A structured review of Cobalt Monitoring Intelligence, an AI layer over the security systems already installed for enterprise security operations and GSOC teams: what it does well, where the constraints sit and who should look elsewhere. Every score traces to something Cobalt AI publishes.
Cobalt AI is doing something narrower than a video platform and doing it with unusual rigor. Cobalt Monitoring Intelligence layers over the systems already installed, reasoning across video, access control and sensor events, and it publishes what most of this category will not: the systems it connects to by name, eight access control platforms and three video platforms plus any RTSP or RTSPS stream, and the numbers its own customers see, with up to 94% of events resolved without human escalation and an average handler response time under 15 seconds. The handlers are named per event type and the design goal is stated plainly, which is to make human operators more effective rather than to replace them. It scores lower on what happens on site, because the automated actions are workflow actions rather than anything a person in a yard would notice, and lower again on what a camera has to be to join, because no manufacturer or model list appears. Choose Cobalt when the problem is an alarm queue nobody can keep up with. Look elsewhere when the problem is what the cameras themselves are failing to notice.
Five criteria, each scored from Cobalt AI's own current documentation as of August 14, 2026. Spot AI competes with Cobalt AI, so no score here rests on an anonymous source, an aggregated user rating or a private benchmark: each one sits next to the documented fact behind it. Anything Cobalt AI does not document is recorded as not publicly specified rather than assumed. Where a fact was not on the platform, handlers or integrations pages, this review says so rather than filling the gap from a press summary. This is not a paid placement, and Cobalt AI had no input into it.
This is a reasoning layer over systems somebody else records with. The integration list and the outcome numbers are published. Anything physical is not.
Cobalt AI sells Cobalt Monitoring Intelligence, an AI platform for physical security operations rather than a video management system. In Cobalt's own description it connects to your existing cameras, access control systems and edge devices, then reasons across all data sources to surface context, identify root cause and automate response workflows. The unit of work is a handler, which Cobalt defines as a dedicated AI workflow designed for a specific security event type, each one following a four-stage chain the company names as Detect, Reason, Act and Resolve. The company states the platform is built on more than ten years of in-house security operations experience.
The handler list is published per event type: door forced open, door held open, tailgating and unauthorized entry, gun detection, person on ground, person detection after hours, crowd forming, person climbing a fence, loitering, rapid egress, runaway alarm detection, periodic audits and automated alarm root causing. Cobalt's own worked example is the useful one to read: a door forced open alarm gets cross-referenced against the video and the badge logs, the platform identifies a faulty request-to-exit sensor as the cause, opens a service ticket and clears the event, which is the difference between detection and judgment in its framing.
What it connects to is named rather than implied. For video, Eagle Eye Networks, Milestone XProtect and Genetec plus any RTSP or RTSPS camera stream. For access control, eight platforms including LenelS2 OnGuard, Genetec, Avigilon, Avigilon Alta, S2 NetBox, Brivo and Acre Feenics. Then Okta for identity, ServiceNow and Salesforce for ticketing, and Slack, phone, SMS and email for notification. Cobalt also publishes a managed service in the Cobalt Command Center and an autonomous patrol robot with more than 60 sensors that feeds the same platform, and it names Salesforce, FedEx and Ally Financial as customers.
Cobalt is strongest where an alarm queue is the bottleneck and the underlying systems are sound, and thinnest where the cameras themselves are the weak link.
Both columns are documented. Nothing here comes from an anonymous review.
Both columns describe documented behavior. Cobalt AI's column was checked against its own documentation on August 14, 2026.
Swipe the table sideways to see every column.
Cobalt AI data comes from Cobalt AI's own public documentation as checked on August 14, 2026. Gaps are marked as not publicly specified.
The honest split, stated the way a shortlist call would state it.
If the bottleneck is volume rather than coverage, if the access control system generates more events than anyone can read, and if the video platform underneath is sound and mid-contract, Cobalt is aimed exactly there and publishes numbers for it. The named integration list also makes it one of the fastest platforms here to qualify or disqualify against a real enterprise stack.
