Notification: Notification: Visit our COVID-19 resource center - Learn More >

Reduce Noise, Improve Outcomes

12 Minute Read

If it’s your job to protect your company’s people, assets, and operations, what’s the most valuable thing you can have at your fingertips? Experienced managers tend to boil it down to actionable intelligence – confidence in the facts on the ground.

But to be actionable, that intelligence must be not only comprehensive, accurate, and timely; it must be relevant to your situation.

With so much at stake, managers err on the side of collecting as much data as possible: no head of security wants to miss something important. They typically have some combination of human analysts and technology providing a steady stream of reports. The consequence of this approach is that global operations centers are peppered with alerts, many of them irrelevant. Data overload can be as paralyzing as not knowing. Your ability to effectively anticipate and respond requires that you isolate those events that affect your world – and filter out the rest.

The sheer volume of information available today makes this job of filtering data especially tough. Tracking the proliferation of news about critical events that could impact your company has grown beyond the capacity of humans to analyze. Even if you had the resources, throwing more people at the problem won’t result in a clearer picture. To be actionable, every critical event needs to be understood in the context of your company footprint: the whereabouts of your people and the movement of your products and materials.

Know Which Critical Events Demand your Attention

At any moment in time, a map of the world is dotted with events of local, regional, or international significance. Each of these events matters to somebody, but not necessarily you. An event is critical to your operation only to the degree that it impacts your people, places, or pipeline. Big events get big coverage. But a hyper-local event like a gas leak or street protest could be far more consequential if it’s happening near one of your facilities.

Stabilitas from OnSolve recognizes the importance of critical event coverage: our purpose-built artificial intelligence (AI) solution ingests more than 17,000 sources of structured and unstructured data. But consolidating that global data is just the first step. To refine it into actionable intelligence, our Critical Event Intelligence pipeline follows a seven-step process:

‍1. USEFULNESS

Not all news stories mean what they say. A story about Tom Brady throwing a long bomb is different than a bomb scare in downtown Boston. Stabilitas has trained its AI engine to tell the difference.

‍2. GEOPARSING

If an event is deemed useful, the next step is to accurately place it on a map. This, too, is challenging. A given story may simply reference Greenville – a place name found in every state in America. Stabilitas AI outperforms the industry in geoparsing.

‍3. CATEGORIZATION

Once the lat/long of an event is known, the next step is to assess category. If it’s a fire, is it a building fire, an explosion, a chemical fire, a controlled burn, or a forest fire? The difference often matters in mounting an effective response (e.g., minimally, you need to know if the blaze is contained or burning out of control). Stabilitas recognizes 53 distinct event types in its ontology.

‍4. TIME NORMALIZATION

When did this even occur? Is it still unfolding? Depending on where you are in the world, the same event may have happened on October 15 or October 16. Stabilitas AI normalizes the time signature based on context.

‍5. SEVERITY

These last three elements comprise the unique fingerprint of a given event: type, timing, and location. But no two events are created equal. An aircraft incident may be deemed of moderate or critical severity based on the number of people on the plane and where the accident occurs. By synthesizing information from mulitple sources, the Stabilitas AI engine assigns a severity score.

‍6. ENTITY DETECTION

Once an event has been fingerprinted, the next step is to isolate the people and places mentioned in the coverage. This process of entity detection begins to inform who and what is affected.

‍7. CLUSTERING AND SUMMARIZATION

By establishing a unique event fingerprint, Stabilitas AI allows for the clustering of articles around that event. Instead of getting 638 disparate articles on a hurricane, you get a single event profile with 638 supporting stories. This programmatic consolidation creates useful structure and massively reduces noise.

That seven-part pipeline describes a global process by which raw data becomes a consumable event profile. But that event profile may or may not be relevant to your unique footprint of people and assets. To address that last requirement, artificial intelligence overlays the lat/long of events with the location of your employees, travelers, facilities, and designated vehicles. This is correlation – the linchpin of actionable intelligence. To bring that to life, let’s look at a typical map of events around the world.

This first view shows the universe of critical events happening at a given moment in time. But your job is not to save the world; it’s to protect your people, assets, and operations. This unfiltered view doesn’t help.

Example: 12,451 Events

This next view shows all critical events with a severity of high or critical. Noise is reduced but not all of these events apply to you. Actionability is still a step away.

Example: 1,537

This final view takes that subset of severe events and correlates them with your assets uploaded into the Stabilitas platform. Finally, you have actionable intelligence: you know precisely what’s happening where as well as who and what is affected.

‍‍Example: 21

21 is a vastly more manageable number of events to track and monitor than 12,451. Our situational awareness solution provides you and your team with the ability to focus on just those events which might impact you, your employees, travelers, assets, facilities, operations, suppliers and key customers. We fuse broad coverage with unparalleled speed to bring you the most accurate, actionable picture of your world. Through the solution you can provide instructions to your response team and issue alerts to affected people on the ground.

SPEED: The time between when an event happens and when it’s picked up by your team is called “time to detect” (TTD). Rapid TTD can make the difference between pre-warning personnel and reacting to an event in progress. In a recent benchmarking study, our AI detected events 90% faster than human-only teams.

COVERAGE: Our solution ingests more than 17,000 sources from government bureaus, weather and geological services, local and international press, and social media. This includes sources that security and intelligence professionals rely on for comprehensive global coverage. By cross-referencing multiple sources – including input from your people on the ground – the solution is able to cluster and validate events.

ACTIONABILITY: Our solution not only tracks critical events in real time, it helps you understand historical context by maintaining dynamic maps of your facilities, logistics, and personnel movements. This means your team can anticipate what will happen, warn of potential impacts, take proactive security measures, and re-route resources in a timely manner – assuring operational continuity and uninterrupted customer service.