A homegrown social media platform, ShareChat is India's social networking space founded in 2015 by Ankush Sachdeva, Bhanu Pratap Singh and Farid Ahsan. It has 180 million monthly active users with the average daily time spent being 31 minutes. With the digital space in India booming, it's no surprise that ShareChat has made waves over the past 5 years. They've managed to do this through extensive research and adoption of AI and ML tools. Today, to stay ahead in the digital game, it's important for tech startups to leverage all the available tools to scale the business.

NLP (Natural Language Processing)

ShareChat is designed to be a platform that brings content and social media to non-English speaking users. With Facebook, Instagram and Twitter becoming a major part of the country's lifestyle, the regional population were taking a backseat on social media. ShareChat fixed this by creating the platform in 15 regional languages. It understood the needs of regional India and built a platform that caters to those who prefer conversing in their native languages, especially those who were first-time internet users.

It was a challenge to build a social media platform in multiple languages for millions of users. An AI framework helped ease the process. The primary focus was that the AI model understands text as well as audio-video data since the content on ShareChat revolves around messages, images and videos just like any other social media platform. The framework needed to interpret data effortlessly for the platform to succeed. An NLP (Natural Language Processing) was built by the team for regional languages. It posed challenges since regional languages are not digitized yet like other widely spoken languages. Hence, the Transfer Learning needed to be strong - the ability of the framework to interpret and recognize semantics of languages where data in a digitized format is minimal. The NLP is what ShareChat mastered and that's pretty much what got them more than 180 million users. It's also why ShareChat has an edge over its competitors like Roposo, Chingari and Josh.

Google Cloud

When you're catering to over 180 million users, there's going to be the challenge of data storage. There's the risk of running out of storage and dealing with high-level traffic on the platform. A smart move was relying on Google Cloud to provide a solution. ShareChat uses a number of Google tools to manage data; the app is deployed on Google Kubernetes Engine, data analysis is conducted on Pub/Sub, BigQuery, Cloud Spanner, and Cloud Bigtable, and to distribute high-quality content, Cloud CDN is used. The Google Cloud detects challenges and uses the necessary tool to fix them. This way, the team at ShareChat doesn’t have to keep switching between technologies as and when there's a hurdle. There's a DevOps pipeline to release changes on the Cloud. Bugs can be fixed and new features can be added without writing large codes and scripts for testing. Deployment has become extremely easy since the migration to Google Cloud. The migration itself was smooth because of CloudCover and Google Cloud Professional Services.

The users are constant posting videos, pictures or using the messenger feature on the platform. All this data needs to be managed in real time. When ShareChat moved 120 of its 220 database tables to Cloud Spanner, the platform could seamlessly replicate data in real time across locations. The cost was also reduced by a good 30%. This was an excellent move because when new features need to be added, the team doesn’t have to think twice about the time being wasted in testing and deployment.

AI-Driven ShareChat Ads

Currently, like other platforms, revenue is generated through ads. However, the next step is to showcase ads that will be of use to the users as well as deliver value to the investors. To implement this, ShareChat has connected content creators with brands to create advantageous regional campaigns. To connect with rural customers on the platform, hyper-local content is being curated that will, in turn, enhance the brand value and user experience. With the help of BigQuery and Data Studio, advertisers can strategise campaigns and optimise their spends. Performance metrics and resource usage can also be tracked with the cost breakdown feature of BigQuery.

Conclusion

The timely adoption of multiple AI and ML tools is what has allowed ShareChat to grow into one of the biggest regional platforms for social networking. The tools used for real-time ML predictions and minimal latency for content delivery even in tier-2 and tier-3 cities have helped them scale the business. There are 5 verticals that the team works on, namely feed ranking, monetization AI, content understanding, camera AI, and AI Infra, and it is the consistent improvement in these verticals that has boosted the platform to become what it is today.