Why You Should Spend More Time Thinking About Paytm?

One97 Communications, the parent company of the well-known financial services and payments startup Paytm, has started a new advertising campaign to highlight its payment offerings and user base to increase current user engagement. "Our ad campaign reflects our commitment: every time India plans its future, we’ll be at the center of it, and as you move forward in life, Paytm will always be by your side," stated the business. The commercial films are scheduled to premiere on television, digital platforms, and social media.

We have a wingman for you: Paytm Assistant

It goes without saying that we require more than just individuals to handle inquiries for a business on the scale of Paytm. Because of this, we started the difficult process of developing a solution for a user base that numbered in the millions at some point last year.

In the preceding cycle, the user selected the most appropriate option from a half-dozen prepared selections in the customer assistance window. There was a subset of alternatives for each of these generic options. Before the inquiry was sent to a human agent, there were just two tiers of questions.

Mohit encountered an issue, for instance, when he was recharging his phone. He would encounter two options when he went to the customer service window.

1. Defining our user

Three different user types have to be taken into account: agents, customers, and product/process teams. The customer's motivation is obvious. The customer assistance section exists because of them.

We required an assistant to help agents become even more productive. Since these agents are human, mistakes and delays are inevitable. We required a technological framework that reduced the possibility of mistakes while improving the working environment.

We conducted some moonshot brainstorming for product teams. Every week, Paytm as an organization undergoes changes. We regularly introduce new verticals, and the range of tasks available on the app is always growing. Therefore, the customer service needs to be flexible enough to handle this extensive

2. Choosing the bot

The second challenge we had to tackle was determining the kind of bots to use throughout the company. At the moment, the industry uses two different types of bots: rule-based and conversational AI-based.

Consider how Alexa interprets your orders to get an idea of how an AI-based conversational bot system works. The system extracts pertinent information related to a question by identifying specific phrases from sentences in the query. For instance, when a user asks, "Where is my order?" or "Why is my order delayed?" the system will recognize the word "order" and retrieve data from the user's past purchases.

3. Defining flows for every query

As an illustration, consider this: User A opens the customer support window and chooses the order that was placed most recently. This signals the system to display frequently asked questions to the user. Assume that the user chooses the "Where is my order?" option because he has a question about the status of his delivery. This is an additional cue for the system to display the order of the travel, by the rules that we have specified. Rules about delivered and pre-delivered orders will differ. Now, based on the selection of option A, the system will consult the backend and present the user with the choice to follow the order in real time — all without involving a human.

4. Futureproofing customer support

Considering this wide (and getting wider!) range of tasks, it may seem daunting to have to build these many rules for every single query that a user may have, but our years of being in operation held us in good stead. This was not built from scratch, we built on the scores of queries we had accumulated over the years. This comprehensive database was the foundation for adding multiple daunting levels and layers of “conversational” elements.

5. More agency for agents (and beyond).

On the agent side, the same bot architecture is operational. It is normal for a human agent to occasionally be unable to provide a user with the exact solution they need. In these situations, the agent's problem is assisted by the bot infrastructure. The bot acts as the agent's assistant, doing the grunt work of gathering all the information in one location so the agent can act swiftly.

The Paytm Optimizer- ASG Module

Although EC2 spot instances are less expensive than ordinary ones, they carry risk and are only permitted in Auto Scaling Groups. the potential for abrupt termination or interruption by AWS at any time. Spot instances are perfect for hosting micro-services that are fault-tolerant and stateless. An intermediary platform is needed to increase these instances' dependability. This platform makes use of the data from hundreds of additional spot instances from all the onboarded accounts and guarantees correct infrastructure configurations. The platforms anticipate the disruption early enough to make the required infrastructure changes for a smooth transition from one machine to a newly deployed machine based on the data gathered. This intermediary platform, the Paytm Optimizer, was created to guarantee the greatest degree of dependability in discounted/Spot virtual

Other services offered by PayTm

Rather than consolidating all of their products onto the PayTm app, they need to have grouped related products across several apps. As an illustration, consider their travel options, which let customers reserve tickets for buses, trains, airplanes, hotels, and taxis. PayTm ought to have combined all of the aforementioned services into a single app called PayTm Travel. Including PayTm Travel (booking), PayTm Money (stockbroking services), PayTm Mall (e-commerce vertical), and PayTm (the main app which would offer payment services, recharge & bill payments, and add financial services to the main app like offering loans, credit card, insurance), they now have a total of 4 consumer-facing apps. I've decided to incorporate financial services within the main app for two reasons:

Conclusion

These days, any payment platform company can simply offer large cashback to attract customers and win them over. This is why, in contrast to other businesses, operating a fintech company is even more difficult because you need to offer more to customers than just a compelling value proposition. After all, they may be enticed to use another app in exchange for cashback on a previous purchase. I had this crazy concept, but I'm not sure what value-added services could pique people's interest and make them devoted.