In Florida, locate requests scheduled for what month are the most likely to result in an incident?
Do you know the answer to this question? By analyzing extensive datasets using artificial intelligence (AI) and machine learning, Irth’s Insights with AI risk model determined that January digs are the most likely to result in an incident.
This is one example of what artificial intelligence can discern when analyzing extensive datasets for Irth Insights for Damage Prevention.
Matthew Abbitt, Chief Product Officer, and Kyle Murphy, Director of Science & Architecture, led the “Leveraging Artificial Intelligence to Improve Damage Prevention and Safety” webinar, in which they discussed this, other advanced strategies, and Irth’s tools for risk mitigation.
Over the years, customers have influenced Irth’s damage prevention offerings. Irth’s customers wanted:
Irth’s offerings for assessing risk are built to accommodate every customer, no matter how much historical damage data you have.
You don’t get just a number; you get an explanation for why that number is what it is.
This is where you can configure risk scores — if-this, then-this statements. Most clients use this for location-based scores. For instance, if you have a high-profile or high-consequence area, you’d configure a higher risk score for that type of ticket.
Customers use a configurable risk score or location-based risk score when there is not enough historical data yet for an AI model to capture the risk in those areas. For example, a new subdivision is being built and micro-trenching needs to be done.
You can have the system search for certain keywords on an 811 ticket, such as boring or blasting activities, which would cause the system to identify it as a higher-risk ticket. Alternatively, you can also reduce the risk score on a ticket by keywords. For example, you could set a rule that if your team is the excavator on the ticket, reduce the score by so many points. Anything on a ticket from facility location to distance to facilities can be configured to increase or decrease a risk score.
You can also use Smart Score with other data besides ticket details, including industry trends, locate audits, depth of cover, and more. Irth is the only product in the industry where you can look at external activities outside of excavations and have the risk of an excavation modified.
This analysis is based on your tickets, your damages, and your locators. It learns over time and modifies scores in real-time.
Irth’s DRA model refines risk score calculations using your organization’s performance data.
The risk score still evaluates the same variables for correlation with risk, but scores are customized to provide the most insight into how your organization operates — allowing you to better inform decisions, refine performance, and develop targeted preventative actions.
We’ve found the three most impactful variables in DRA to be:
At Irth, we use AI to:
Two main factors provide the best AI results:
Irth quantifies and predicts risk using proven methods also used by NASA and the U.S. Department of Defense (likelihood versus impact). Machine learning allows us to identify risk and quantify what information predicts an incident and how much riskier it is.
Our models are getting retrained every single day. This prevents drift, the systematic degradation of predictive models over time.
Irth Insights with AI is fast, easy to integrate, scalable, and effective. With a 2-second average response time, a ticket is received and scored in almost real-time, and then automatic notifications are submitted and follow-up activities are created based on a risk matrix.
With AI, our tools can proactively identify risk with predictive models, to give our customers immediate visibility into high-consequence/high-profile facilities, help focus damage prevention resources using real data, and automate steps to notify and assign the correct individuals or risks.