How to offer customers the same user experience across all channels?1 This is the challenge that all companies are striving to achieve. Why do they focus their efforts only on “services” and leave “customer service” in the traditional format (telephone calls and mail)?
Why are companies making a big effort to make the interfaces of their apps and websites more user-friendly if users have demonstrated that they feel more comfortable using instant messaging apps (WhatsApp, Facebook Messenger, Instagram or Telegram)? Why don’t companies use this model to offer their services and customer service?
Companies can achieve the goal of a uniform customer experience by making it possible to interact in natural language, which is customers’ inherent form of communication and how they feel most comfortable.2 In other words, offering their services and customer service through conversational interfaces, either through voice or text:
What is my account balance?
I want to change my cellular plan.
Does my insurance cover the repair of the lock on my door?
My Internet isn’t working. I need it to be fixed.
Therefore, conversational interfaces3 will allow users to consume the “services” the company offers and access their “customer service” in a quick and convenient manner that will earn their satisfaction and loyalty.
These conversational interfaces will make the omnichannel challenge easier to attain – not only for internal channels (mainly app and web), but also on external channels (social networks, WhatsApp, Instagram, Telegram or Facebook Messenger). Furthermore, they can be integrated in any device: PCs, smartphones, clocks, smart speakers, televisions or cars. Let’s look at two examples:
With text
When a customer opens their bank’s app on their cell phone, they will not see the traditional interface of options and buttons. Instead, they will see an instant messaging app where they could ask:
“Has my paycheck been deposited?
“Why was I charged the last fee?”
With voice
A customer can use their smart speaker ask to speak to their insurance company to ask:
“Does my insurance cover the repair of damp spots?”
“When will they charge me for the next bill?”
And what lies behind these conversational interfaces? Who answers? Companies have two ways of serving the customers who use them.4 First, through human experts in every area of knowledge who respond to each of the customers requests. Allowing them to interact in natural language will enhance customer experience and therefore lead to a greater volume of users for the service. This exponential growth of the service can represent a problem in terms of cost and availability, issues for which the company must find solutions in order to support the scalability of the service.
The solution is to not hire more human agents because the costs will rise exponentially. This is exactly where artificial intelligence comes in and becomes the company’s best ally to address this scalability without costs soaring uncontrollably.
Natural language processing (specific area of artificial intelligence) makes it possible to automate and comprehend the requests that customers make on conversational interfaces, simulating the comprehensive capacities of human agents without the need for any humans to intervene.5
Therefore, it will be software that responds to the requests customers make on any channel. But it is not sufficient to simply understand the customer. The software must also have the intelligence to take the requested action and generate with a response for the customer.6 Below we look at these three tasks to be conducted by the software in greater detail and in the indicated order:
1º Understand the request made by the customer in natural language (text or voice) though a conversational interface.
For example, if the customer asks the bank’s software: “Have I been paid yet?”, the conversational interface will understand that the customer wants to know if his or her company has deposited his or her paycheck.
2º Take the action requested by the customer using cognitive skills.
Continuing with the example from above, the software will access the database to check whether the company has deposited the customer’s paycheck.
3º Respond, composing a response to the customer in natural language (text or voice).
Continuing with the example from above, the software will generate a response to the customer: “Your paycheck has not been deposited yet. It will probably be deposited tomorrow between 8:00AM and 10:00AM.”
This software has three skill levels to address customer requests (accumulative levels or incremental levels, from the lowest to the highest skill level):7
Level 1: Perform simple and repetitive tasks based on a specific domain of information on the company. With this skill level the software is called a chatbot.
Level 2: Perform simple and repetitive tasks based on different domains of heterogeneous information on the company. With this skill level the software is called a virtual assistant.
Level 3: Perform any complex task based on any of the company’s products or services. With this skill level the software is called a cognitive conversation system.
Furthermore, as they are based on artificial intelligence, cognitive conversation systems8 allow companies to offer new, unprecedented services that add value thanks to their analysis of the data customers generate on a daily basis
Descriptive analytics services
Analyzing patterns in customer behavior. The cognitive conversation system offers customers an indicator of how their behavior compares to similar anonymous profiles (same family size, live in the same neighborhood and similar income). It is currently an unprecedented tool that will be highly useful to customers. For example (banking sector):
Customer question: “How can I save?”
Software response: “You are spending 10 percent more on your electricity bill than those similar to you. If you adjust your behavior and home you could generate that 10 percent savings.”
Predictive services
Anticipating customers’ future behavior based on their activity patterns. For example (banking sector):
Customer question: “How much will I pay for water next quarter?”
Software response: “I estimate that you will pay approximately €150 as you usually fill your pool in June.”
Prescriptive services
Recommending customized products or services for customers based on their activity patterns. For example (banking sector):
Customer question: “When do you recommend that I buy the new iPhone?”
Software response: “You should buy it in the month of September because in June you will be charged for your car insurance and you will have expenses from summer vacation in July and August.”
Conclusions
Cognitive conversation systems allow companies to offer their “services” and “customer service” in a uniform manner through the simplest interface possible: natural language. Customers make their requests or ask their questions with their own way of expressing themselves through text or voice. Clicks on a screen are a thing of the past.
They will make it possible for the challenge of an omnichannel experience to become a reality, integrated into any channel and offering users the same experience on the company’s channels as well as those that belong to social networks – and on any device (tablet, cell phone, watch, smart speaker, television or car).
Cognitive conservation systems are the solution that companies need to meet their customers’ needs in the channel that is most convenient to them, when they need it and providing the specific solution they require – all of this for an affordable cost.
Bibliography
- Can We Talk? – The Impact of Conversational Interfaces on Human Autonomy Teaming Perception, Performance and Situation Awareness. Adam Bogg, Andrew Parkes, Mike Bromfield. IHSI 2020: 938-944
- Do Conversational Interfaces Kill Web Accessibility?. Marco Furini, Silvia Mirri, Manuela Montangero, Catia Prandi. CCNC 2020: 1-6
- How to Design and Evaluate Intuitive Conversational User Interfaces (USABLEBOTS). Andreas Bleiker, Kyoko Sugisaki. IUI Companion 2020: 1-2
- Submitting surveys via a conversational interface: an evaluation of user acceptance and approach effectiveness. Irene Celino, Gloria Re Calegari: CoRR abs/2003.02537 (2020)
- Communicability of traditional interfaces VS chatbots in healthcare and smart home domains. Stefano ValtolinaORCID Icon, Barbara Rita BarricelliORCID Icon & Serena Di Gaetano Pages 108-132 | Received 16 Nov 2018, Accepted 22 Jun 2019, Published online: 30 Jun 2019
- Analysis of conversational listening skills toward agent-based social skills training. Hiroki Tanaka, Hidemi Iwasaka, Hideki Negoro, Satoshi Nakamura. J. Multimodal User Interfaces 14(1): 73-82 (2020)
- A Smart ChatBot for Specialist Domains. Roberto Canonico, Giovanni Cozzolino, Antonino Ferraro, Vincenzo Moscato, Antonio Picariello, Fabio Raimondo Sorrentino, Giancarlo Sperlì. AINA Workshops 2020: 1003-1010
- Sistemas Cognitivos Conversacionales. Inteligencia Artificial: Sistemas Conversacionales Cognitivos: Procesamiento de Lenguaje Natural, Computación Cognitiva y Sistemas Inteligentes Chatbots, Voicebots y Asistentes Virtuales. Javier Porras Castaño. January 2020. ISBN-13: 978-6200343871
Comments on this publication