Indian AI Startups – Comparing 4 Companies – Applications, Teams, and Fundraising

Ayushman Baruah

Ayushman is an experienced journalist covering innovation and technology in India's startup ecosystem.

Top 4 AI Startups in India - Comparing Their Applications, Teams, and Fundraising 1

According to research firm Zinnov, India is the fourth largest geography globally in terms of AI talent pool. Zinnov’s research says India is home to more than 200 AI startups that attracted about $200 million+ funding in 2017. A few major companies that are implementing AI include Amazon, AIG, Equifax, Infosys and NVIDIA – all of which we’ve covered in depth in our recent “AI in India” research.

Atit Danak, Engagement Lead at the formerly cited research firm, Zinnov, says:

“At its core, AI promises an opportunity to lower the cost of delivery of solutions which the masses need but cannot access. …  We have very interesting work happening in logistics, retail and commerce, industry 4.0, financial services, and healthcare, already from India. A lot of these solutions are global in nature and are witnessing positive customer traction.”

In this report, we’ll note the top four funded startups based originally in India using AI technology based upon funds raised and listed in descending order.  These companies were identified by startup intelligence firm Tracxn, based on their research.

While three out of four of these companies now have a US presence through a second headquarters or work with prominent US clients, all companies listed were founded and still have open offices in India.

Gupshup

Product Summary: Messaging and bot development platform

Funding Year: 2006

Funds Raised: $47 million

Funding Round: Series E  

Gupshup.io is a platform used for building chatbots and conversational experiences. The platform includes a variety of bot-building tools that enable users to create bots quickly. This includes a coding tool for developers (called IDE Bot Builder), a graphical tool for designers (Flow Bot Builder) and pre-built templates for business users (Template Flow Builder). The company claims to provide tools to help with deployment, hosting, publishing and monitoring, along with analytics for its chatbot applications. It is built to scale on top of Amazon AWS. It claims to handle over 4 billion messages per month and over 150 billion cumulative.

“Because many of our bots are conversational, we use a lot of Natural Language Processing (NLP) tools to engage in a human-like conversation,” Beerud Sheth, co-founder & CEO, Gupshup told Emerj. “We also use AI for image recognition in cases like an accounting chatbot where the bot needs to detect and extract data from the photograph of an invoice.”  

Sheth adds that unlike plain-text messages, Gupshup says its smart-messages contain structured data and intelligence, which could enable advanced messaging workflows and automation.

“Gupshup’s smart messaging platform is also highly scalable, available, and reliable with intelligent message routing capabilities.”

This 2-minute video demo shows what Gupshup looks like to a user when it has been implemented:

According to GupShup’s website, it targets solutions to marketers looking to engage new customers, sales teams aiming to suggest products to gain higher purchase conversion rates,  and support teams who may need assistance with troubleshooting issues with multiple users.

In a case study, Gupshup reveals that they created a message-based “deals” portal, where users could get deal suggestions for Naaptol, an online Indian eCommerce site. Gupshup says that, since building the platform, Naaptol’s portal has had over 1 million “deal-seeking” visitors. The case study did not not the amount of time that Gupshup was implemented before reaching 1 million visitors.

According to its homepage Gupshup’s other clients include McDonald’s, Twitter, WhatsApp, LinkedIn, Facebook, Kellogg’s and Dish TV.

Gupshup has grown at a CAGR of over 75 percent in the last 4 years. It claims it has captured over 40 percent market share of the Rs 10-billion Application-to-Person (A2P) messaging market. Some of Gupshup’s customers include FlipKart, OLA, Facebook, Twitter, ICICI, HDFC and ZeeTV. Gupshup claims to help over 81 million users across 27,000 businesses, handling 3.6 billion mobile interactions per month.       

Gupshup is funded by investors such as Charles River Ventures, Helion Venture Partners and Globespan Capital Partners.         

NetraDyne

Product Summary: Provides deep learning-based computer vision for drone and automotive navigation

Founding Year: 2015

Funds Raised: $16 million

Funding Round: Series B  

NetraDyne claims to be the first commercial vehicle technology provider to combine AI with video to detect, reason and determine causality of events, providing commercial fleets with timely meaningful data. Founded in 2015, Netradyne has established technology innovation centers in San Diego, California, and Bengaluru, Karnataka, India.

The company has developed a vision-based driver safety platform known as “‘Driveri” that is deployed in commercial vehicles.

