The Indore-based startup harnesses Artificial Intelligence, Machine Learning, Blockchain and image processing technology to identify and tag various products, and locations featured in a video.
At a Glance
Founders: Saket Dandotia, Alok Patil and Vinayak Shrivastav
Year it was founded: 2016
Where it is based: Indore
The problem it solves: The platform allows users to shop while they are watching a video
Funding raised: Undisclosed early-stage funding
Are you guilty of pausing a video you are watching online every time you see a person wearing a lovely dress or the perfect shade of red lipstick that you have been hunting for? Then there are those of us who want to own the exact things we see in a video, or visit the locations where a show is shot at. Is it, then, possible to shop while streaming a video?
Indore-based Toch, started by Saket Dandotia and Alok Patil in 2016, has strived to make this a reality.
Toch is an Artificial Intelligence-based video object detection and tagging platform that can detect various categories like products, actors, activities, locations etc, thereby providing tag-relevant information/products from various ecommerce platforms. Saket says,
“Toch’s Machine Learning is the first in the space to provide AI-based tagging with over 85-percent accuracy of the tagged products using its proprietary algorithm. Blockchain helps to keep the identity and maintain sharing logs for content ownership.”
In the year 2014, when Saket and Alok were working on machine learning and object detection for some US clients at the other organisation they ran, they realised they should use AI/machine learning capabilities on some domains that are actually looking for this tech stack.
“We started our research and realised that video is that domain where we should apply our skills of AI and machine learning along with image processing to detect all the possible objects inside the videos and then tag those with the relevant products/information,” Saket explains.
Now, any video uploaded on their platform gets converted into content where shoppable product discovery is made available within a few minutes.
“Tagging millions of videos manually will take ages to convert. We thought that going with AI is expensive during the early days without any user experience and feedback. Therefore, we decided to publish over 40 manually-tagged videos on our platform to gauge user experience, and we were surprised to get a huge level of interest in our beta launch,” explains Saket. This later motivated them to work on machine learning and build a 100-percent AI-based platform for detection and tagging.
Saket and Alok had known each other for over a decade and had worked in a couple of other IT ventures related to offshore software services and consultancy business. Later, the duo got together to start Toch. The idea resonated with Vinayak S, Saket’s friend, who later joined them as the third co-founder.
While Saket is a B Tech graduate from Indore and runs a mobile and web solutions consultancy company, called Linkites, Alok is a computer science graduate and also a partner at Linkites. Vinayak is a mechanical engineer from Pune.
The platform currently has over 500 videos through which users can shop while they are watching them on the platform itself.
“A user can also add the YouTube link of a certain video, without the need to download, and our algorithms can tag the relevant items.” Saket explains.
The tagging and detection process usually takes about twice the amount of time of the video duration. Once the platform completes the process, the user can shop for products on the video through relevant ecommerce sites.
Some videos on the Toch platform are auto-curated according to user interests and run through the Toch dashboard for adding real-time analysis. Once the analysis is done the video is auto-uploaded onto the Toch platform for users to watch from the library and people who have subscribed to the specific domain under which the video falls get a push notification with the content. It also allows advertisers and over-the-top (OTT) content players to license the technology to provide widgets or real-time analysis for engaging their customers.
Toch charges brands and OTT players for the technology integration. For any transaction on the platform, Toch charges affiliate commission from brands and ecommerce websites.
It claims to have been growing 100x month-on-month on the B2B side. Some of its B2B clients include fashion designer Masaba Gupta, labels Crocs, Scott, and ICICI Bank, as well as other fashion designers.
“On the B2C side, Toch is growing organically at the moment with 15,000+ registered users,” says Saket. The major B2C clients working on the affiliate model are marketplaces like Amazon, Flipkart and Snapdeal.
“Toch is also doing a proof of concept with Cathay Pacific to provide in-flight experience to the passengers so that when they travel from India to other parts of the world, they can experience the shoppable videos and can bookmark their favourite products, and while waiting for their baggage at the belt counter they can place the order,” he adds.
Technavio analysts forecast the global digital video content market to generate $121 billion by 2020, and will grow at a CAGR of close to 26 percent.
Besides, interactive videos have become one of the major sources to learn about customer preference and to engage viewers, both online and offline. Viewers not only watch the videos but also click, swipe, and engage via the call-to-action overlays that give them an enhanced customer experience.
Some of the other companies in the same space in the West are Wirewax, Cinematique, Clicktivated, The Take, Smartzer, and in India there is Charmboard.
“Toch AI detection of people, object, scenes, activities, and emotions in real time differentiates it from its competitors in the market,” says Saket. It has raised undisclosed early-stage funds from Z Nation lab and JITO. It was also part of the ZEE Esselerator programme. Its current team size is nine.
For the future, Toch plans to expand in the B2C market by getting users to watch short content from its platform directly.