Using NLP based Voice recognition for trading in kite mobile

@Zerodha I think it would be nice to have voice based trading…where I instruct the app to place trades etc. This is very easy to implement and the experience would be akin to using the google assistant. Check out the following:





May be not any time in near future, I believe technology has to evolve lot more to integrate voice based with trading.

1 Like

NLP trading huh? do you realize the consequences if the voice processor picks up wrong instructions? Trading is a highly specialized skill, not meant for ipads, mobiles or even web based platforms. The colorful snowjob by Indian broking platforms onto unsuspecting clients is already too much.

1 Like

It could be designed such that it fills up all info and waits for your final click, so it’ll seek final approval from the user…the app will be ever improving on account of machine learning used across zerodha’s sizable client base reducing error rate at a very fast clip

ML can be used to create indicators, some people have managed to create neural network indicators for other platforms. I am currently researching into boltzmann distribution for <1 min charts…But still…day trading on mobile will make you lose money, maybe its ok to push delivery trades on phone while watching some nutjob on cnbc.

Good luck getting ur ML models to work… there’s some places online where quants have posted their work on neural nets, deep learning etc… I have always wanted to try it out but holding out for want of time
And I was not opining that voice trading be only mobile based, coz it can be deployed on desktop/pc on pi and even on kite

1 Like

This might be very interesting option for traders, but I think that there are many challenges to it. These challenges would mostly related to correctness of order execution by system and written proof that trader did actually ordered specific trade

The system will display all details for final confirmation, which has to be approved by clicking. This will minimise errors wrt order placement.

Google cloud speech API is already flawless and improving with every passing minute. So errors are less likely. The trader can claim even today that he didn’t place a particular trade, still a voice recording can be stored for future reference