Video analytics is now often talked about in mainstream conversations. Many of us have a very Orwellian view due to some governments reportedly using facial recognition technologies to study the social behaviour of individuals. But video analytics is much more than just that. The businesses that employ it require technological capabilities in computer vision and the use of machine learning.
Based in New York and founded by Aaron Rhodes and Amit Dhand in early 2019, Collective Intel believes that there is a need for the corporate world, the hospitality industry, and government bodies to use software to analyse terabytes and petabytes of data to observe and address human behaviour.
Collective Intel’s platform extracts data from videos and provides actionable intelligence that solves challenging business questions. These challenges vary: from a product defect to managing the shop floor.
“All companies have to do is connect the videos to our platform and tell us what they want. Our platform will solve it for them,” Aaron Rhodes, CEO, and Amit Dhand, Chairman, of Collective Intel tell YourStory.
Amit says video data will drive the future of humanity.
“Think about driverless cars, automated retail, remote healthcare, and entertainment on all platforms. Imagine that Facebook and YouTube have hired thousands of employers to watch and moderate videos. Collective Intel can extract video data more efficiently and accurately than any other company. It offers accuracy rates approaching 99 percent, which is not possible even with the best AI,” Amit says.
Aaron and Amit have been working in the Big Data and AI space for well over a decade. Aaron pursued MBA from John Hopkins University and graduated in 2017. He graduated from Juniata College in 2009. Amit has an MBA from Virginia Tech and graduated in 1998.
Aaron was at ComScore (a provider of online data) while Amit was a serial entrepreneur who tasted success with companies such as Affinnova (acquired by Nielsen) and NAILBITER ( a CPG technology company).
Both Amit and Aaron have spent years working with data products to improve video analytics. They knew each other through common friends and realised that they could bring Artificial Intelligence into the mix when they discussed ideas.
How the startup works
The startup’s video platform is deployed in convenience stores to increase display engagement and employee presence. In large grocery chains, they are studying checkout operations and display engagement. They also help drive-through stores in the US understand customer wait times.
Collective Intel’s platform attaches to any fixed or drone camera, transmits the video to ready-to-code users who decipher and solve for the objectives needed for the company. The data goes through layers of quality analysis, and is then collated and shared with clients.
“The platform is also highly flexible. When COVID-19 hit, we were able to easily pivot into a model where we help cities, retailers, and businesses safely reopen,” Aaron says.
They have had many trial customers, but the product was officially launched on July 1, in the middle of the COVID-19 outbreak, with an undisclosed number of clients. “We have done many pilots for leading retailers and restaurants, and the data is promising,” he adds.
“I need to stress that the system has the highest privacy and security, but is also transparent. We blur faces and licence plates of individuals. Also, our video is open to the public so there is nothing to hide,” Aaron says.
The business model
Three undisclosed investors have invested in the startup. In the next 18 months, the founders are looking to expand their product in India. The company charges customers based on the size of the videos and client objectives.
Collective Intel does not want to divulge revenues, but believes its revenues can cross over $1 million in ARR.
Many companies are working in different areas of analytics. AITech provides solutions for security cameras, Acquifi replaces repeatable tasks such as identifying objects, and Deepomatic and Cloudsight provide video analytics for field force.
Video analytics and AI industry
Video analytics is a global business. Most countries heavily rely on video cameras for theft prevention or to deter incidents. These places typically store only seven to 30 days of video archives. This is because videos are very expensive to store. To illustrate, QSR chain McDonald’s has day-long videos, from stores and marketing that are stored in their data centres, for a short while, and later erased to avoid huge costs.
Companies like that need a solution to choose which video is important to store. “The video is only saved if it captures an incident of some sort,” Amit says .
There are three parts to the ecosystem today. The client installs hardware such as cameras. Software like VMS (Video Management Systems) allows people to watch and store videos over the internet. Then come the analytics, which converts video to data.
In 2021, AI augmentation is expected to create $2.9 trillion of business value and 6.2 billion hours of worker productivity globally, according to Gartner, Inc. Gartner defines augmented intelligence as a human-centred partnership model of people and AI working together to enhance cognitive performance. This includes learning and decision-making.
“In the video analytics industry, we build accurate and affordable systems, which is why every installation that requires high accuracy still uses large teams of people (think casinos) and we can make AI solve video analytics,” Amit says.