According to BBC, while India does not allow driverless cars, Driveri uses similar technology in attempt to prevent regular accidents that may happen in the country. A user can place a Driveri device, with built-in cameras and sensor technology, to their dashboard. When I driver is too close to another vehicle or may be doing something dangerous themselves, Driveri beeps with an alert. The device can also show the driver a video of their dangerous behavior or of what objects are close to them immediately following the incident, according to BBC.

The company also claims that the Driveri device can also collect computer vision-based data that is used to generate dynamic 3D, HD maps for autonomous vehicles.  

This short video from BBC Click follows a reporter in the car with a driver using the device:

Avneesh Agrawal, founder and CEO of Netradyne, shared:

We have disrupted the legacy video-based fleet monitoring systems that required constant monitoring to identify actionable events. Our systems eliminate costly, inefficient business processes and encourage drivers through positive reinforcement. We see great success of our products and an exponential demand as the word gets out and drivers become our evangelists.”

In December 2017, Netradyne launched an advanced driver assistance system (ADAS) solution trained for Indian road conditions as part of its Driveri platform. “With its launch of ADAS product recently, it is leading the path to autonomous vehicles in India,” Agrawal said.

Clients include Vascor, an Indian transportation company; Load One Transportation and Logistics; and trucking company W.N. Morehouse. Case studies for Driveri could not be found on its website.

Netradyne is primarily funded by Reliance Industries. Investments are made through Reliance’s subsidiary, Reliance Industrial Investments and Holdings Ltd (RIIHL).     

Belong.co

Product Summary: AI-enabled candidate sourcing platform

Founding Year: 2014

Funds Raised: $15 million

Funding Round: Series B  

The founders of Belong.co said they shared a common challenge – that of hiring talent to build teams. As they spoke to several talented but discontented people, they said it became clear that there was a mismatch between passionate people and the places where they worked. This stood in the way of them finding satisfaction in what they did and unlocking their true potential. That’s the problem Belong.co is says it is solving.     

Tavant Technologies chose Belong.co to help them scale their team.

Belong.co’s talent search suite, Belong Hire. claims to have saved 17 hours per recruiter each week, translating to $100,000 saved weekly. It consists of various tools, including:

  • Curation Engine: Curates talent data sourced from 90 different professional networks, communities and forums to offer hiring managers an untapped pool of ideal candidates
  • Adaptive Search: Uses machine learning and predictive analytics to uncover people similar to the most successful candidates based on the company’s sourcing and interview patterns
  • Belong Pick: While Adaptive Search uses historical sourcing and hiring data, Belong Picks analyzes current and past employee data to identify talent who are most similar to the high-performing employees.
  • Job Description Scanner: The company claims this automatically understands a job description and generates relevant, skilled candidates.

Exact role titles are not clear on the company’s team page, however, Belong.co’s Data Science representative, Devashish Deshpande, has received a Bachelor’s degree in Computer Science from the Birla Institute of Technology and Science, Pilani. One of its Tech representatives, Kanishka Pahwa, received a degree from the Indian Institute of Technology and has held software engineering internships at Amazon and Symantec, according to his LinkedIn page.

Belong.co clients include Adobe, Accenture, PayPal and Cisco.

Its customers page also includes testimonials from clients including Adobe India’s Head of Talent Selection, Prasad Rao:

“Belong has helped us identify great senior engineering talent. They are a key partner in creating and executing our pan-India diversity hiring plan, as we work towards our goal of becoming the most diverse workplace in the country.”

Belong.co has also released a case study on recent work with Tavant Technologies. While the case study does not go into detail about the time period or previous numbers, it notes that Belong’s software led to a 20 percent decrease in the time employees needed to search for the best candidate.

In a more detailed case study, Michigan-based Altimetrik, a digital solutions and consumer technologies company, came to Belong.co after issues hiring Full Stack Engineers. The case study adds that Amtimetrik also wanted to create and scale a different candidate experiences and technical evaluation, in order to measure and “engage the best-fit candidates for these roles.”

The company used Belong.co’s customer success manager to evaluate compensation trend reports, that would assist them in planning a hiring process. They then used Belong.co’s automated services to email and keep contact with candidates throughout the process. Altimetrik also used other Belong.co features to identify fitting candidates for current position openings that may have declined a previous position offer. While this case study similarly didn’t mention how much time the product Belong.co used for, it did note that Altimetrik’s was able to make 36 offers following job interview processes using the platform.

