Engagement

33:04

Leveraging Technology to Scale Personalised Customer Engagement

Moderator – Sweta Duseja, Director, Customer Success – META, MoEngage

Speakers:

Nitish Jalan, Director of Growth, Swvl
Rohan Kapoor, Digital Marketing Director, Careem
Abdulaziz Algain, Senior Growth Executive, Altibbi

03:09 – Role Of a Unified Engagement Platform In The Use of AI and ML

Moderator – Let me just kick it off with a question. I am going to ask all of you. What role does a unified engagement platform play in the use of AI and ML? What does it help you deliver?

Nitish Jalan – Okay so probably I will start on this and honestly in the current world how could companies talk to their customers easily, know more about their customers, and engage with their customers on a very regular basis? I think in that world, a unified platform such as MoEngage itself helps us a lot in terms of multiple factors like data-driven customer insights, automation, journeys to reach out to customers, and smart segmentation.

I think these are three things that play a very important role in defining how we actually talk to our customers, how we ensure that we are engaging with the customers on a regular basis, getting customers to know the kind of services that we are providing, and ensuring that they come back to the platform, they engage with us, their user service, and continue to be with us. I think that’s the most important thing that a unified platform like MoEngage helps us to deliver on a day-to-day basis.

4:43 – What Is Customer Engagement?

Moderator – What is customer engagement? Define that even for us.

Rohan Kapoor – For us, for Careem, customer experience is the top priority. To enhance customer experience at Careem, we know that we need to build best customer journeys. Along with that, our communication across channels needs to be personalized and it needs to happen in real-time, based on real-time customer data. This is where I say that a unified platform, a customer engagement platform really helps us do this and at Careem we have seen some great results when we have this customer-centric personalized journey planning strategy in place. It has really worked for us and we heavily invest into all the ML that goes behind and all the rules that go behind making the strategy work for us.

5:56 – Role Of AI and ML In Healthcare

Moderator – Mr. Algain, what do you think from a patient’s perspective? How much of a role do AI and ML play in healthcare as opposed to commercial activities?

Abdulaziz Algain – Yeah, definitely. One of the main advantages that Altibbi has is that we are exclusively an Arabic platform. Though we deal with different dialects, one of the challenges we have is that when it comes to consultations, they can come from Jordan, different parts of Saudi Arabia, Egypt, or whatever. Without ML or AI, it is very difficult to have a unified diagnostic system the way we help doctors as well as give treatment.

So Altibbi has worked on something we call Altibbi Vec which is a way for us to vectorize different terms. In Arabic, we have different names. We use Altibbi Vec to help different doctors from different regions understand what the customer is saying so that from the earlier part already AI is playing a big role.

Moderator – And what about from an ongoing customer page because you definitely do want to necessarily mark it but how do you think it goes from a customer engagement. Is it primarily playing with the information provided to the doctors to have a better engagement with the patient?

Abdulaziz Algain – Yes, for sure. For example, when it comes to healthcare startups, especially telemedicine, we can’t just send push notifications, talk to a gynecologist today, or talk to an eye doctor today. It’s different than logistics or delivery apps. It’s mostly around following up with what happened with them and trying to give them something that makes sense based on what problem they have because it is not something that people consult on a periodic level. It is not something that people don’t want to talk to doctors about every week but we are trying to work on wellness programs, different things, and future integrations to make sure we can integrate that part and hopefully have MoEngage help us in that kind of thing.

Moderator – #ShamelessPlug Rohan, talk to us about two things. Explain to us what the super app is because I am sure when people start thinking about it. Gone are the days of acquiring brick-and-mortar to establish dominance across multiple industries or verticals. Now it’s being done via the app. Talk to us about what a super app means and how you think AI and ML can help supercharge multiple apps within the app. What beats that requirement?

