The hidden cost of blind troubleshooting
When support teams cannot literally see the customer’s setup, issues are harder to diagnose. Customers do their best, but they may not know the right terminology, or they may describe the wrong device or miss something that would be obvious to a trained technician.
It gets harder as product ecosystems expand. A service representative may need to support dozens of device types and hundreds of troubleshooting scenarios. A wrong diagnosis could lead to a wrong shipment, a repeat call, a non-fault-found return, or an unnecessary truck roll.
Imagine an agent correctly determines that a battery needs to be replaced, but cannot identify the exact model. The wrong battery is shipped, the issue remains unresolved and the customer has to call again. Multiply this small mistake across thousands of interactions, and it is easy to imagine the increasing operational costs through unnecessary shipments, repeat contacts, non-fault-found returns, technician dispatches and ultimately customer churn.
Adding connectivity into the mix complicates the situation further. Sixty-eight percent of U.S. households reported recurring Wi-Fi issues earlier this year, and 76% experienced weak or unreliable Wi-Fi in specific rooms. For a smart home security provider, those numbers matter. A device that appears defective may simply be located in a weak coverage area in the home.
The result is familiar to every service operation: longer calls, repeat contacts, avoidable returns and lower customer confidence. In security, the stakes are higher. If a customer experiences repeated false alarms, cannot reconnect a camera, or waits for a technician to solve something that could have been resolved remotely, customer confidence in your organization falters.
Visual intelligence has a service role
In security, visual intelligence is usually discussed in the context of surveillance. But there is another important use case: helping providers understand whether the security system itself is installed, configured and operating correctly.
When a customer can show the issue through a phone camera, support teams gain context that would be difficult to capture through a traditional call. Visual AI can identify the device, read indicators, recognize installation mistakes, validate setup steps, and help determine whether the problem is related to placement, power, connectivity or configuration.
There are already organizations taking advantage of visual intelligence, and they’re seeing success. For example, ADT has completed more than 2.6 million visual support sessions, achieving over 80% CSAT, preventing more than 75% of truck rolls and increasing Net Promoter Score to around 50.
The next step is Spatial AI, which adds context by visualizing the customer’s space. It helps interpret where a device is located, what surrounds it, what may interfere with it, and how the setup relates to the area it protects. A doorbell camera issue may involve more than the device itself. The field of view, mounting height, lighting, Wi-Fi strength and nearby obstructions may all affect performance. A sensor issue may be caused by placement rather than product failure. A smart lock issue may come from door alignment rather than software. These are all concerns that a Spatial AI system would help diagnose.
