VisionSentinel | PlantIQ AI

Two Visual-AI Monitoring Appliances

Both are autonomous, agentic appliances. The primary difference is where AI processing runs.

All-in-One: AI On Board CapEx

Integrated AI processing: Dashboard Captures, Images, and Visual Assessments never leave the site
Suited to air-gapped sites, or wherever data governance prohibits outside AI services
The larger appliance: Higher CapEx, no per-assessment AI OpEx
Powered from a standard outlet

Compact: AI On Demand OpEx

AI processing by a site-approved private service: Dashboard Captures and Images are sent to it, and Visual Assessments come back. The collaborative workspace and its records stay on site. Only questions put to the optional AI assistant may also go to that service
Suited to sites whose data governance permits only approved, private AI services
The smaller appliance: Lower CapEx, with AI OpEx billed per assessment and per optional AI-assistant use
Powered from a standard outlet, with Power over Ethernet (PoE) coming soon

Which Appliance Fits?

Your data-governance policy ultimately decides. If assessments must stay on site, choose All-in-One. Where approved private AI services are allowed, Compact costs less to start. Either way, the hardware is supplied at cost, with no markup. All-in-One needs high-capacity memory, which is currently in short supply globally, so its availability is confirmed during planning. Prices for the appliance, the setup engagement and the annual software subscription and support are on request.

What VisionSentinel's Visual-AI Sees, and What It Says

Three demonstration dashboards and one sample assessment result. Lettered regions mark what VisionSentinel was configured to watch most closely, while the rest of the image serves as supporting context. The fourth tile is what VisionSentinel posted about the pulp mill to its built-in, shared collaboration workspace. Click any to enlarge.

AI data center · GPU cluster liquid cooling

What VisionSentinel posted about the pulp mill

Plant operators watch screen after screen across long shifts, yet emerging problems rarely announce themselves with an alarm. They begin as drifts too slight to notice, or as readings across unrelated systems that only your most senior people know to read together.

VisionSentinel gives every shift a second pair of agentic-AI eyes that never tire. A VisionSentinel appliance on your network looks at your dashboards and field-camera images on a schedule, much like an expert would, and posts what it sees — NORMAL, WARNING, or CRITICAL — with a plain-English explanation to a collaborative workspace, where your team reviews, discusses, and acts on each finding. It only observes: Nothing in the plant changes for it to run, and no decision is handed over to it.

VisionSentinel Capabilities

Scheduled dashboard captures and field images assessed by Visual-AI, with results
posted to a shared collaboration workspace for team discussion and review.

Dashboards and Field Images

VisionSentinel assesses dashboard captures, and images your systems supply from field cameras at the edge. Each assessment has its own schedule and evaluation criteria.

Configurable Visual Assessments

The Visual-AI evaluates each image set against its configured criteria. An image set can contain a single image, a set of correlated images from disparate sources, or time-series captures of the same image to spot slow drifts. Results include a NORMAL, WARNING, or CRITICAL status, a confidence score, a technical reading, and plain-English wording.

Collaborative Workspace

Appliances include an integrated workspace for team collaboration. AI assessments are posted to the vs-runs channel, with WARNING and CRITICAL findings repeated in the vs-alerts channel. Staff can acknowledge findings or flag false positives with emoji reactions, then review them in the vs-discuss channel.

Rolling Assessment History

Recent assessment results remain available for review, including the AI's responses and the original and annotated images. This history is retained for the configured period, then automatically deleted by the appliance.

AI Assistant for Results

An integrated AI assistant supports follow-up questions about assessment results. Plant administrators can add context, such as a runbook page. The assistant considers this context alongside the returned findings to help explain whether further review is warranted.

Low-Touch Appliance Operation

After setup with PlantIQ-AI, the appliance runs with minimal administration. Assessments run autonomously, and built-in commands change settings or pause and resume individual assessments. When support is needed, PlantIQ-AI can request log exports, which plant administrators download through the appliance interface.

See It In Action

Request a demo to see VisionSentinel at work.
The full datasheet is for customers under NDA.

Request a Demo Datasheet

VisionSentinel Patent pending