How Gupshup Uses AI

Product innovation and development is driven in-house allowing speed, flexibility and scale.

Gupshup is a conversational messaging platform. Beerud Sheth founded Gupshup in 2004. In July, the company raised $240 million in follow-on funding from an industry-leading group of investors at a $1.4 billion valuation.

Analytics India Magazine got in touch with Beerud Sheth, Co-founder and CEO, Gupshup, to understand how the company uses analytics and machine learning for a better customer experience. An alumnus of IIT Bombay and the Massachusetts Institute of Technology (MIT), he is a serial entrepreneur and investor with over 25 years of experience in the tech space.

“While conversational AI, on the one hand, helps the brands understand their customers, their preferences, their requirements and their choices, on the other hand, it provides the most efficient platform to the users to interact with the brand on their choice of the channel at their preferred time, language and availability,” said Beerud.

AIM: How does your company implement AI/ML/Analytics to drive growth? Discuss briefly by citing customer success stories or use cases.

Beerud: Gupshup implements AI, ML, and analytics to understand and process human language inputs and respond to them in a near-human manner with the right context and smooth flow. We have predefined verticalised and industry-specific templates with domain knowledge to ease building intelligent, interactive bots. 

With a full-fledged bot deployment, we have automated the complete customer support experience for our clients. 

Citing an example for Panasonic: We have reduced the agent handover by ~40% and the manual requests for registration, address change, and other similarly mundane tasks by over ~20% in the initial deployment days. 

We have seen a significant increase in customer experience, more leads generated and more efficiency across marketing and sales verticals across all industries. From Kotak Bank, ICICI Lombard, Bajaj Housing Finance Ltd, DBS, Khan Academy, TVS Credit, Lakme, to Zomato; Gupshup powers <> conversations every day over <> clients.         

AIM: What percentage of resources are focussed on implementing state of the art? 

Beerud: We have a team of 25, including full-time employees and contractors as engineers, data scientists, developers, AI leads and experts. We are also on a continuous hiring spree to grow our AI team.

AIM: What kind of skills do you look for while hiring data scientists/ML engineers?

Beerud: Major skills that work in the conversational automation space are :

  • Strong understanding of linear algebra, optimisation, probability, statistics.
  • Experience in data science methodology from exploratory data analysis, model selection, feature engineering, deployment of the model at scale and model evaluation.
  • Understanding of latest NLP architectures like transformer models(BERT, BART, T5)
  • Experience in adversarial attacks/robustness of DNN is good to have.
  • Experience with Python Web Framework (Django), Analytics and Machine Learning frameworks like Tensorflow/Keras/Pytorch.

AIM: What is the major challenge when it comes to recruiting Indian talent?

Beerud:  Major challenges include:

  • Lack of academic institutions offering AI/ML as a specialisation.
  • High attrition due to increasing demand and short supply.
  • Lack of India-specific AI models and engines that are equipped to deal with multiple languages and dialects.
  • Lack of tools to assimilate India-specific learning.
  • Limited understanding and perception about AI and ML with the consumers.

AIM: What are the predominantly used programming languages by your data science team? What does the toolstack of your team look like?

Beerud: Python and Java are the prominent programming languages our data science team employs.

Our tool stack includes Pytorch, MLFlow, ElasticSearch and Airflow.

AIM: What kind of AI/ML deployment challenges does your team face?

Beerud: Challenges we face include:

  • Hosting and serving large models at scale. 
  • Data cleaning and management. 
  • Annotations and training data preparation. 
  • Model life cycle management.

AIM: Do you prefer having your own in-house R&D team or outsource the innovation?

Beerud: Gupshup prefers having product innovation and research driven in-house. However, Gupshup continuously supports and sponsors academia in premier institutions across the globe. The need to sync innovation and product release cycles are pressing now more than ever. Product innovation and development is driven in-house, allowing speed, flexibility and scale.

AIM: How do you see the landscape of AI/ML evolving in India with regards to your domain?

Beerud: AI/ML play a vital role in evolving the relationship between the brand and its customers across the globe. While conversational AI, on the one hand, helps the brands understand their customers, their preferences, their requirements and their choices, on the other hand, it provides the most efficient platform to the users to interact with the brand on their choice of channel at their preferred time, language and availability. 

Brands have realised now that Conversational Messaging with the right mix of AI/ML tools and techniques has the power to completely redefine the entire customer interactions across marketing, commerce and support, and are using this to their advantage through vendors like Gupshup. To provide a better customer experience, brands need to deploy Machine Learning (ML) algorithms and Natural Language Processing (NLP) capabilities to automate mundane, repetitive tasks by driving insights from huge amounts of data, evolving the bot to understand and automate. 

More Great AIM Stories

kumar Gandharv
Kumar Gandharv, PGD in English Journalism (IIMC, Delhi), is setting out on a journey as a tech Journalist at AIM. A keen observer of National and IR-related news.

More Stories


8th April | In-person Conference | Hotel Radisson Blue, Bangalore

Organized by Analytics India Magazine

View Event >>

30th Apr | Virtual conference

Organized by Analytics India Magazine

View Event >>

Yugesh Verma
All you need to know about Graph Embeddings

Embeddings can be the subgroups of a group, similarly, in graph theory embedding of a graph can be considered as a representation of a graph on a surface, where points of that surface are made up of vertices and arcs are made up of edges

Yugesh Verma
A beginner’s guide to Spatio-Temporal graph neural networks

Spatio-temporal graphs are made of static structures and time-varying features, and such information in a graph requires a neural network that can deal with time-varying features of the graph. Neural networks which are developed to deal with time-varying features of the graph can be considered as Spatio-temporal graph neural networks. 

Vijaysinh Lendave
How to Evaluate Recommender Systems with RGRecSys?

A recommender system, sometimes known as a recommendation engine, is a type of information filtering system that attempts to forecast a user’s “rating” or “preference” for an item. In this post, we will look at RGRecSys, a library that performs constraint evaluation of recommender systems.

3 Ways to Join our Community

Discord Server

Stay Connected with a larger ecosystem of data science and ML Professionals

Telegram Channel

Discover special offers, top stories, upcoming events, and more.

Subscribe to our newsletter

Get the latest updates from AIM