Skip to content

Chatbots basics

Understanding Chatbots

Quote

A chatbot is a software application used to conduct an on-line chat conversation via text or text-to-speech, in lieu of providing direct contact with a live human agent.

-WikiPedia

Chatbots are designed to converse with humans. They are meant to simulate human way of conversing. It needs a lot of training to reach a an acceptable level.

It was Michael Mauldin (creator of the first Verbot) in 1994 who coined the terma "ChatterBot".

Application of Chatbots -

  • Customer Service
  • Information gathering
  • Responding to FAQs.
  • Virtual assistants like Google Assistant, Amazon Alexa etc
  • Educational chatbots
  • Entertainment chatbots
  • News chatbots
  • Health chatbots
  • Banking chatbots (Finance)
  • e-commerce chatbots
  • Healthcare etc

Info

Turing test-

In 1950, Alan Turing's famous article "Computing Machinery and Intelligence" was published, which proposed what is now called the Turing test as a criterion of intelligence. This criterion depends on the ability of a computer program to impersonate a human in a real-time written conversation with a human judge to the extent that the judge is unable to distinguish reliably—on the basis of the conversational content alone—between the program and a real human.

-WikiPedia

Limitations of Chatbots -

its still an active area of research as there's no perfect chatbot availble till now. They have the following limitations -

  • They can fail when an unsaved or untrained query is asked.
  • Their efficiency highly depens on the Natural language processing. Hence sometimes accents and minot spelling mistakes may lead to a bad response.
  • They have some sort of tunnel vision i.e. they can only focus on current query. They can't reply to multiple query at a time.
  • They need a lot of data to get trained.

And there can be other limitations as well but these were the major ones.

The need for chatbots as per Business perspective -

  1. Easily Accesibile 24x7: -

    They are availble 24x7 for the customers. So they can tirelessly work without taking off unlike humans. And they can easily answer some generic queries plus can take information from the users effectively.

  2. Efficiency: -

    Any customer can get a lot of basic detials about their stuff like where is there food order? or ETA of the delivery? Train schedule status etc in their mobile devices any where any time.

  3. Scalable: -

    It can handle multiple customer without getting tired. Plus waiting time also get reduced unlike the IVR calls where you have to wait untill the customer support agent gets free.

  4. Cost effective: -

    Certainly it can save you a lot of time and money with this much of efficiency and availibility.

  5. Insights and lead genration: -

    Most of the time sales team can get their leads and insights from the chatbots. With this insight it helps them to know their client before hand like with their occupation, age, gender etc we can anticpate a lot about the customer.

Point to consider before building a chatbot

  1. is problem solvable by simple QnA ?
  2. is it highly repetitive which requires fetching and anlaysis of data ?
  3. can it be automated ?

If answers to above queries are yes then your business problem can be solved with a chatbot.

Common terminologies used in the chatbots

  1. Intent

    This means what could be the intention of the customer when he/she is asking certain questions.

    Example

    Q1. tell me about today's weather?

    Q2. how's the weather today ?

    the intent in both the questions user wants to know weather

  2. Entity

    This are the parameters or the values that we want to extract from the queries asked by the users.

    Example

    Q1. Book me a flight from India to US?

    the entity is source and destination of the flight intent is to book a flight

  3. Utterances

    This is nothing but the response of the chatbot for the query.

  4. Fulfillment

    This is nothing but processing an information given by the user and sendinf them a proper response. Here processing may involve interating with several APIs and then responding back.

    Example

    Booking a flight or cab.

    Ordering food etc

This terminlogies will come into picture when we will build our bot