The evolution of conversational AI has marked a paradigm shift in customer engagement. The emerging technology has enabled machines to decode human language, analyze its context and generate a response that’s easier for humans to understand. To get a sense of how new-age entrepreneurs are viewing this development, we reached out to Rahul Mehra, co-founder of Delhi-based logistics automation platform Roadcast.
Mehra believes conversational AI is the only way in which enterprises can acquire scale. He attributes his conviction to the widespread availability of smartphones and cheap internet.
“If I’m running a big company with lakhs of users, it would be very difficult for me to set up a network of support agents who can help customers in real-time while ensuring timely delivery of services. Considering India’s huge population and post-pandemic digitization, AI is the only way in which we can scale up,” he adds.
In an interaction with Ozonetel.com, Mehra pointed out the crucial role played by AI in customer engagement and added how emerging technology has taken off the burden from human agents, thereby allowing them to focus on high-priority tasks.
How can organizations leverage conversational AI to improve customer services?
If we see all the major apps that we’re using today, right from Uber to Zomato to Swiggy, they are leveraging conversational AI for customer engagement to great effect. 10 years back when conversational AI was in a very early stage, customers had to wait in queues for long, to speak with executives.
But now if you look at the entire situation, it’s completely transformed. So, calling is probably the last option that customers will use in any of these situations. They would usually go to the app’s query section and use chatbots to raise the query. The AI mechanism has become so advanced that it already has pre-fed answers that can probably solve the customer’s query in the first interaction only. There is no need for human intervention in that case. I would rather say, proportionately, if a company receives 10 queries from customers in a day, out of these 10 queries, only two or three calls are the ones which actually need human intervention, all thanks to artificial intelligence. In our start-up too, since we are dealing in telematics and have an entire ecosystem of delivery boys and riders using our platform, we have a high volume of inbound calls. We have also built a chatbot, which essentially, tries to automate the responses based on what the user is keying in. So, it helps us reduce the number of calls that are coming into the system, and it just creates a better customer experience.
Consumers expect brands to show empathy and seek hyper-personalized services. What are the major challenges to personalization at scale and How can Conversational AI help enterprises in this regard?
The key to customization or personalization is identifying user behaviour. As a service provider, we need to understand what the user is thinking or what is the user’s preferred choice of a product. Till the time we don’t identify those preferences, we will not be able to personalize at scale. When we talk about identifying this kind of information at scale, the only problem is that there are so many threads or data sets that we are dealing with.
In case of a single user too, probably there could be 100 data parameters that should be taken into account to identify whether a user has traits of a particular type of personality and this is probably the kind of content he would like. To achieve a high level of personalization, one would have to compile a huge amount of data and divide users into segments. Manually, it would be very difficult. With AI at our disposal, we can process large volumes of data in a short span of time. Once the user gives an input, the AI bot or the AI algorithm is able to identify that this user fits into this bucket the best. Now, if a human was doing it, the entire process would have been tedious and longer. But, in case of a chatbot or a conversational AI bot, the personalization process can start right from the first answer.
As the co-founder of a start-up., can you share some examples where AI technology has helped you in customer retention or provided them a good customer experience?
Since we are into B2B space, our customers are enterprises or businesses that are using our platform for their operational automation. So, if we talk about e-commerce or hyper-local customers, the biggest driving factor of their success would be whether the deliveries are being done in an optimized cost-effective manner while ensuring that the customer experience is also great. If a delivery company is not giving a good service, nobody would want to order food or groceries from them.
Our current technology helps these companies identify which are those delivery boys that are not performing sufficiently well. Okay, if those delivery boys are not performing, maybe they’re delivering late, maybe the feedback that they’re getting from the customers is not right. So, we’ve built certain AI around this, which helps us identify those users in the system who are not suitable for the business and resulting in poor customer experience. So, currently, we are providing this technology to businesses.
Apart from this, we have enough data that helps us identify the tentative requirement of delivery partners during the festive season. For example, during Diwali, the number of deliveries goes up exponentially. So, the companies need to plan beforehand the number of orders they are likely to receive and the number of riders they will need to deliver those orders. This is something that can be done only through AI, if we leave this to manual work, there is always a chance of a discrepancy and maybe we will not be able to sufficiently handle the demand.
Please tell us about the changing landscape of conversational AI in India in terms of adoption, growth and future possibilities.
I strongly believe that this (conversational AI) is the only way in which the companies can acquire scale, the reason being, that a majority of our population across cities have access to smartphones and cheap internet. [AR1]
So, if I’m running a big company with millions of users, it would be very difficult for me to set up a network of support agents who can help customers in real-time while ensuring timely deliveries. AI is the only way in which we can scale up.
The huge inflow of calls especially when people are working across different geographies poses a major challenge. If I have to set up a call center in the north and the customer is from South India, the language could be a problem. But if AI comes into the picture, then it becomes very scalable. It will help customers get the resolution in the local language also.
So, for a business owner, it makes a lot of sense to implement certain tools, which basically help us minimize costs and enable us to scale up exponentially.
Post-pandemic, due to fragmented workplaces and lack of engagement, we also witnessed a high turnover rate. How conversational AI can help in employee retention?
I believe there are two sides to this. Since I am a start-up founder, I would have a perspective from a start-up’s point of view. For us, I personally feel that in a start-up, innovation is very much dependent on collaboration.
