Niklhi Dhar: what AI changes—and what it does not

I think what it really brings to the game is efficiency at this point of time. We rightly said we also have a massive call center. 60-70% of our queries whether they were on social, email or calls were just information. We moved all of that under the whole chatbot engine on our website, WhatsApp, etc. Working with ChatGPT at this point of time to figure out what is the new caption. Where AI currently is not there is this whole concept of tone of voice, right? Nikhil, Nikhil, so we have seen some impressive use cases of AI in conversation marketing and something not necessarily in India but global examples. It's going to be easy to adopt once it succeeds over there.

So like the Domino's AI powered chatbot that allows customers to place orders via voice commands also. There was an Indian case study with the peer and life sciences where they have done a voice assistant and we just published a blog on it. So those are the use cases that are coming. So with that said, what would be the top three recommendations for harvesting the power of AI in the next 30 days to elevate? So it's not even far. Next 30 days what would have come of the top three? Sure. Thank you for first of all inviting. But I think for AI, I think for next 30 days, I look at three things from at least from marketing perspective. So there is education, there is engagement and there is energizing the customer.

So we look at three things when we do anything. At this point of time, mostly what we do and I feel AI does really well is the education space. Any information, anything so for next 30 days to three months, I would recommend that if anyone and at least we are also doing the same. Leveraging AI chatbots just to give information. And for us as airline, there are massive amount of information people need. What is the flight on time? Which gate? How much baggage can I carry? What are the rules for infants? Et cetera, et cetera. So these are very simple FAQs, something which is very, which you don't need humans to talk about. We are leveraging that such information for AI for such information. Simple basic stuff, very difficult to make errors in this.

AI can simply reply to them. Obviously, when cases become complex, they move to a live agent and they leverage it. Second thing I think for which we are leveraging AI and I think most of should do is primarily what also Grant mentioned is efficiency. We, and you rightly said, we also have a massive call center because we fly a free two and a half lakh passengers every day. So we have a massive call center. We can't really take everyone's call every day. So what this has done at least for us is, and when we looked at all our data, we realized 60 to 70% of our queries, whether they were on social, email or calls, were just information. People were just calling up, tweeting us. Just want to know is my flight on time? What are the baggage rules, etc.

So what we have done is we moved all of that under the whole chatbot engine on our website, WhatsApp, etc. Remue brought in massive amount of efficiency. And as I said, when it moves into sort of any complex cases, the humans then start taking over. And third, which is what also we are doing. And I think it's interesting, a lot of marketers have working with chat GPT at this point of time to figure out, oh, what is the new caption that I use? What can be that post copy and what is that new email heading that I use? I think a lot of good things for your content writers to at least get initial framework for whether it's creating an SEO blog, whether writing that perfect email headline, etc.

I think it just gives them a good framework then to build on top of it. But I think one of the things that we actually even tried working with a partner, I think where AI currently is not there is this whole concept of tone of voice, right? We Indigo as a brand have a very specific tone of voice. We talk to our customers in a very specific manner. We have a style and tone and we are very proud of it. Unfortunately, AI doesn't give you any of it. It just gives you a very basic construct. That's where then you take that basic construct, the content writers put the brand tonality on top of it to create something. Hopefully, AI will reach a point where it is able to create something which is very unique to a brand in its own tonality.

And you don't really have to then do all that human intervention. I just like to share one example and join in before I move on to this next question because it's an Indian Jugaad. And Grant, for your interest, the Jugaad is the quick fix way to do something in a very cost efficient manner, yet the same results. I was talking with my dear friend Amit Sarra of Soulflower and he permitted me to share this example and therefore I am sharing it. He hired about 100 interns across India. If you go on to ChatGPD and ask about Soulflower, his brand, which is into oils, aromatic oils and things like that. ChatGPD is more learned about his brand and his tone of voice. So it was forced learning.

Every time ChatGPD, they would ask a question to ChatGPD, what Soulflower? The interns would prompt the correct answer. Indian Jugaad? Of course there is model training, APIs and he understands it all but he used brute force till the time it was open and available. No, just to add to that Vassar because you asked about 30 days but that's something that we are doing. Hopefully in six months, we are doing the same. Taking the ChatGPD source code and now we are putting in data to train it to hopefully be able to take out something which has our tone of voice because at the end of the day, that plays a big role, right? Because if we remove our tone of voice, then it's any airline, right?

What's happening, conversely, and that could be my personal very strong point of view. I remember as a child, my grandmother, because mom used to go to work and I was supposed to memorize tables and Gujarati household is even more. So for example, the table is not allowed to be 20 by heart and it was not allowed till 10, you had to multiply till 12. So what is 12, 12 or 144, you had to be able to rattle it out. So when calculators came, you lost 25% of it. When Excel came, you lost the balance 25% of it. After that, another 25. So today, if somebody asked me, even though I am an enthusiastic stock trader at times, it becomes difficult to compute because we have the basic method to compute and derive the answer faster.

On ChatGPT, I feel it's so rhetoric and repetitive when you try to get answers or questions or the tone of voice. For every marketing email, I ask it to write for this thing. It will start with unleash your marketing or unlock this thing and it will say stay ahead of the curve. And end with something like that. It's just very semantical and very similar. So it's tough. So the other thing that's there is what I get happy about every time I'm reading, me reading the same rhetoric from AI is that human intelligence is not going to die. We may not really have a CEO as AI as a CEO boss. I'll see you later. See you later.