According to a study conducted by Oracle, 80 percent of the respondents are planning to use chatbots for customer interactions by 2020. At the same time, 36 percent of them have already implemented a chatbot service. Considering that the mass use of chatbots has just begun, the expected growth is quite impressive.
Previously there was the 60’s command line interface (CLI). A command-line interface or command language interpreter (CLI)(also known as a command-line user interface, user interface and character user interface (CUI)), interacts with a computer program where the user (or client) issues commands to the program in the form of successive lines of text (command lines). A program which handles the interface is called a command language interpreter or shell.
In the 70’s there was the second generation Graphical User Interface (GUI). A GUI (pronounced as either G-U-I or gooey) allows the use of icons or other visual indicators to interact with electronic devices, rather than using text via the command line. For example, all versions of Microsoft Windows utilize a GUI, whereas MS-DOS does not.
Graphical User Interface (GUI) is the third generation of NLUI. Natural language understanding (NLU) or natural language interpretation (NLI) [1] is a subtopic of natural language processing in artificial intelligence. Natural language understanding is considered an AI-hard problem Bot, is one of the third generation that can be run with language “more natural” to humans, where humans feel they can discuss issues with the computer system.
When Will It go Live?
Nowadays, some companies have utilized chatbot to empower their own web and mobile-based apps. However, whether this is effective or not should be evaluated because it is only a few companies that have taken this step. The companies should consider that chatbot has three main functionalities:
- Customer Service
In terms of business and industry, chatbot can be used to reduce the cost and effort in organizing a customer service or call center. To date, it cannot 100% replace the function of customer service, but if the function of customer service and chatbot are combined, up to 94% of business needs can be solved through chatbot. This has been experienced by Kata.ai, which developed Veronika, as a virtual assistant. - Productivity Bot
A chatbot that functions as knowledge discovery will help the user to receive information faster. It leads to better and faster decision-making, which automatically improves productivity. Perhaps the most sophisticated example of this function is shown by Alexa, a virtual assistant owned and developed by Google. - Engagement Bot
As an engagement support, chatbot can be used by an organization to interact personally with the user. IBM Indonesia predicted that what is exciting about future chatbot is its potential to be leveraged into conversational advertising.
What did Companies Prepare to Develop Chatbot System?
1. Intent detection
Whether the machine can understand the customer’s words correctly. For example, I am hungry. What kind of hungry? The definition of intent depends on whether the bot is created from scratch or not (existing / already running). Here we take the example of bots made from scratch such as CS bot. Firstly, we can collect call center transcripts. Next, an analysis of the existing conversation transcripts is performed. Lastly, we perform clustering for existing conversation data. From there we get the TOP 5 intents with which customer’s make enquiries, which is useful for companies in terms of dealing with said enquiries.
There are two main types of intent:
- Casual Intent — These queries should direct your bot to respond with a small talk reply like “Hello what can I do for you today”. The casual intent queries also consist of affirmative and negative statements like “Ok”, “yes please”, “nope” and so forth.
Having General Casual affirmative and negative statements helps you handle all such queries and take them in context with the conversation the bot just had with the client. - Business Intent — These are queries that directly lead to business transactions. They are more complex. Ideally, you should put more thought into business queries for this reason and because the small talk statements such as saying hello or affirming choices is taken care by general casual intent queries.
2. Entities
Business intents generate metadata about the intent called “Entities”. The process of finding the entities can be understood at Part of Sentence (POS) tagging. As a user of NLP and as a service you don’t need to get into the technicalities of knowing how POS tagging works. However, if you would like more information you can read about it here.
Whenever a user is thinking about designing their intents the entities must also be identified and labelled accordingly. Again, in entities you can have general entities labelled for use throughout the intents like metrics (including quantity, count, volume, dates). NLP as a service allows you to tag entities of such general types without any big hassle.
3.Determining The Target Customer and Their Behaviour
As Peter Drucker, the father of modern management, said, The aim of marketing is to understand the customer so that the product or service can be provided in accordance with the customer’s needs and requirements. This means the product can really sell itself. There are the elements you must consider:
- Who is your target customer?
- Describe their life situation or your business situation
- What do they want?
- What are their problems?
- What are their needs?
How to Choose The most Suitable Social Media Channel for Clients?
- The third-party must know the most existing media data whether in Line, FB, or other.
- Companies should be able to see the market demand most often present in social media.
- Customer segmentation. Focus on the channel that often gives more impact. Companies do not need to play on all social media channels.
- Security aspect of social media. Particularly for the financial industry, the company should consider terms and conditions even a privacy policy.
- Social media capabilities when integrated with engine and bot.
How to Implement A Strategy for Making Bots Work Effectively to Help Humans?
- Training must be provided by the bot trainer in order to increase the excellency of the bot.
- Escalation of the call center. Create an empathy element in the bot. So it can operate on the same frequency as the humans.
- Do not lie to the customer if the companies responses come from a bot.
How to Maximize Return of Investment (ROI) through Chatbot?
- Maximize infrastructure related to chatbot development implementation.
- Analyze in detail the company’s needs regarding chatbot usage.
- Maximize the number of customers that bots can handle.
- Analyze in detail the percentages directed to bots and humans.
- Analyze financial calculations related to the number of customers and cost of development bots.
Why Bot is The Right Tool and How to Integrate It with The Existing Company System?
The first advantage of using a bot is the ROI gained. One example is the Talkabot.id experience, with chatbot implementation used by one of our clients, can increase ROI up to 1.1M (BCA API). In addition to ROI or efficiency, in the business world, there is a term comparing the quality of the company with various improvements.
This chatbot’s function can help improve the company’s quality to prepare future needs. Through the use of a bot, user experience (engagement new model) can be further improved. By using bots, companies can more compatible than their competitors.
On the other hand, there are some points to consider before integrating a bot with an existing system:
- We need to know the existing data from the company first.
- Integrate bots from third-party.
- How many customer conversations are already conducted through company apps.
The Future of AI and Something must be Prepared
Every innovation nowadays in the IT world has a very high impact. Currently, in the changing world of IT, innovations can happen almost every month. As it happens, almost all aspects in our everyday lives, we interact with AI. However, please note that humans will not be replaced by AI. AI and humans will work side by side. AI will help to streamline human performance and help companies to be more productive as more work can be completed with this extra resource.
In the future, there will be many bot trainers. This is a way to make the bot’s ability to keep improving(getting smarter) and respond to human needs. To be able to coexist with AI, humans not only need high IQ quality, but also EQ and SQ. Since human IQ could be dominated by AI ability, it makes empathy and sympathy elements of human beings e the most important thing.
Last but not least, based on the predictions of technology research firm Gartner, in 2020 the public will tend to have more conversations using chatbots. Meanwhile, according to Accenture, a global management consulting firm, chatbots will boost economic growth by twice the current growth and up to a 40% increase in labor productivity by 2035. Deploying chatbots will enable companies to automate simple processes so that up to 94% of business needs can be solved through chatbot and they can be an effective tool to collect data.
On Monday 23rd of April 2018, Qiscus held a monthly event, #TechTalk 107. Taking place in Kolega x MarkPlus Coworking Space Jakarta, the event was attended by participants from various professional backgrounds. At this event, Hokiman Kurniawan (CEO Bahasa.ai), Bobby Pratama (CTO Talkabot.id) and Gito Nugroho (Software Client Architect of IBM) attended as speakers. TechTalk #103 discussed introductory matters related to chat and AI. Following up on this, TechTalk #107 discusses “5 Practical Questions CXOs Must Ask When Adopting Chat + Al”. We summarized the session of TechTalk #107 through this article.
