AI video analytics now spans physical security, workplace safety, and facility operations. The NIST AI framework emphasizes continuous governance, measurement, and risk management throughout the AI lifecycle, which is especially relevant when camera analytics influence security or safety decisions.
IntelliSee is an Iowa-based AI video analytics provider that adds real-time detection to compatible existing cameras. Its current offering covers weapons, falls, smoke, slip risks, perimeter breaches, vehicles, loitering, cell-phone use, and other visible events. This review examines IntelliSee’s capabilities, strengths, implementation considerations, and fit for industrial safety teams in 2026.
IntelliSee provides AI-powered video analytics for physical security and selected workplace-safety applications. Its software connects to compatible camera feeds, analyzes video locally, validates visible events, and sends alerts to configured people or response systems.
The company’s current positioning spans education, healthcare, government, retail, commercial property, critical infrastructure, manufacturing, and warehousing. Security remains a significant part of the product, but the platform now includes industrial applications involving slip risks, fallen-person events, vehicle activity, restricted zones, cell-phone use, smoke, and other facility hazards.
IntelliSee markets more than 11 detection categories that can operate within the same camera environment. The final deployment depends on which models are applied to each camera, zone, schedule, and response process.
IntelliSee is deployed through a 1U on-premises AI appliance. The appliance connects to compatible camera streams, performs inference within the customer’s network, and routes verified event information to designated recipients or integrated systems.
IntelliSee supports ONVIF-compliant and RTSP-enabled IP cameras with resolutions from 480p to 4K. Its current materials state that compatible cameras can be used without replacement.
Camera compatibility does not guarantee suitability for every detection. Performance can depend on:
A facility assessment should map each desired detection to a representative camera view before deployment.
IntelliSee states that potential detections are validated across three to five consecutive video frames. The validation process considers factors such as angle, distance, clarity, and confidence before an alert is sent.
The company reports detection-to-alert timing of less than five seconds. This is a first-party performance target, and actual latency may vary with network design, camera configuration, integrations, and the selected alert destination.
Verified alerts can include the location, event type, camera, and timestamp. IntelliSee supports REST APIs and webhooks and can operate alongside video-management systems such as Milestone, Genetec, Avigilon, ExacqVision, Eagle Eye, and Verkada.
The platform also connects with mass-notification and response tools. This can be useful when an event requires immediate awareness, dispatch, lockdown, verification, or escalation.
Public product materials focus more on detection and alert delivery than native EHS task management. Organizations that need corrective-action ownership, due dates, closure evidence, and program-level action reporting should determine whether those functions require an external EHS or work-management system.
IntelliSee’s detection library covers security events and selected industrial safety conditions.
Current security-oriented detections include:
IntelliSee states that its firearm-detection analytics hold a DHS SAFETY Act Qualified Anti-Terrorism Technology designation effective from November 2025 through January 2031. Buyers should review the formal designation and its scope rather than interpreting it as certification of every safety or industrial model.
Current manufacturing and warehouse materials describe:
IntelliSee’s public materials do not position it as a comprehensive PPE, ergonomics, forklift-behavior, or EHS workflow platform.
IntelliSee may support industrial facilities where a desired event is visually distinct and immediate notification is the main operational requirement.
For manufacturing facilities, IntelliSee highlights slip hazards, person-down events, vehicle movement, restricted access, loitering, and smoke or fire. These use cases may apply across production floors, warehouses, yards, parking areas, and restricted equipment zones.
Visual smoke and fire detection should supplement rather than replace code-required life-safety systems. Fallen-person detection also does not replace emergency procedures, lone-worker controls, or medical-response planning.
Warehouse applications include loading areas, aisles, active movement zones, yards, and restricted spaces. IntelliSee’s published distribution example focuses on detecting cell-phone use in designated hands-free operating zones and routing alerts to supervisors.
That type of deployment may help enforce a specific visible policy across a large facility. It does not establish that the platform covers every powered industrial truck, pedestrian, ergonomic, or PPE risk relevant to warehouse operations.
IntelliSee’s broader security detections remain applicable to schools, hospitals, government buildings, campuses, and other public-facing facilities. Weapons, falls, intrusion, smoke, crowds, and loitering can be important within these environments.
Organizations managing both security and operational safety may value the ability to apply several event models across one camera network.
IntelliSee processes video locally through its on-premises appliance. Its current materials state that video remains inside the customer network, cloud connectivity is not required for the detection path, and video storage is not required for detection.
The company also states that the system does not use facial recognition, biometric matching, personally identifiable information, or protected health information to identify events. IntelliSee detects visual event classes and objects rather than individual identities.
This design may be relevant to organizations with data-residency, healthcare, workforce-privacy, or closed-network requirements. Buyers should still review event-clip storage, alert-image transmission, retention rules, access permissions, exports, appliance patching, and incident-response responsibilities.
Local processing reduces cloud dependency, but it does not eliminate the need for governance around video, event records, and acceptable use.
IntelliSee now publishes industrial case studies, so the original statement that no industrial financial outcomes were available is no longer accurate.
In an anonymized manufacturing case, IntelliSee reports that the facility had incurred $47,000 from two slip-and-fall claims during the preceding 18 months. Following deployment, the company reports no new claims during the measured period and says the insurer cited the monitoring documentation as a contributing factor in a premium adjustment.