If what the cameras notice is the problem, or if something has to happen on site before anybody arrives, a layer over the current platform inherits the limits of that platform and stops at a workflow action. Teams that also want safety and operations answers from the same feeds are looking at another product entirely, and the arithmetic of paying for both belongs in the decision.
Spot AI ships 15+ pre-trained Video AI Agents spanning vehicle break-in, fire, intrusion, personal protective equipment, forklift near-miss, falls and hazard-zone crowding, so a mixed estate gets shipped coverage on both sides. Iris builds anything else in natural conversation in about eight minutes, full-resolution video stays on the Intelligent Video Recorder in the building and the platform is SOC 2 Type II, NDAA-compliant and HIPAA-aligned.
These two are stacked rather than parallel, which is the thing to establish before comparing them. Cobalt reasons over the events a security stack already produces and hands operators a resolved case. Spot AI is the platform that produces the events in the first place, running named AI coworkers on the cameras a business already owns, keeping full-resolution video on the Intelligent Video Recorder in the building and acting at the moment through standard speakers. One reduces the queue, the other changes what enters it.
The cheapest way to settle it is to count the renewals. If the video platform underneath is staying for three more years, Cobalt is the lower-disruption purchase and the honest recommendation. If both renewals land in the same year, price the combined stack against one platform that records and reasons in the same place, using the identical site list, and let the arithmetic decide rather than the pitch.
Ask both vendors to describe the same 03:00 event end to end, from the first frame to the last action. The two answers stop at different points, and that difference is the decision.
See what the AI catches on your live feeds before any platform decision.
Customer-reported outcomes from named Spot AI customers.
Cambridge City cut footage search from two hours to 30 seconds after consolidating seven municipal locations on one platform.
Unique Industries covers more than a million square feet with a three-person safety team, catching near misses and falls on the cameras already installed.
Blackmon Oil runs a vehicle loitering filter to keep parking lots clean and safe at its around-the-clock stores, with a smaller overnight crew.
"The biggest benefit of Spot AI is how easy it is to look up incidents, see the footage, and then share it and collaborate."
Yes, for the right problem. It scores 3.6 out of 5 here, carried by a named integration list, published operating statistics and handlers enumerated per event type. It fits an enterprise security operations center whose bottleneck is alarm volume rather than camera coverage. Teams whose problem is what the cameras notice, or what happens on site at the moment, should read the published action list first.
For video, Cobalt names Eagle Eye Networks, Milestone XProtect and Genetec, plus any RTSP or RTSPS camera stream. For access control it names eight platforms including LenelS2 OnGuard, Genetec, Avigilon, Avigilon Alta, S2 NetBox, Brivo and Acre Feenics. It also connects to Okta for identity, ServiceNow and Salesforce for ticketing, and Slack, phone, SMS and email for notification, and states it layers over existing infrastructure with no overhaul required.
A handler is Cobalt's term for a dedicated AI workflow built for one event type, following a chain the company names as Detect, Reason, Act and Resolve. The published set includes door forced open, door held open, tailgating and unauthorized entry, gun detection, person on ground, person detection after hours, crowd forming, person climbing a fence, loitering, rapid egress, runaway alarm detection and automated alarm root causing.
Three figures, all its own: up to 94% of events resolved without human escalation, an average handler response time under 15 seconds, and more than 6,000 hours of operator time returned per month at a single enterprise deployment. They are the vendor's reported results rather than audited numbers, which is what makes them useful: a figure a vendor prints is a figure you can hold it to. Turn each one into a pilot specification, naming the sites, the event types and the measurement window before the trial starts.
It depends on which layer is actually failing. If the video platform underneath is the constraint, or the same cameras need to cover safety and operations, a camera-agnostic platform such as Spot AI fits, because 15+ pre-trained Video AI Agents run across the fleet already installed, full-resolution video stays on the Intelligent Video Recorder on site and deterrence runs through standard speakers. If the queue is the problem and the stack is sound, Cobalt is hard to fault. The alternatives roundup compares the field side by side.