AWS for Industries

Building a Scalable Loss Prevention and Video Intelligence Service using AI and AWS

Many security camera systems are accessed when a security incident takes place, being either an insurance policy or a reactive measure for unexpected mishaps. Solink is delivering a more proactive type of commercial loss prevention solution by pairing video footage with transaction data to help companies assess and identify risk by making intelligent decisions that lower costs.

Solink is a video security-as-a-service (VSaaS) product built on Amazon Web Services (AWS) and is using artificial intelligence (AI) to automate important tasks, such as video annotation. This solution has greatly improved their customer’s ability to monitor transactions, gather insights, and prevent losses. For many businesses, Solink delivers a return on investment of three to five times. In addition, customers have seen a 2-percent increase in profits by controlling and ultimately reducing shrink using Solink’s offering.

Building a Scalable and Secure Loss Prevention Solution on AWS

Solink is making it simpler to match security footage with point-of-sale transactions and other events like access control, inventory, and other sensors. Many of its customers are retail stores, restaurants, and financial institutions that previously had limited access to security footage and often only reviewed it reactively when an incident took place. And while identifying the most important risk factors is critical to strong security, it previously required manually watching hours of footage. That’s where Solink saw an opportunity to innovate.

“Our approach has always started with being secure, and our solution needs to be always-on and always-up,” says Jim Farrell, vice president of sales at Solink. “We then bring innovation to those principles so that everything we build can be easily accessible and deliver value to our customers.”

Part of Solink’s innovation was building a cloud-based software as a service. The company knew it needed a scalable, highly available, and secure foundation that it could use to cater to its global customer base, and it chose to use AWS. “As we started architecting our platform in 2015, we found that AWS services gave us a lot of the infrastructure that we previously would have had to build ourselves,” says Norm Wong, Chief Technology Officer at Solink. “We have high availability and scalability built-in.”

The company also ingests data continuously. Much of the data is sent directly from sensors and other devices that can integrate via APIs or local integration. Solink, which continues to grow, currently ingests and correlates 1 billion transactions from tens of thousands of locations across more than 40 countries per month. Making this data always securely available in the cloud is necessary to ensure the real-time insights we can deliver.

Using AI to Power Intelligent Security Monitoring

Solink uses Amazon Elastic Compute Cloud (Amazon EC2), which provides secure and resizable compute capacity for virtually any workload. The pipeline annotates videos and images with additional contextual information—such as the time and date of each transaction and the device’s location—to make them searchable. To access the video remotely, customers use a secure relay technology that distributes video across AWS Regions to minimize latency and maintain security – similar to how a CDN works.

To manage its current network of over 300,000 connected devices, Solink uses Amazon Relational Database Service (Amazon RDS), which makes it simple to set up, operate, and scale structured data in the cloud. “We don’t have to worry about redundancy or database backups because it’s managed for us through AWS services,” says Martin Soukup, VP of Engineering at Solink. Solink’s customers have convenient access to this data through dashboards and widgets, and they can view and filter the data and generate performance reports. Then, they can drill into potential security events—sales, drive-through transactions, deliveries, and so on—to watch the video produced by any relevant cameras.

Strengthening Security While Reducing Costs on AWS

By using Solink, companies unlock compounding internal benefits, from savings on loss of goods to reduced internal theft. “Our customers get to know their businesses better,” says Farrell. “They have better control over their staff and a better ability to coach and train them.”

Companies also save in other ways. By using Solink’s Video Alarms solution, they can be much more confident in determining whether a given alarm was triggered by a person and whether the police should respond. The system uses Sage Maker to build custom models that yield high confidence, which automates and lowers the cost of image recognition and video analysis with machine learning, to outperform legacy “sensor/panel”-based offerings. As a result, Solink reduces the chances of false alarms that can result in fines for a business. And when the police are called, Solink customers often see faster response times because the video verification helps responders better understand the situation.

Moreover, companies in the US are improving compliance with Occupational Safety and Health Administration regulations by using Solink-connected cameras and AI to identify potential obstructions to exit doors. Purchasing dedicated systems could cost hundreds of thousands of dollars, but Solink offers this capability as a built-in feature.

Continuing to Enhance Solink’s Software as a Service on AWS

On AWS, Solink has achieved virtually seamless scalability, which means it can continue to grow and meet customer demand without worrying about infrastructure management or costs. The company is continuing to explore innovations in security and using generative AI technology to build new offerings, such as Solink Sidekick AI, which the company launched in 2023.

“The scalability of our system grows based on usage,” says Soukup. “Having access to a lot of the services on demand has freed us up to adopt and implement software much quicker than we’d traditionally be able to do.”

Read more about Solink’s story here.

Martin Soukup

Martin Soukup

Martin Soukup leads Engineering and DevSecOps at Solink Corporation. Martin’s special talent is being able to see order in the chaos. His current focus is bringing his experience in cybersecurity, IT/OT convergence, and AI/ML in SaaS and IoT to the VSaaS world. Martin has over 25 years of R&D experience, 17 in the cybersecurity space, 18 in video, and 10 in AI/ML. He has led large teams in electronic design, embedded software, high performance networking, cloud, data analytics, and machine learning industries. He led ethical hacking teams for almost 10 years and wrote his first hack as a teen. He has run startups and held leadership roles in Fortune 500 and medium-sized businesses around the world. Martin holds a Master of Science in Information Technology from the University of Liverpool (Computer Security) and a Bachelor of Arts Psychology from Carleton University. He has also studied at the University of British Columbia, the University of California at Berkeley, and the Wharton School of Business. Martin has experience speaking about cybersecurity topics and technology trends at various industry events and forums, has numerous academic and trade publications, contributed to robots.txt, IETF, IEEE, ITU-T, Java, and W3C standards. He also has over 30 granted and pending patents.

Justin Swagler

Justin Swagler

Justin Swagler is worldwide head of Physical Retail at AWS, where he leads the global strategy and thought leadership for physical retailing. Justin has 15+ years of consumer packaged goods, retail, and strategy experience spanning innovation strategy, retail operations, product development, and executive leadership. He is passionate about shepherding organizations to strategically innovate and reinvent consumer experiences. He holds an undergraduate degree from the University of Illinois at Urbana-Champaign and an MBA from the Kellogg School of Management.