AI Leadership Insights: Smarter With AI

Mike Vizard: Hey folks, welcome to the latest edition of the Techstrong.ai video series. I’m your host, Mike Vizard. Today we’re with Dustin Radtke, CTO for OnSolve, and we’re talking about how to be prepared and how AI might make us all smarter when it comes to business continuity. Dustin, welcome to the show.

Dustin Radtke: Thanks, Mike.

Mike Vizard: What is the challenge that people run into here? I mean, I think it was Oscar Wilde said originally that the sign of intelligence is to say you should be able to expect the unexpected and then that wound up in a Monty Python skit from there. But what is it that people are challenged with today? Can they not just walk through and go, well, if A is gone, what would the impact be downstream on B, C, D and E?

Dustin Radtke: Yeah, I think they definitely can understand what they would do. It’s to understand when they actually have to do it. With everything going on across the globe nowadays, whether it be weather events, whether it be riots, protests, or even wildfires up in Canada, there’s so much happening. It’s just more than what any single human or team of humans can keep up to and keep up with as it unfolds, to make sure that you’re actually taking the events so you can have the best outcome possible for your company or community.

Mike Vizard: So what exactly is the role of AI in all this then? What does that enable me to do that that team of humans wouldn’t have been able to do before?

Dustin Radtke: Yeah, I mean, if you think about what it’s able to do, it’s able to actually scan thousands and thousands of sources of intelligence on a continual basis, but not only scan that and understand what’s actually happening, but actually tell you what the actual impact of what it actually is. So if you think about a hurricane versus an earthquake versus a shooting, to determine the severity of one of those is very different. So AI is able to actually determine what’s actually happening, the severity of what’s happening. And then it’s actually able to go farther and actually tell you what the impact to your business is by understanding where you actually have people, places, or assets so that you can actually take the appropriate action versus trying to wade through everything going on and understand if it actually happens to you at any time.

Mike Vizard: Can I play what if scenarios? Can I model various types of things so that I can compare it? What does that look like?

Dustin Radtke: So If you think about, let’s use a Canadian wildfires for example. As the winds start to shift, you’re starting to get news and different alerts coming out of many different types of systems to different areas. Let’s say it expanded up into the Midwest into cities like Chicago or Milwaukee. So if you can get advanced alerting on that and you happen to be an event venue operator, you need to understand what the impact’s going to be to any of your events. You may have a concert scheduled at a certain venue, what is going to happen to the crowd so you can better manage the crowd, maybe deploy the right resources, cancel events.
So you’ll be able to get that information very quickly so that you can take the action. And if you think about some of these things, it may be a first mover advantage. If you’re thinking of a freeze that’s happening and you happen to be somebody who’s sourcing strawberries from a certain region that needs to know about a freeze and you have a supplier in that region. If you can get alerted first, you can start to source from another region, another provider to be able to get the strawberries that you may need.
So it’s all about first mover advantage and being able to react before something actually happens that’s going to interrupt your operations.

Mike Vizard: How hard is it to set all this up? Do I need to be collecting all these data sources into some data warehouse somewhere, or do you do that for me? What’s involved?

Dustin Radtke: Yeah. The great news with a tool that OnSolve provides is we actually have all those sources pre-configured. We have all the proper training of our AI models. The only thing we have to do with our customers is actually set up their different facilities in our system and it starts monitoring and alerting them for them immediately.

Mike Vizard: Will it, therefore send an alert to me before I even know I have an issue and it may say, “Hey, there’s a storm forming off the coast of Florida somewhere, and this is likely to have an impact,” rather than me having to just kind of log in and hope that I discover something.

Dustin Radtke: Exactly, exactly. That’s how a lot of customers actually use the tools. They actually set it up to monitor for things that matter for them so that they can take action. And in things like that, whether it be a weather event, even protests or something that people like to monitor, because while some protests stay civil, some protests in different regions actually turn into more violent disruptions. So they want to be aware of different things that they can actually start thinking about actions they want to take. So they set up those alerts, they go to the proper people and they can make up the proper action plans and put them into place.

Mike Vizard: Well, how smart is smart then? Does it serve as suggestions for what I might do for my business processes, or does it tie into some sort of workflow automation process? How savvy is this thing?

Dustin Radtke: Yeah, you can definitely tie it into different business rules. Maybe you have rules set up that when something happens at a certain criticality, you auto alert the different people in that region. Maybe you have employees in a building, you may want to auto alert them to evacuate that building. If you get an alert around a bombing within a 500 feet radius of a certain building, those types of things can be automatically configured to actually take those rules.
And as you start to apply those rules, that’s where we’re continuing to invest and do some research on how can we actually automate through AI and learn from prior behaviors what somebody actually does, so that we can actually automate that in the future for them and provide suggested outcomes and suggested action plans.

Mike Vizard: How much is this real-time information? Because a lot of times we will build a plan and then, as they say, when first contact comes with the enemy, the plan isn’t worth what it was originally started. So how quickly can I adjust and navigate to changing circumstances?