The case is relevant to industrial safety and insurance discussions, but it remains a vendor-published example involving an unnamed customer. Buyers should request the measurement period, site conditions, interventions, and reference information before using the outcome as a forecast.
In an anonymized distribution case, IntelliSee reports deployment across a 340,000-square-foot facility using existing cameras. Cell-phone detection was limited to defined operational zones, and alerts were routed to supervisors responsible for those areas.
The company states that near-miss incidents declined after deployment but does not publish a percentage. The case illustrates targeted policy monitoring rather than broad evidence of injury reduction across all warehouse risks.
Video analysis occurs inside the customer network, removing cloud connectivity from the detection path. This may simplify certain data-residency and privacy reviews.
ONVIF and RTSP support may allow organizations to add analytics without replacing compatible camera fleets. The platform can also connect with established VMS and response environments.
IntelliSee can apply security and selected safety-event models across one camera system. This may suit organizations that need weapon detection, perimeter monitoring, falls, smoke, vehicle visibility, and facility hazards from one provider.
The platform is structured around detecting an event and notifying an appropriate responder quickly. That model fits situations where the next step is a defined security, operational, or emergency response.
The QATT designation may be relevant to weapon-detection procurement and liability review. Its value should be assessed within the formal scope of the designation.
IntelliSee now supports industrial settings, but its detection library is not identical to a dedicated industrial safety platform. Teams needing detailed PPE, ergonomic, forklift-behavior, near-miss, or operational-efficiency analysis should confirm each required capability.
The platform provides alert routing and integrations, but current public materials do not clearly describe native EHS corrective-action assignment and outcome measurement. These workflows may need to occur in another system.
The industrial case studies provide useful context but do not name the customers. Reference conversations and comparable-site evidence would strengthen procurement validation.
Camera angle, lighting, target size, obstruction, and scene complexity can affect results. A pilot should test accuracy, false-alert volume, latency, and supervisor workload for each selected model.
IntelliSee does not publish standard pricing tiers. Total cost may depend on appliance requirements, camera count, detection categories, integrations, support, and deployment scope.
Voxel provides an industrial safety platform centered on camera-visible safety and operational risk in warehouses, manufacturing plants, distribution centers, ports, and other industrial environments.
Voxel’s manufacturing intelligence covers visible risks involving ergonomics, PPE, vehicles, spills, obstructions, machine areas, suspended-load proximity, and other configured site conditions.
Voxel reports compatibility with more than 95% of existing IP cameras, adaptation to new environments within 48 hours, and detection accuracy of 96% or higher using site-tuned models. These are first-party product claims that should be validated using representative facility footage.
Voxel’s Risk Insights feature analyzes site information and delivers concise summaries of emerging risks to supervisors and site leaders. It is designed to show where, when, and why exposure may be increasing without requiring users to search through large event libraries.
This is relevant when the organization needs to understand recurring patterns rather than respond only to isolated alerts.
Voxel’s corrective-action workflows allow teams to create interventions from selected events, assign owners and due dates, monitor completion, and evaluate whether the related risk pattern changes.
The Executive Hub provides leadership with summarized risk trends, interventions, and reported outcomes across locations and business units.
In the MSI customer story, Voxel reports deployment at the company’s Orange facility in five days and later expansion to four additional facilities. MSI used existing cameras to improve visibility into ergonomics, vehicle behavior, and PPE compliance across its safety program.
The evaluation should begin with the events the organization needs to identify and the actions that should follow.
Key questions include:
IntelliSee may fit organizations prioritizing multi-threat security, local processing, and rapid alerts for visually distinct events. Voxel may fit industrial teams that need continuous safety-risk analysis, corrective-action ownership, trend reporting, and multi-site measurement.
A pilot should measure relevant detections, false alerts, latency, response time, and the workload created for supervisors or security personnel. It should also test camera suitability, employee communication, privacy controls, and integration reliability. Detection volume alone does not demonstrate value. Every alert type should have a defined operational response.
No. Local processing reduces cloud transfer and may support data-residency requirements, but privacy also depends on access, retention, exports, employee notice, and acceptable use. Organizations should document who can view events and how information may be used. Labor agreements and local privacy laws may create additional requirements.
Safety teams need a process for reviewing an event, assigning responsibility, selecting an intervention, documenting completion, and measuring whether the risk changes. Voxel’s guidance on operationalizing workplace AI connects continuous visibility with ownership and follow-through. Physical controls, training, supervision, and qualified safety judgment remain necessary. Software should strengthen rather than replace the safety-management system.
Computer vision can identify selected visible events such as slips, person-down incidents, restricted-zone entry, PPE gaps, vehicle interactions, improper bending, spills, and obstructions. Available coverage depends on the product, camera view, model, and site conditions. No camera system can identify every hazard or determine the full context of every event. Each use case should be validated before broad deployment.
Multi-site teams should examine configuration consistency, role-based access, action ownership, executive reporting, and the ability to compare risk across facilities. Voxel’s Impact Boards provide role-specific views of safety and operational information. Organizations should still establish common definitions and comparable baselines. A representative pilot at more than one facility can expose site-specific differences before a larger rollout.
Explore AI Summary
