Unlocking hidden insights in your incident data

How can your 911 center unlock insights to improve quality, compliance, and investigations? The answer might surprise you – it’s hidden in your voice recordings and CAD incident data. 

Until now, these insights were difficult to get to. But by combining voice recordings, CAD, and audio analytics, these insights can bubble to the surface. In this Q&A interview, NICE’s Director of Marketing & Business Development for Public Safety, Diamond Chaflawee, explains how.

Both CAD and voice recordings are two essential solutions that every PSAP employs – so how can they work better together?

Chaflawee: You may have heard the saying: “the whole is greater than the sum of the parts.” That’s the case when you integrate voice recordings and CAD data. Combining CAD status updates on one timeline along the 911 and dispatch radio recordings creates an authentic incident reconstruction that shows when the 911 call came in, how long it took to dispatch a response, how long it took units to get to the scene, and so on. This makes it possible to easily see whether certain response standards were met. For example, according to the National Fire Protection Association standard for emergency services communications (NFPA 1221), certain types of calls (including those which require EMD questions and pre-arrival instructions) must be dispatched within 90 seconds 90% of the time and within 120 seconds 99% of the time

Without this capability, piecing this information together can involve a lot of manual work – listening to the voice recordings, marking the start times, and measuring the time it took from when the 911 call came in to when a unit was dispatched. On the other hand, being able to visualize all of this information on one timeline is a significant timesaver because you can get to those answers faster. 

You can also identify gaps between what was said and what was done by listening to the voice recordings and comparing them to juxtapositioned CAD status updates. You can see if what was said aligns with actions taken. For example, there have been cases where a dispatcher communicated instructions over the radio but logged different information in the CAD system. By integrating voice recordings and CAD data, it’s easy to identify these disparities. The only other way to do this analysis would be to manually compare the recordings and CAD data.

Are there any other benefits to integrating the two systems?

Chaflawee: Yes, another benefit is that it gives you more ways to find the calls you’re looking for, faster. 

For example, today, most searches are a two-step process. First, you look up the incident in the CAD system, then you use the incident’s time and date stamp to then go and retrieve the voice recordings from a separate system. It’s a multistep process and it’s not exactly precise, because it can return a lot of recorded audio – and much of it may not even relate to the particular incident you’re trying to reconstruct. 

When you consider the dozens or even hundreds of incident reconstructions that PSAPs may do every week, the inefficiencies can be staggering. Now imagine the time savings if you could find the voice recordings simply by typing in a CAD incident number. An integrated solution also makes it easier to retrieve specific types of audio recordings for training, investigations, or QA, because you can search for calls based on their classification, and other CAD criteria as well.

You mentioned audio analytics as another way that PSAPs can unlock hidden insights. I thought audio analytics were mainly used in commercial contact centers. What’s the relevance to public safety?

Chaflawee: True, audio analytics are more commonly used in commercial contact centers to address customer satisfaction issues, and generate business intelligence on products, competitors, and services. But providing the audio analytics solution is designed the right way (more on that in a moment), audio analytics can serve a useful purpose in public safety settings as well. 

An investigator can use audio analytics, for example, to find a 911 call where a specific word was spoken, or to find calls that were part of a build-up to an incident – calls that could reveal key insights and even, potential eyewitnesses. 

Say for example there’s a shooting incident in a public location in broad daylight. One eyewitness reports a man in a white SUV fleeing the scene. But are there other leads that could be missed? Using audio analytics, you can search for other calls where the phrase ‘white SUV’ was mentioned. Now let’s say you find another call, you listen to it, and you now learn that the suspect was wearing a gray hoodie. They you can use the phrase ‘gray hoodie’ to potentially find more calls and more leads. 

This brings me to another key difference between contact centers and public safety agencies. Managers who are evaluating calls in the contact center typically know what they’re looking for in advance. So it’s’ easy to set up the rules ahead of time and let the audio analytics engine flag calls that meet preset criteria (for example calls with that contain words like ‘upset’ or ‘mad’ to identify unhappy customers). 

The public safety environment is much more challenging. Investigators usually don’t know what they’re looking for (for example the ‘white SUV’ or the guy in a ‘gray hoodie’ until after the fact). First something has to happen, and then they investigate it. The implication is that the audio analytics engine needs to process every word in every call as they’re recorded, and then index calls accordingly so they can be searched when needed.

Can you think of any other applications for audio analytics in public safety environments?

Chaflawee: Yes, Quality Assurance (QA) is another area that can be significantly enhanced through audio analytics. For example, PSAPs can use audio analytics​ to identify specific types of calls for QA review (for example, by isolating specific words and phrases, such as ‘CPR,’ ‘cardiac arrest,’ and so on). They can also isolate problem calls (for instance, calls that contain unacceptable language).

Managers can also use audio analytics, coupled with CAD data, to determine if call-takers are adhering to proper procedures. So if they want to evaluate ‘cardiac arrest’ calls, they can use the CAD incident classification to first find those types of calls, and then use the audio analytics to detect the presence or absence of key words / phrases, to determine whether specific protocols were followed.

In and of themselves, audio analytics and CAD-voice recording integration are developments that promise significant insights and productivity improvements. But together, they’ll go even further toward helping PSAPs unlock hidden insights in their incident data.
Share this:
Twitter LinkedIn Facebook Email