What are voice and chatbots, and what is the situation today?
A bot, either a chatbot or a voicebot, is an autonomous program able to interact with users on a specific channel, textual or vocal, through a natural conversation. Today, we are seeing a major rise of vocal devices like Google Home or Amazon’s Alexa. You can ask them to do many day-to-day actions like playing music, getting a recipe, an internet search… This rise has been seen at different levels. Alpine.ai published a report indicating more than 1 million searches by voice per month, and this number keeps growing. Another significant number is that there were 16.8 million smart speakers sold in the second quarter of 2018, a major rise compared to last year. This really shows that interacting with bots is a major trend these days.
What are the benefits and hurdles raised by bots for end users?
For end users, chatbots and voicebots are the ultimate level of instantaneity and personalization they can expect. The solutions are available 24/7 and are ready to provide any information, but they also can take action in regards to users’ personal information. If a user asks a voice device to play a playlist, the bot will recognize that playlist in the user’s profile.
On the downside, the technology is very new and there were a lot of initiatives in the past to provide such solutions, but they were mostly disappointing at the time because bots were not mature enough and could not properly understand the requests users made. This resulted in customer frustration and therefore the dereliction of the channel. Today, the challenge is to provide a mature solution that can meet users’ expectations.
What are the benefits and hurdles raised by bots for companies?
For companies, the main benefits would be the automation of customer support, or at least the handling of redundant and simple requests. Up to 30% of customer support costs could be saved through bots. Indeed, customer satisfaction participates to the improvement of banks’ image, and consequently clients’ satisfaction towards companies. As a matter of fact, we are adding a new channel through voice for companies, and that enhances cross-sales opportunities. Finally, companies can use the data that goes through the bots to enlighten business opportunities. Indeed, it will help them understand their clients better and therefore address them with more relevant inputs and offerings.
However, on the downside, customer service teams could be frightened by bots as a solution because providing an automated solution may lead to thinking it will replace their jobs. On a more internal perspective, since this is a very technical solution, companies need specific resources to train, maintain and enhance this kind of solution. This is why they need to be accompanied through the implementation of these solutions. Finally, the last hurdle for these companies is a user problem: they are still not sure of users’ interest in their markets, so they want to try a small solution instead of going big and implementing a complete bot solution.
Are there any additional challenges with voice and chat bots?
The technology of chatbots and voicebots has already been tried, but the results were disappointing. Therefore, one of the main challenges relies on the maturation of the solution: a good understanding (NLU = Natural Language Understanding) of end users’ inputs and the capacity to answer the request properly (NLG = Natural Language Generation). To do so, chat and voice bot solutions must rely on the most relevant NLP (Natural Language Processor = NLU+NLG) engine, and also properly design conversations to ensure the most natural interactions between users and bots. Another challenge is to understand which channel is used to provide an ergonomic answer (graphs for visual, summary on vocal).
The second biggest challenge with the democratization of chatbots, especially for retailers, is the security of the solution. Indeed, users need to be at ease with sharing their personal data and processing payments, and this requires a specific effort and focus from chatbot providers. These providers are therefore expected to deliver a high level of services, including a recurring update of the formulations and the training of the NLP engine. The finality of this constant improvement is to lead to a completely automated machine learning process.
What is Worldline’s position in regards to bots?
Worldline has developed a multi-channel Conversational Platform, which processes chat and voice bots on the most used devices (Google Home and Alexa), or which can be integrated into existing web/mobile pages, including specific expertise in the banking sector. The WL Conversational Platform provides pre-trained conversations on 3 topics:
- Daily banking operations, like account consultation, or credit card management, through natural conversations;
- Dynamic frequently ask questions, so that users can ask for an answer instead of looking for it on a dedicated webpage;
- Financial Assistance, which allows users to better understand their spending, the way they could save money, and even achieve projects thanks to detailed financial insights and advices. To do so, the engine integrates a PFM (Personal Financial Management) solution.
On a more architectural note, the solution is agnostic from the NLP engine, which means it can adapt to several engines for the training and use of conversations. Indeed, depending on the engine, intentions can be understood differently, and in different languages. This asset ensures that WL Conversational Platform provides its clients with the best fitting solution.
WL Conversational Platform is continuously evolving in order to help its customers deliver compelling and engaging experiences that strengthen the customer relationship by providing additional services, such as a teacher who could interact with kids to teach them simple finances notions, or also travel assistants, that would help you get through difficult processes like international insurance or currency exchange.