As AI technology develops, its industrial use cases range from transportation enforcement to workplace safety. AI can process large amounts of data more quickly and accurately than any other system, helping businesses make more informed decisions. It can also identify patterns and trends, allowing companies to analyze and respond to data in real-time.
But there are industries where AI technology is lacking. Intenseye co-founder, Sercan Esen, discovered this when he and his co-founder had created a video processing engine and were trying to decide how to apply it.
“When we founded Intenseye in 2018, we were trying to decide which problem to solve,” said Esen. “We were visiting all these Fortune 500 companies’ facilities where you can see drones flying around doing various tasks. But when we talked to the health and safety teams, we realized that their technology was basically pen and paper at best.”
Given the importance of employee health and safety, this is a significant gap. Workplace illnesses and injuries have been found to cost the US
Looking at this gap, Esen and his co-founder have seen an opportunity to use AI as a problem-solving tool. They had previously worked in AI and saw this as an opportunity to put the technology’s problem-solving capabilities to work. Using AI and machine-learning technology, they created a solution that allows companies to connect their existing cameras, and leverage their CCTV network to detect any unsafe conditions.
Investors have recognized the opportunity for growth as well. Inteneseye finalized its seed round in 2020 with $4 million in total seed funding and raised $25 million last year in its series A round. The company has a team of over 70 people and is growing really fast. Its technologies are in use in more than 23 countries, protecting tens of thousands of workers.
Employee health and safety is certainly an industry ready for AI development, but what is needed to adapt AI technology to this industry? How are companies adapting this technology to keep their employees safe?
Developing Industry-Specific AI Models
AI is revolutionizing a variety of industries by helping to automate manual tasks, streamline job processes, and analyze data faster. However, different industries require their AI systems to do different things. An AI system for the financial sector might need to do complex calculations, while an AI for marketing might need to have stronger language processing capabilities.
In the case of employee health and safety, an AI system can act as an observer, constantly looking out for indicators of unsafe practices or potential for accidents and injuries. The AI must be trained to recognize the machinery, local laws, and workplace regulations for each company. It can then analyze images taken through cameras in the workplace and compare these against the data, noting if anything is different or contradictory to health and safety laws or company regulations.
Although security and privacy are often a concern with AI, especially systems that use facial recognition, Intenseye is committed to protecting the anonymity of workers. Its systems use face anonymization to ensure that no personally identifiable information is captured, and are SOCII and GDPR compliant.
One of the advantages of having an industry-specific AI developer like Intenseye is that each model the company builds can help improve the effectiveness of all its models. Analyzing different versions of similar objects, machinery, and environments makes it easier for the AI to learn, making its process of recognition and analysis faster and more accurate.
In addition to creating models specifically trained to process employee health and safety-related data, Intenseye has used the latest AI technology to develop its own tools for data labeling and accuracy. These tools are industry-specific and have helped the company receive its impressive 92% accuracy rating in the last quarter of 2021. Having industry-specific data labeling and accuracy tools is especially important as Intenseye's AI processes over 22 billion images every day.
Intenseye’s Maia system is an AI-powered tool for annotating images. This tool automatically labels video feeds received from cameras, detecting machinery, objects, employees, and other aspects of the image determined by the company. Intenseye’s in-house labeling experts also check the video feed and add missing annotations, helping the AI to improve as it does its job.
Intenseye has also developed an industry-specific accuracy-checking platform called Maat. This platform allows the company’s EHS experts to measure the accuracy of its alerts. This allows Intenseye to monitor and improve the effectiveness of its AI, and also provide transparent results on a continuous basis.
Predicting Outcomes with Leading Indicators
With advanced recognition and labeling capabilities, AI is transforming the way that data is used across industries. AI is making it possible for analysts to
Currently, most EHS compliance systems that don’t use AI rely on manual checks. Workers must input data manually, logging accidents and injuries that have happened and using that information to help prevent the same thing from happening in the future. These are known as lagging indicators, data points used to analyze an event after it has happened.
AI opens up the possibility of using leading indicators, and predictions based on data. This is made possible by the ability of AI systems to constantly observe a workplace and process and analyze all the data it receives in a matter of seconds. With this vast amount of data and advanced analysis capabilities, AI systems can make predictions about which elements of an environment are likely to lead to injury before it happens.
The shift to leading indicators has already led to major increases in effectiveness and efficiency in the marketing sector, and
“To us, the most the health and safety team at a company has the most important jobs,” said Esen. “So it’s important that they have the tools they need to protect their frontline teams.”
Using leading indicators, Intenseye’s machine-learning technology allows teams to create different workflows that will help them prevent accidents. They have already sent over 15 million real-time notifications, protecting workplaces across the globe and preventing millions of potential accidents and injuries.
Increasing Adaptability for Continuous Improvement
One of the key advantages of AI technology is that it is constantly learning and improving. Intenseye maintains a focus on constantly improving its systems with teams dedicated to checking their accuracy levels. If the system’s accuracy levels drop, the team will add any missing labels to the data and work on retraining the AI to improve its accuracy.
The ability to retrain AI systems is essential in an industry like employee health and safety that is constantly incorporating new technologies and workflows. Esen noted that Instenseye’s teams are prepared for this scenario, saying that given a new use case, the company can retrain its AI models within 24 hours to detect those threats.
This type of adaptability is one of the most exciting aspects of AI. With real-time data and the ability to predict outcomes, companies can anticipate new scenarios and develop solutions more quickly. In fact, one of Intenseye’s clients has shown the scale at which this can take effect, integrating over 50 facilities in the last 5 months. This capability can help businesses keep up with an increasingly complex and fast-paced landscape, no matter what industry they’re in. And when put to use in an industry like employee health and safety, it can save lives.