Thinking about Chat bots?
There is a lot of buzz about Artificial intelligence (AI) and Machine Learning (ML) these days. It’s the cool kid in town. However, when people talk about AI and ML the first thing that generally comes to mind is conversational bots. The growing general sentiment is that having a conversational bot is better for getting relevant information than traditional website sections, apps, or other digital means. Just think about how easy and convenient it would be if we could just ask a bot for any information instead of scouring web pages. Right?
Just like for website and mobile apps the most important thing is user experience, the same goes for conversational bots. There are however some other factors as well which impact a bot’s success.
Let’s look at some of Characteristics, Challenges and Applications of successful bots:
Characteristics of successful bot:
1. Is bot available on platforms where it can reach maximum number of users?
2. Can user get intended information with minimum steps?
3. Does the bot provide engaging or guided experience?
4. Does the bot provide more humanly experience?
5. Does the bot have ability to pass control to human or know when human intervention is necessary?
6. Can bot empathize with user?
7. Knowing ahead that which types of problems Bot will solve? and also able to set the user expectation while conversation so for e.g. Instead of asking "How can I help you?", it can say "I can help you on reserving flights, trains and buses".
Challenges:
1. Humans don’t talk linearly they tend to change their mind or flow of conversation however they still expect to keep the context of conversation.
2. Sometimes humans want to drive the conversation instead of following the bot’s questions especially if the bot tries to force a specific order or questions or repeat questions.
3. Bot has to provide interactive interface as user experience matters the most in cases where mere intelligence or knowledge doesn't help.
4. Bots should have active learning capabilities so with time it improves and able to answer most of the queries related to context.
AI/ ML capabilities:
To achieve success in modern bot world there are following cognitive capabilities are required.
1. NLP (Natural Language Processing)
2. Active Learning
3. Visual Recognition
4. Sentiments Analysis
5. Voice recognition
6. Personality characteristics Identification
7. Language Translation
NLP (Natural Language Processing):
Natural language processing is used to make computer understands human languages.
Active Learning:
Cognitive model which helps to ingest documents and build understanding to answer queries. It helps bot to learn progressively and improves with time.
To make sure Bot always improves with time, system should run in hybrid mode so bot can learn with human experience.
Visual Recognition:
Visual recognition is another input which bot can use to empathize more with humans and also there are some other applications of it such as
1. Face detection, can be used for authorization/ authentication.
2. Identification of objects, visual diseases, color etc.
3. Able to judge user emotional state such as anger, happy, sad, fear etc.
Sentiments Analysis:
Identify sentiment of text (Positive, Negative) it can help bot to identify either continue conversation or pass the conversation session to human agent (depends on implementation), connect with user more humanly and also able to gauge general sentiments of whole conversation session.
Can be used to gain information like if people are happy to chat with bot.
Voice recognition:
It helps computer to recognize voice, It can be used for
- Authentication/ authorization through voice.
- State of user through voice like fear, resting, running etc.
Personality characteristics identification:
Generally bot gets this data while doing conversation with humans and through machine learning this data can be processed to understand user habits and preferences on an individual level. It helps computer to connect with humans more naturally.
Language Translations:
Language translations can be utilized to have conversion from one language to another, it helps to reach wider audience.
Applications of Chat Bot:
Here are some industries where chat bot can be useful.
1. Health care
2. Customer Services
3. Education
4. Financial
5. Social Networking / Entertainment
6. Product recommendations/ Marketing
7. Weather
8. Online Retail stores
9. Ticket Booking
10. News
11. Tourism
Chat Bot Platforms:
Here are some interesting chat bot platforms
Chatfuel, Gupshup, ChattyPeople, BotKit, Octane.ai, Flowxo, Wit.ai, Facebook Messenger Platform, Microsoft Bot Framework, Google Dialog Flow, IBM Watson.
Conclusion:
Even though most people go gaga about integrating chat bot on their web applications or mobile apps it is very important to keep in mind what problems the bot will be solving and what capabilities of bot are required to solve these problems before. Hopefully this article gives you good information to start your bot creation journey
