Wednesday, July 18, 2018

Solution for rumors going viral in Whatsapp


First of all, I would like recognize and appreciate Government of India for acknowledging the problem of virals. Because acknowledgement by a Government authority brings the problem on to the center stage for discussion, debate, solutionizing, addressing it - no problem can be solved without first acknowledging that there is a problem and needs to be solved.

Also, for the stern ask to Whatsapp to do something to curb the virals being spread via its platform without thinking about its feasibility or practicality. Because, the one expecting solution should probe people to solve the problem without getting into its nitty gritty details, else there will be no one asking for a solution which inturn means there will be no one thinking about solving it - no ideation or innovation will occur. (There can be exceptions where few can volunteering think about finding solution but they will be very few of them)

For this ask of GoI, Whatsapp has also reciprocated it with equal seriousness and is trying to find a solution to its best possible and has not washed off their hands saying like others that they are just a platform to share for people and so it is not their mistake but that of people's using it. It is true that Whatsapp is just a medium for sharing and it is the people who are not using it in a right way for right things. Need to appreciate Whatsapp for deciding to find the solutions for addressing it.

After going through Whatsapp's call for researchers/scholars to come up with solution for this problem, I was thinking about it. After a while, I got this simple idea if implemented by Whatsapp will definitely curb the spread of mis-information or rumors (though may not stop it completely).

This idea is using the human psychology aspect and more specifically Indian's mindset (may be applicable to people of other countries but I am talking about only whom I know).

This idea or solution needs only a small change (in its minimalistic basic approach) in the app by Whatsapp.

The Idea is:
When a Whatsapp user tries to forward a video, a popup can be shown asking for user to confirm for proceeding with forwarding or not. This popup should contain few notes in the language chosen by the user.

The notes can be something like for example:
  • Are you sure this information is true and not fake/rumor?
  • Please note that you can be held accountable and legal action can be taken deducing that you are a participant in, if any, untoward incidents (link which will take to a page listing the various incidents/cases that are considered severe and punishable and has to be avoided in terms of public safety) arises due to forwarding of this information.
  • Link pointing to general guidelines of do's and dont's w.r.t forwards.
Note: The above wordings of the popup is just for example. I suggest Whatsapp to utilize the appropriate experts in finalizing the content of it.

Note: Incidents to be listed should be done by working with GoI officials which would have discussed, debated among public and approved via proper legal route.

The content should be something that will discourage people from forwarding information carelessly, without validating it or second thought to others.

Since currently it's free and easy to forward the information to others, people often end up forwarding the information irrespective of whether it is valid or not, it is true or fake, often thinking that "what cost is there in forwarding something that he/she has got from someone else - if it is useful, it is useful else the recipients will ignore it". This is the mindset of people who forward everything carelessly.

But the moment they are told that they are accountable and can be held legally while forwarding, then all of a sudden, the mind will start re-thinking and validating about the information that is being forwarded and Whatsapp user will take a conscious decision about proceeding or not.

Since freedom should also be respected, we should not stop people from forwarding the information. Just enable the user to make conscious decision on forwarding instead of carelessly.

IMHO, this will definitely curb the wide spread of virals and also, bring consciousness among users on what to forward and not.

Also, Government can take appropriate legal action on those people if caught involved in spread of mis-information which caused harm to public safety.

This solution as described above is in its minimalistic approach or basic approach. However, this can be extended further by implementing algorithms that can make this solution contextual and intelligent rather than being dumb on asking for every forward.

I will talk about few of those enhancements in more detail in my next blog if anyone shows interest about them.

Tuesday, February 20, 2018

AI and ML for enhanced customer service

Oflate lot of customer service mostly online chats have been taken over by bots who uses AI, ML and NLP capabilities to enhance the interaction with customer.

Will they replace the need of human customer support representatives (CSR)? No way!


Bots will mostly be an add-on to the human CSR.

The majority of the areas that bots can contribute w.r.t customer service will be:

1) 24/7 availability reducing the need for humans to be 24/7 available

2) Handle initial interaction with the customer

3) Handle trivial simple interactions with the available knowledge base

4) Handover the interactions which are complex, new or just need human interaction to human CSR

5) Help the human customer representative in addressing the service request of customer

6) Sort the service requests into various categories as needed. For example, based on sentiment, urgency, category, etc.

7) Flag and follow up on specific service requests either with customer or human CSR

8) Retrospection of past service requests and solutions to derive the enhancements or changes to be done in the ecosystem of a company so as to improve its user experience and bottom line.

What other areas of customer service do you think AI/ML/NLP capabilities can help with?
Please comment and let me know...