Dustin Radtke: Yeah, I mean that’s one of the things that we’ve really focused a lot on with our platform is really being able to support that agile response process. Because if you think about it, everybody sits down and they have a group or team that actually creates response plans, but when actually something’s happening, what actually gets done doesn’t always follow what was written in some book five, 10 years ago. So we want to continue to evolve and help companies create and keep those action plans updated by creating these agile response plans to be able to report back, here’s actually what somebody did when this certain type of event happened.

Mike Vizard: All right. Should I just throw out all those books? Seems like they were written four or five years ago and maybe this whole thing can just go into an AI platform.

Dustin Radtke: Definitely, that’s where the future’s going for this, along with many other things.

Mike Vizard: As we go along, do you think we’ll get to the point where the platform will evaluate the plan that I created and give me a score for that as well and say this plan that you created is not as good as the average plan, or are there going to be scores and risks that I can assess?

Dustin Radtke: Yeah, there’s definitely opportunity to take and learn across a larger ecosystem as more and more people start using similar systems. We already do that today with some of our communication stuff. We can help companies actually craft the right type of message to get the right type of response when there is an emergency, so that you know certain keywords and trigger words that you do and don’t use based on historical patterns of response levels when people get an alert or get some type of action that they need to take.

Mike Vizard: What role is generative AI going to play in all this? Because of course, you can’t walk down the street these days without somebody leaping out telling you about their new great generative AI thing. But does that have a role to play in this? And I’m assuming your AI platform is built more on machine learning algorithms, but what’s the underpinning of this thing?

Dustin Radtke: We actually use a lot of different types of models. We do use a lot of the models that underpin some of those larger generative AI models. We use large language models. We’ve got a convolutional neural networks within our system, and we do all that, but we do it in a way that we make sure that it’s purpose trained because while tools like ChatGPT, they’re really great at very broad use cases, when it comes to going depth in a certain area, whether it be physical threat, you need to make sure that you’re actually getting very high accuracy with information and high relevancy because that’s critical for customers and they’re not able to do that because they don’t train on specific areas. And that’s where our differentiation really comes in, we continue to invest looking across all the right models for every specific use case. And we invest a lot in our AI pipeline.
Any true strong companies can have a very strong AI pipeline pulling in all the right models for the different use cases that they need to do that could expand and grow as their customers grow. And to your point just a second ago, it’s all about the data as well. So you can the greatest AI, but if you don’t have the data, which is where we’ve invested very significantly as a company, being sure that we pull the sources from very broad areas. Because a lot of times when it comes to AI, you want to make sure you have ground truth.
So we focus on how many different sources are reporting the same type of thing, what are they all saying? Where’s their commonality, where’s their differences, so that we can make sure that we use our AI appropriately and that we truly are giving the most relevant information in the fastest time possible. Because when it comes to any type of threat, speed is critical when you come in to do it, but also relevance, meaning making sure that actually is important to me, otherwise it just becomes noise to anybody using a system like this.

Mike Vizard: Who’s in charge of making these decisions within an organization, because it clearly goes well beyond IT? So who’s taking the leadership in having this kind of forethought to invest in something like this?

Dustin Radtke: Yeah, I mean, it all really depends on your organization. A lot of times it’s led by the security operations center or the CISO of a company. But sometimes depending on the size and scale of a company, because we really focus on making sure that our solution is available to all levels of company, it all depends really where their business continuity or their security operations or their duty of care lives. Sometimes duty of care lives under an HR function, matter of fact. You’re starting to see that change has become more and more important, but it all depends on the company, and that’s where we really need to focus as an industry on making sure that our tools are consumable by the average human to be able to run and actually maintain and operate.

Mike Vizard: What have you seen people do that kind of makes you shake your head a little bit and go, “I think we’re better than this, folks.”? What’s that one piece of advice you have for folks about how to go about really planning this stuff in a way that the results are what people expect?

Dustin Radtke: I think first is you have to realize really where you’re at. You would be amazed. We’ve done some reports and 60% of people have no plan to address some of the more severe threats to their business. So that’s really crazy. And that’s from a survey that we did across 250 US-based CEOs and only 37% said they have a plan for extreme weather and 29% have a plan for an infrastructure failure. So really it’s understanding where you have risks and making sure that you have tools so you can at least handle the risk. Maybe you can’t invest in a full business continuity plan for everything, but at least start to invest in understanding and monitoring for the threats that are out there, because that’s critical.

Mike Vizard: All right, folks. Well, you heard it here. When it comes to business continuity, there’s no substitute for, as the Boy Scouts used to say, being prepared. Dustin, thank you for being on the show.

Dustin Radtke: Great. Thanks, Mike. Have a great day.

Mike Vizard: And thank you all for watching the latest edition of the Techstrong.ai video series. You can find this one and all our other episodes on the Techstrong.ai site, and we’ll see you all next time.