Facilities security leaders must be deliberate in how they deploy AI-enabled features to unlock value or they risk bloating their operations with more data, alerts and complexity without outcomes that justify their investment, according to a report by Convergint, a large global systems integrator.
Many facilities already have complex security systems in place but they’re not used effectively because they are comprised of legacy tools that are largely operated manually and others that aren’t integrated well, among other weaknesses. “The result is an enormous installed base — some of it capable technology operating well below its potential,” says the report.
“When it comes to the security business, [many] legacy systems are not really connected, easy to use or easy to extract data from,” Amir Schechter, Convergint’s executive director of innovation and technology solutions, said in an interview with Facilities Dive.
In its report, Convergint outlines a framework it says can help security leaders understand what they already have, what they need, how to make everything work together and how to leverage AI so security becomes preventive rather than reactive.
“This is where a lot of customers are struggling,” said Schechter, “especially if they have complex environments that have multiple access control systems [and] multiple video management systems.”
The company organizes the framework in five progressions: detect, describe, explain, recommend, and act.
“Detect” refers to systems that can detect previously defined security events in real time across cameras and sensors. “Describe” pertains to systems that classify and summarize incidents by evidence, locations and timelines. “Explain” and “recommend” refer to systems that validate incidents by correlating data with risk patterns and recommend interventions, respectively. “Act” refers to systems that can execute pre-approved response workflows while maintaining audit trails.
Most organizations operate at level 1 or 2, meaning their systems can detect and describe incidents, but they struggle to reach the higher levels because of poor integration and other problems, according to the company.
It’s “not because the underlying AI capabilities are immature, but because the industry has been unable to connect capability to outcome via a structured path,” the company says.
“This is where it's a little bit harder to get all the data in an easy format, or to maintain even the integrity of the data in a way that's going to be clean and synchronized,” Schechter said. “This is where you see a lot more challenges to really aggregate all the data from multiple systems.”
System integration challenges are compounded by facilities not having enough skilled people, time or budget to assemble an end-to-end solution while maintaining uptime, cybersecurity and operational performance, the report says.
Teams are being asked to support more sites, devices, alerts, reporting and governance without proportional headcount growth, it says.
At the same time, security leaders are being pressured to articulate an AI strategy, “adding a strategic communication burden on top of the operational one,” the report says.
How to deploy AI for security in a smarter way
To make the framework actionable, the report says, security leaders should follow a series of strategic logic steps: prioritize, contextualize, build, govern, diagnose and execute.
“The path is not self-executing,” the report says. “The organizations that will realize these returns are not those that purchase the most advanced technology. They are those that build the foundations to absorb it — clean data, integrated systems, defined outcomes, an operating model that evolves with capability, governance that earns trust in automated decision-making, and a roadmap that sequences investment by impact and readiness rather than by vendor release cycle.”
“We end up having to have … conversations with our customers … and say, ‘Okay, here's the big fluffy stuff that you hear about,” said Mark Schweitzer, manager of solutions strategy at Convergint. “Here's how the rubber actually meets the road in your instance.’”
Schweitzer recommends facilities managers learn about what their peers and colleagues are doing to integrate security. “A lot of times it's knowing what others are doing in industry,” he said. By knowing what use cases others in the industry are applying AI to, facilities managers can better respond to requests from senior management to implement the technology, he said.
“The whole key [is] being able to come in and provide some practical use cases for them,” he said. “Rather than just saying AI is the greatest thing since sliced bread, actually understanding these are some use cases. This is [the] product. This is how you go about implementing it. Just to get them started and get them thinking about it, because right now it's such a high level concept,” he said.
Schechter said he sees the industry moving to provide tools that will make the process more simple while creating more interoperability, which will enable security leaders to get easier access to data.
“The expectation from a lot of the technology partners is to have a much more open interface and API so you can actually get the data and not go through hoops,” he said. “We see that now starting to be front and center when it comes to evaluating what technology you choose.”
