Background
I'm the CEO & Co-Founder at Solink, a company we founded in 2010 with the purpose of detecting and reducing fraud in banks. We've since grown to over 1000 customers globally and growing at 10-15% monthly. The focus has always been on using data analytics to make video a proactive source of insights. In 2015 we pivoted the company to provide an end-to-end cloud solution rather than managing integrations with outdated DVRs/VMS' on-premise. We've seen time and again that the value is in the data, not the security infrastructure.
I wanted to introduce Solink to the IPVM community to get feedback and questions. Happy to treat this as a Reddit-style AMA (Ask Me Anything) or feel free to email mmatta@solinkcorp.com if you’d rather have a private discussion.
Why another "video" company
First-off - we are a data company and use video as a verification source. Some customers have us in parallel to pure-play VMS/DVR products and that's fine. We are seeing a commoditizing video market - the value is in applications around data, video is another input to data.
Surveillance video is primarily reactive, it's main use is to record and provide evidence to support an incident. We love the concept of "video analytics" but in practice there is a lot of tuning and noise from computer vision that results in a lot of data but little actionable information. Traditional Exception-Based Reporting (EBR) tools have been helpful but they lack the ability to seamlessly integrate to video or adapt to new techniques employed by internal theft or external fraud. Data analytics using machine learning, on the other hand, is more practical at identifying suspicious activity and improves with use & quantity of data.
Offering
Solink is a complete solution for retail and financial institutions to identify and reduce losses. Best described, Solink brings Video + Exception Based Reporting + Case Management in a single end-to-end offering. We offer the solution on a RMR basis and rev-share RMR with partners for the life of the customer.
Video Discovery
Solink ingests transactions and "time-based" events from existing systems (like POS & ATM/Teller, Access Control) to understand "context" and we use video as a verification tool. We enable search in our cloud to identify patterns or specific events at scale. Solink retains transaction data and snapshots/clips for up to 1 year. You can search through millions of events in milliseconds. We call this "Video Discovery" - here is an example:
When an incident is identified you can watch the full footage from our web or mobile application. You can share evidence or push a specific clip to our "cloud" for extended retention via a "bookmark". The video controls are intuitive and user-friendly.
Mobile
Solink has native apps for both iOS and Android. Users can search, view, share, and watch any event or camera live. We focus on our mobile experience heavily and have high adoption and engagement from our user base on mobile.
When an incident is identified you can share it to our "Case Management" add-on. The embedded video clip and transactional data is shared to a cloud-based document store as a case. You can manage the workflow of cases and share internal & external as well as collaborate "in-case" - similar to Slack but around a case workflow.
Reports & "Daily Digests"
We identify patterns of suspicious activity through reports. Users can filter events by selecting a report category and narrowing the events in view to say "successive discount transactions" or "High number of Deleted items per order" or, in banking, "multiple declined transactions for common accounts". Users receive a daily email, that summarizes these reports and provides a snapshot & link to allow direct access to the application.
Machine Learning
As the system collects more data it gets smarter and can identify issues earlier. We tag customers by peer group and benchmark activity for each peer group (QSR, banking, specific geography) to develop a real-time "index". At scale, this index becomes "normal" and we can understand the summary data that distinguishes high and low performers. So, as a very simplistic example, if a particular retail chain has an average of 0.5% discounting or declined transactions and one store spikes to 1.5% this points to 3x higher activity that we can alert on. This is the biggest bet we are making - as we collect more video, data, and transactions our focus is on "picking winners" - similar to how Amazon can recommend future products based on your past purchases and benchmarked activity of similar users.
Hardware
As part of our solution, we include a NAS device - called "Connect Gateway". Once Connect is flashed to the device all ports/services are locked down and user management is managed from our cloud. Connect is a lightweight VMS - we can integrate to any ONVIF or (rtsp & h.264) compatible camera on the network. We have native camera integrations and recording rules for motion, schedules, Low/High-Def recording strategies - the basic requirements of any VMS.
Management
The application is managed through our secure device management layer called "Call Home" where each device can be configured, updated, and managed for health at scale from the cloud. Partners can manage their entire customer base from a single portal. This should look similar to other IoT MDM applications like Cradlepoint. Along with hardware/network health we can push down connectors to data systems (like POS') and reports. Each "tennant" can be bubbled-up to the reseller level to manage ALL customers. All Identifiable Data & Video is isolated from this portal for PCI & SOC compliance.
Happy to take any questions, thoughts or comments
Thx
Mike Matta - CEO @ Solink (www.solinkcorp.com)mmatta@solinkcorp.com