How should a Bot behave? As smart as your Mom!
I just landed in a User Experience Design Industry and exploring artificial intelligence in the new millennium. In this quick expanding world of AI, Bots are the next “big thing.” Excited to discover human vs machine conversation!
I and my team are building a Bot design platform that helps one to create Bots and train them. The user can define the conversation flow and streamline the amount of information sharing to the user.
Training a bot with all the guidelines is the challenging task, but this is how your bot will be clever and meet business goals.
I can relate this to the conversation I have with my mom. In spite of all the improper details and queries I ask my mom, she easily understands my need and helps me out.
She understands each and everything very well and tries to fulfil it. Mom is the best example of Natural Intelligence in every house. Natural intelligence (NI) is the opposite of artificial intelligence: it is all the systems of control that are not artefact, but rather are present in biology. Understanding NI requires understanding all of these influences on human emotions, behaviour, and their interactions. Can we design a bot that smarter? Maybe!
A Bot needs to provide information in terms of text, images, videos, locations, codes, etc as user demands.
Understanding a users intention with too many conversational errors is a challenging NLU (Natural Language Understanding) training. Thus developers trying out the best practise to make sure the product tool delivers accurate user expectations.
Many synonyms, typos, conversational pattern and subject matter vocabulary is supposed to be added to the Bot so that it replies correctly and meets the User expectations. It might have the following issues to be addressed.
- Spelling and grammar check
- Finding synonyms, antonyms, homonym, acronyms, etc
- Parsing information from the internet.
- Understanding meanings of words, phrases, sentences, documents, and queries
- Machine Translation (e.g Translate text from English to Hindi)
- Question Answering and doing tasks (e.g scheduling appointments, ordering a laptop, Booking Tickets, etc.)
“A consistent experience is a better experience.” — Mark Eberman
Once you provide a better User Experience, there is no way back!
Making each word related to logic
For e.g.
- The word “Rose”, has two meanings. One is to have gotten up and another is a type of flower.
- The word “Current” has two meanings. One is up to date and another is the flow of water.
The bot should behave smartly to understand which meaning to pick up as an intent and trigger them to its entities with the concerned expression.
Making each sentence relate to the logic
For e.g.
- If you mess with the bull, you get the horns.
- Weeds are easy to grow and hard to kill.
The sentences convey a meaning which a human brain easily understands but for bot its a mess.
These sentences have an independent clause, thought of the phrase can relate to anything. The dependent clause is used to turn a simple thought into a more descriptive, informative one.
Question answering
For e.g.
- What is the capital of India? Ans. Delhi
The bot can easily understand the intent of the question and answers it correctly.
- How could I present myself better?
The bot has too many suggestions for this question and now Bot needs some more clarity to answer it for a specific purpose.
These Bots will become more clever and human-like, hence we would be addicted to their intelligence. Following a proper information study and involving better User Experience Design methodology, your Bot will be more engaging to your audience.
Cheers!
Image Credits http://www.positivemomdiary.com/ https://apparat.cc/world/jobs-for-machines-life-for-people/attachment/robots-army/ https://giphy.com/ https://emilyhagedorn.wordpress.com/2012/01/11/2012-will-be-my-year-to-code/