Belong.co is funded by investors such as Sequoia Capital, Matrix Partners India, Blume Ventures, Kunal Bahl, Rohit Bansal, Phanindra Sama, Raju Reddy and InnoVen Capital.    

Unbxd

Product Summary: Product Search Software for e-Store

Founding Year: 2010

Funds Raised: $14.5 million

Funding Round: Series C    

Founded in 2011, Unbxd is an eCommerce product discovery platform that claims to apply advanced data sciences to connect shoppers to the products they are most likely to buy. It does this through several functions, namely: intelligent site search, personalized product recommendations, and real-time, predictive merchandising insights.

According to the company, the software “tokenizes” each search query. It then maps it to the most relevant products and desired attributes to showcase relevant results that understand the shoppers’ intent; even when they use long tail queries.

Unbxd reportedly serves 1,200+ online retailers as clients in over 40 countries, handling over 1.5 billion interactions each month. These clients include Ashley Homestore, Cabat Lane and Burkes Outlet.

Unbxd says it also profiles each shopper in real time which may improve conversion rates and improve customer loyalty. Unbxd adds that its predictive Visual Guided Autosuggest shortens the path to purchase, right from the search box. It delivers suggestions that the platform estimates has the highest likelihood of conversion through predictive analytics, and machine learning algorithms.

Unbxd has a handful of short tutorial videos on their YouTube channel, but the volume is unusually low and the videos are hard to follow along with for this reason.

The firm does have a number of webinars with better audio and a more in-depth look at their products and case studies – the video below is a case study highlight webinar with Unbxd and their client, U.S. Patriot Tactical:

In a case study with The Kewl Shop, the retailer was looking to add a stronger customer recommendation feature to its site. Prior to using Unbxd, The Kewl Shop’s website used a recommendation feature which did not analyze user behaviors, such as purchases or viewed items.  According to the study, this recommendation platform was limited to some products. The company wished to highlight more of its product lines at once.

The study notes that by adding Unbxd’s “You Might Also Like…” widget to product pages on its website, The Kewl Shop was able to speed up a customer’s purchase by 1.5x. According to Unbxd, the widget also collected customer data that could be seem on the retailer’s Unbxd dashboard. This dashboard also allowed web managers to customize aspects of the widget, such as the number of products shown.

This case study did not note how long the product had been used or more specific numbers behind the results.

“Realizing that it was actually easier to find products on Google than on the websites offering them, we developed a product which makes the experience of finding products on e-commerce websites a pleasant one through,” the company said on its website.

Unbxed is funded by investors including Eight Roads Ventures, IDG Ventures India, Inventus Capital Partners, Nirvana Ventures Advisors, Indian Angel Network, Orios Venture Partners and Singapore Angel Network.        

Common Themes Among India’s Well-Funded AI Startups  

These top four startups in India are using several different AI functions, namely messaging and bot development, drones and automotive navigation, candidate sourcing in human resources, and product search in eCommerce. One common factor across these startups is that they are all based in the top metropolitan areas of India. While Netradyne, Unbxd and Belong.co are based in Bengaluru, Gupshup was founded in Mumbai, the financial capital of India.     

Although all companies on this list were based in India, Unbxd, Belong.co, and Gupshup have gained a US presence through work with clients or new office locations. While Belong.co’s offices are still in India, they have worked with American companies including Adobe. After raising $12.5 million in Series C funding, Unbxd opened offices in Mountain View, Calif. and Oak Brook, Illinois. According to its contact page, Gupshup has also opened a US headquarters to San Francisco, Calif., with American marquis clients including Facebook, Twitter and Ebay. It is not clear of when Gupshup opened its US office or what level of funding the company had reached at that time.

While these startups are not founded by leaders with a robust career or academic background in AI, they hire teams with educational expertise or some experience in the field, which may gain them further credibility. Danak of Zinnov told Emerj:

“As an offshoot of their experience, growing AI startups [in India] tend to focus on building processes and algorithms to expand data sources, expand raw data sets, and handle unstructured data. In an Indian context, and based on the startup activity, we do believe that teams with a narrow focus but solution depth, have the highest odds of success, given the global competition for AI talent.”

While the focus areas of the top-funded startups are an indication of where the money is, India holds a lot of potential for building niche solutions that may be a low priority in other geographies. Per Zinnov, some of the sectors where India has tremendous opportunity are agriculture, healthcare, education, financial services, logistics, and energy. 

 

Header image credit: talenteconomy.io

Subscribe