Today, if I’m talking to the people in my team, I’m able to innovate better, and I’m able to gather more ideas together. Due to the pandemic, employees are not able to sit together, collaborate or have a good conversation. But at the same time, if we use AI in this, the technology should be able to bring us together or say, it may suggest that in the last two weeks, the teams have had 10 conversations together, and give us a sense of how much collaboration the teams are having. As a business owner, it can help me increase or decrease the frequency of team meetings and encourage more interaction within teams. So, if I see that somebody is online, it would automatically pop up a message on my window saying that you are working on this project; do you want to have a conversation with them because they are live right now.
This is something which is at a very early stage but as a start-up founder, I would love to have a tool, which helps the teams get together and have more productive interactions.
As per a report by Red Box, the majority of businesses want to use conversational AI data to resolve training gaps and upgrade skills. So how does this technology address the issue of training gaps and allow human agents to tackle high-priority tasks?
I will share a small example here. I enjoy gaming and have played a lot of games on the PlayStation. When we download a game, AI technology basically guides us on reward points and challenges and this trend has been there in the gaming industry for at least a decade. So, in a short span, you will learn how all the controls work, how to cross levels, and the objective of the game. Now, this gamification is available in the form of AI in different industries.
Today, if I’m running a factory, I would need a tool that guides me. This machine needs to be operated in this particular way. Also, there is an entire AR or VR-based training module with AI integration which enables businesses to automatically rate employees based on their performance, whether they perform those tasks accurately or whether they are following the correct steps. In simulation games, when you play a character, you are required to answer certain questions to proceed further.
So, this gamification can be brought into customer support. If I am hiring somebody for tech support, I would need that person to have good conversational skills. They should be polite, empathetic and able to answer questions in challenging circumstances. With the help of AI, we can create those simulated environments and identify whether he is a suitable candidate or not. This would save us a lot of manual work.
What are the requirements for deploying a scalable conversational AI solution that models customers’ cognitive state?
If we are talking about deploying AI that model’s customer’s cognitive state, it has to be backed by somebody who knows how to build that solution. So, human intervention becomes imminent. To achieve this, we would need a good resource that understands how data can be processed, analyzed, and segregated into different buckets.
For us, our AI team is our greatest strength. They have great understanding of business analytics, data analytics, and coding because it’s a very cross functional role. So, that is the kind of manpower you would require to build a very efficient tool.
Secondly, we would require a robust hardware infrastructure to build a good tool. If we talk about companies like Amazon, they are working heavily in this space right from good infrastructure for AI tools to providing auto-scalable server spaces. Also, there are multiple automation tools that are already available in the market. So as a company owner, we need not start from scratch. So, I would rather take a tool, deploy a team on it, and then customize it accordingly.
Suppose, we are gathering data for the purpose of marketing. Can you share an example where we can use that data to improve a product?
SI will share how we are using our AI tools to market our products. We are working with multiple OEMs which are using our services and we provide them with the entire solution for the products. So, whenever a customer purchases a particular type of product; what can be the upselling that the company can do vis-à-vis that product? Let’s say, if I am selling an electric bike today, I would think of aligning certain accessories to improve the product and evoke customer interest? Why would I need to market 10 products in segregation if the customer is not interested? So, obviously, upselling is very much dependent on an AI-based algorithm where you need to understand what the customer is buying.
I will give you another example, we are working with multiple hyperlocal aggregators, and based on the ticket size, we can identify that in this part of Delhi, the customers spend more on groceries or the customers spend less on food, so I will be able to channelize my marketing spends accordingly.
Despite many business use cases, AI-oriented technology continues to face setbacks. They are far from completely grasping the nuances of emotions, gestures and sentiments. What, according to you, are the major barriers to large-scale adoption of conversation AI?
Conversational AI has a very broad spectrum, it starts from a very basic level of AI which can translate into Self Aware AI, a very advanced version of the technology. Here, we are talking about robots that can actually understand human emotions and grow accordingly and then before that there is a model which trains itself based on certain inputs. Although we aspire to go to the self-aware part, the infrastructure that we have currently, the kind of skill set that is required to build certain capabilities using AI, is still at an early stage and the more we advance towards self-aware models, the cost of implementation gets very high. So, the companies need to make an informed decision about whether they want to spend that much money on building a very advanced AI tool or whether they are okay with the basics and this remains one of the biggest challenges in its deployment.
In India, data privacy is a big concern. How can firms deploy Conversational AI to ensure the safety of customer data?
I believe the responsibility of data protection falls on both the company and the customers. Firstly, it begins with customers; a company will do everything in its capability to ensure that data is not leaked or hacked. But these days, we come across a lot of frauds, especially banking frauds where a customer gives out information unknowingly.
I think conversational AI is a good way of informing customers about the precautions one should take to keep their data protected. A fraud posing as a bank employee may call you and ask you to share the OTP or click on a link. But if a conversational AI bot is active, it knows that these people have been sent a new pin and that bot will immediately alert the customer to not give out the pin. So, there are processes that can be built to make the customers aware and make data secure.
Which segments are the fastest movers in the Conversational AI space, and how has it helped them in improving the overall customer experience?
In our ecosystem, e-commerce and hyperlocal seem to be the fastest movers. During the pandemic, we witnessed a massive growth of hyperlocal businesses. Over the last few years, we have developed a tendency to get almost everything delivered to our doorstep. With this transformation, businesses want to drive the entire customer experience through AI. Five years back, the same company, which was using 20 staff for customer support, can now perform the same task with 5 people and the rest can be tackled by AI. So, I believe growth is happening in those verticals where there is a lot of customer engagement and AI is driving that growth.