts).
Surely in the coming months, we will see a rapid evolution of bots, leading to increasingly evolved and natural bots. At the same time, users will gain more confidence and familiarity in interacting with these tools, creating enormous potential for companies investing in this sector.