Rohan Kapoor – So the first question is, what’s a super app? What we are trying to build at Careem is basically one big utility app that does it for you and is out of your way. We started off with ride-hailing and we did it very well. During COVID, we realized that we have a vision of becoming a super app which means we have multiple services, we want to help our users in more than one way and so now, in more than six-seven markets, we are a super app. What it means is that in Dubai, you see at least 15-20 services that you can avail within Careem.

That is, you have ride-hailing, you have food delivery, you have groceries, we also have a fintech product called Careem Pay and on top of that, we also have a subscription model which gives you value if you subscribe to this model. I am not going to talk much about that. I am not going to market this over here but how AI and ML are working for us is very interesting.

Careem has been in this region for the last ten years and we have millions of people who have installed Careem in this region and so we are sitting on big data, very rich, heavy data. As we are progressing and moving forward into our super app vision, launching it in new markets and most services in new markets, we are monitoring the behavior of our users in real-time and how they interact with these new services. For example, we understand how do they discover new services and when they discover these services. Based on that we have an engine which is our proprietary engine that is working on ML and spits out whether a user is going to graduate into something that we call a multi-service user and whether they do it or not, that insight instructs our marketing strategy.

Why do we need this multi-service user? It’s very simply put. When we were just ride-hailing, we just had only one service. Careem is a platform which has multiple services. One acquisition, you increase the customer’s lifetime value if you are able to sell them across different services.

12:26 – The Growth Experience of Leading Swvl

Moderator – I wanted to ask you based on what your experience has been leading Swvl with this kind of growth. Give us some use cases that you see, a clear play in terms of we have deployed this, it’s working and it’s powered by AI and ML.

Nitish Jalan – I’ll give you probably a very classic example. Just going before that I will give you more context about Swvl is. Swvl is exactly between what an Uber or a Careem is in the region and what our public transportation is, right? We are exactly between that. We are powering both sides of the world but at the very sweet spot of being appropriately priced for the users who are using public transportation and also for the users who are using Uber and Careem but also giving affordability at the same time to both the segments of the users at the same time.

Obviously, we are mass transportation and hence, not giving the same level of comfort that a private will give you but we are almost there in terms of giving comfort with the seven-seater, thirteen-seater, and fourteen-seater kind of bus. What we have done is to ensure that we are at the right price point because that is the most critical thing for us as we need to be very close to public transportation. What we have done is we have built an amazing demand-responsive pricing engine that is completely powered by AI and ML. We have a bunch of data points sitting behind empowering those tools.

What this is doing? This is something very complex in comparison to what the airline industry has in pricing but very simple from a customer-facing perspective, right? Customers see very simple pricing. If there’s a search they’ll see that this kind of a search exists. They’ll know that okay the search is there because there is only one seat left and so on and so forth. For a customer is very very simple but at the back end, we are even more complex than what the airline industry is in terms of pricing.

We have a proprietary app. It’s very complex. We have done a lot of research on our end. We have created a bunch of white papers against this type but it’s amazingly working for us. Giving customers the right pricing, and the right value; giving us the right kind of monetization opportunity wherever there is high demand, and then helping us to serve multiple countries and multiple languages at the same time.

Customer behavior and pricing sensitivity are very different in different markets. I think that this kind of engine that we have developed is completely powered by AI and ML but is helping us enormously in terms of customer engagement and giving the right value to the customers.

15:16 – The Role of AI and ML in Patient Engagement

Moderator – Mr. Algain, we have watched the videos about robots operating us, it scares me a little bit and that’s where true AI ML is going. At some point, I am going to need botox for sure. But where is this going to go from a healthcare perspective? This patient engagement? Where do you see the future trend? Where do you see it leading to and where does it start, where does it stop?

Abdulaziz Algain – Honestly since we are talking about the prediction or the future, it’s very hard to think about the right way to approach something like this because we all in the previous sessions were talking about retention and acquisition. You can make some decisions that may be from a very short-term point of view, push notifications, etc. We get good clicks or get people to buy some stuff.

Altibbi is now starting a drug delivery service as well but since we are talking about the future and what I am very interested in from a personal point of view not as a professional is sequences of decisions. Say, reinforcement learning could play a big part in this and it’s still an area that is not very developed but let’s say that the multi-agent system and the multi-agent reinforcement learning could allow us to only promote things that make sense from a business point of view today but will also keep the customer engaged for a very long time because one of the problems in not just telemedicine but healthcare also is that there is an asymmetry in terms of information between the doctor and the patient.

Sometimes we could doctors over prescribing certain drugs or pushing people to maybe take diagnostic procedures like X-Rays, ultrasounds, whatever. So we could use AI to kind of not only support them in their decisions but also in monitoring and referencing. Where is the point where we are doing too much to the point where the customer feels guilty now. I did this but I don’t want to talk to this doctor or platform or whatever. So, let’s see.

18:02 – Future Of Careem In Terms of AI and ML

Moderator – What about Careem, Rohan? Where is the future of engagement with Careem? Let’s talk about it from a more organizational perspective. Like how is it going to change your team in terms of manpower, woman power?

Rohan Kapoor – Careem has always been very innovative in this area. From very early on, we had this destination prediction model that users find to book rides easier and faster because we already know their behavior, we already know their history. So we are able to bring that up very quickly so that’s really helped. Apart from that in food and groceries, for example, we have a powered recommendation engine for your favorite items or you know similar to your favorite items.

Apart from that on the supply side of things, we have done some innovations which allow us to book riders. We call them captains. But knowing the probability of them accepting the job only to make sure that the customer gets their delivery on time. So these are some of the ways that we do this. In marketing it is core. Machine learning and AI is a core in marketing strategies. I mentioned about recommendation engine that also works for marketing but then again you have, creative targeting, customer lifetime forecasting, you have churn forecasting and all of that is dependent on ML.

We run queries ourselves and we have proprietary kinds of codes. I am not a Data Scientist myself but my team does a really good one. So all of this really helps us and then we use all these insights and put them back into our marketing strategies to optimize across marketing.

20:27 – One-to-One Personalization on Customer Engagement Platforms

Moderator – I have one last question. Nitish, talk to us about is one-to-one personalization actually achievable on MoEngage?

Nitish Jalan – That’s a very interesting question. There are a bunch of things that you can think of from a one-to-one personalization perspective. Nowadays, everyone wants one-on-one communication with customers, telling them what their need is, and so on and so forth. But, the challenges with on-to-one communication as well are that you don’t go to one-to-one communication on day one.

It needs a lot of learning, a lot of experiments at your end to ensure that you are actually communicating the right thing to the customer and you are doing one-on-one communication. You start with segmented communications with the customers. You try to understand what is working and what is not working.

Which type of customer is actually responding to the communication, messaging, the kind of marketing that you are doing for the customer, and then slowly moving towards one-on-one communication? And also, with one-on-one communication, there’s always a risk and fear that you can always over-communicate which can frustrate the customer a lot, can drive the customer away from the platform, and so on and so forth.

You need to be very very cautious when you think about one-on-one communication. You need to have a path towards one-on-one communication through your segmentation, through your initial quote and targeting, and so on and so forth.

When you’re talking about whether this can be achieved through a platform like MoEngage or any other engagement platform then it can definitely be achieved. It’s just that you need to very very cautious, you need to know what you want to communicate at what time, what language, and to which customer. You need to know that very well and that has to come with a lot of experiments, a lot of data analysis, and machine learning algorithms to understand that this is what I need to communicate, this is what the customer will want to listen to and this is if what if I communicate then the customer will go away from the platform.

Moderator – So basically, when you are looking for technology, you should look for things. You know, of course, I think AI and ML were add-ons, it would quite a bit of money for all this. It has almost become something that you cannot invest in without it adding its own flavor of AI and ML.

You might also like...