This article originally appeared as the cover story in the March 2021 issue of Security Business magazine. When sharing, don’t forget to mention @SecBusinessMag on Twitter and Security Business magazine on LinkedIn.
In an increasing number of industries and vertical markets, companies are already using Artificial Intelligence (AI). It is part of their digital transformation initiatives to dramatically change the way their businesses work and bolster profitability. The trouble is that security industry companies are scarcely participating in those transformations.
AI is poised to be become history’s single most transformative influence on the physical security industry – but not in ways that derive from traditional security thinking, which is what gave us the status quo. It is transformation, not status quo, that AI is about for most businesses.
Internal IT functions and their service providers are already engaged in implementing AI for business use-cases in digital transformation projects. AI is a key transformative element, and the IT folks like calling video analytics “computer vision” – something that sounds right up IT’s alley; in fact, that is one reason that Amazon has labeled its AWS Panorama video analytics technology platform as computer vision (read more about it at www.securityinfowatch.com/21203570).
Both IT service providers, as well as end-users’ internal IT, are positioned to provide and provision network cameras on their networks. The good news is that security integrators – as opposed to many IT pros – can easily understand both AI business use cases and their security use-case counterparts. However, it is important to remember that ultimately, digital transformation is about business transformation, not technology transformation. It is about enabling the future vision of the organization, which should include business insights the likes of which most organizations have never considered.
Physical security AI can be a valuable part of that vision. For those in the industry who sit back wait for the trends to unfold and let the impacts drive change in their businesses, the likely outcome will be growing business irrelevance and eventual extinction.
Embracing Computer Vision
The computer vision concept is something that security integrators should embrace, but it helps to put it in the proper context.
“In security we have had eyeballs (cameras) for a long time, but eyeballs are useless unless they are connected to a brain that can make proper sense out of what they all see,” explains James Connor, Head of Corporate Engagements at Ambient (ambient.ai), a company founded in 2017 by AI experts. “Putting AI behind all the cameras at a facility would provide the context for what is being seen. It can be done while also maintaining privacy.”
Computer vision ties everything together – creating what many would consider a smart building. Connor says computer vision is technology’s answer to ending access control false alarms, while also enabling other building systems to interpret and act on data from security-related technologies such as thermostats, smoke and CO2 sensors.
“The extent of this AI utility cannot be understated,” Connor says. “It would put more overall intelligence and business value into intelligent buildings. 2021 (may) be the year that physical security stops its continuous technology ‘iterations’ and begins truly innovating with AI.”
Convergint Technologies is one global integrator company that is not sitting back and waiting for the trends to unfold before acting. “I really like to think about AI in the context of computer vision – obviously, I didn’t invent that term, but I have spent much more time in the last year studying outside of our industry, and I see major initiatives from Microsoft, Amazon, Google and others beginning to systematically enable computer vision technologies,” explains Eric Yunag, Convergint’s VP of Technology & Innovation. “As we advise some of the more forward-thinking customers, we think about what digital transformation, AI and computer vision mean to their businesses in a broader context.”
Yunag explains that he asks these customers: What if you and your smartest people could stand here 24/7/365 and observe what goes on and log it to achieve full insight into the behavior and activities over time, and how would you optimize or better understand your most important customer or employee processes? “We know we are just scratching the surface on how we would turn those observations into data, how we would act on those insights, and ultimately the value that creates for customers,” Yunag says. "That’s the kind of discussion that company leaders want to have about digital transformation and how we can make the case for additional investment in visual intelligence infrastructure.”
Technology Creates Full Situational Awareness
Ava (www.ava.uk) is an AI company focused on the full spectrum of security – starting with Ava Cyber in 2016, and Ava Video (formerly Vaion) in 2018, merging the two in 2020. Ava provides both an AI-enabled cloud VMS and cloud-connectable cameras designed specifically to run machine-learning analytics in the context of full facility understanding.
Ava CEO Tormod Ree uses the term “smart analytics” to refer to AI-based analysis in surveillance cameras. “It is important to provide value beyond security – for example, helping large retail stores optimize their stock levels and staffing based on which department has the most traffic at that moment, or ensuring social distancing,” Ree explains. “Integrating with other systems like access control, you can automatically see the person trying to enter real time or automatically disable a door to prevent additional entrants when maximum occupancy is reached.”
Readers may have noticed that two leading AI entrants into physical security, Ambient and Ava, have designed their systems from the start to run AI on all cameras and provide full facility situational awareness.
Several years ago, Bosch Security begin building video analytics into every camera – a basic set of analytics on all cameras, with additional analytics on its more advanced line of cameras. Bosch, too, believes that analytics should be run on every camera.
Although running AI on all video streams has always been the ideal situation in the minds of security practitioners, the technology to do so at feasible costs simply has not existed until now, thanks to Open application APIs.
Open APIs are a fundamental element of the digital transformation, which leads to achieving facilities that are truly operationally intelligent. Interestingly, many of the security industry companies with advanced and secure open APIs – Ambient, Arcules, Ava, Brivo, Eagle Eye Networks, Lenel Systems, Milestone Systems – all entered the security industry from various IT domains, and all of them have a goal of bringing expanded business value to their customers through their products and open API integration.
We are well past the time for the physical security industry to fully embrace information technology and its development and deployment practices – including facilitating business value through open API integration. Manufacturers and security service providers who fall short in this regard will find their businesses falling short of what their now-tech-savvy customers expect both in technology deployment and in business results.
Security AI and ROI for End-Users
Within many customer businesses, security AI-related projects will be viewed from the context of the organization’s overall digital transformation AI initiatives.
The retail industry is often the first stop in illustrating the value of video surveillance technology coupled with AI and machine learning. A recent whitepaper from STATWORX, a consulting company for data science, machine learning and AI, points to 35 different AI and machine learning use-cases for the retail industry – two of which are about camera-based computer vision (access the free whitepaper at www.statworx.com/en/data-science/whitepaper). The whitepaper outlines how video cameras can be used with AI in cashier-free stores, as well as for automatic detection and correction of empty retail shelf space.
Many security industry vendors, including Axis Communications, Bosch and Eagle Eye Networks, have already proven that retail business use of security video cameras is highly profitable for end-user customers. One use-case from Eagle Eye features a Domino’s Pizza franchise that recently deployed body-worn cameras for delivery drivers via the Eagle Eye Cloud VMS – which found that the ensuing positive changes in customer interactions resulted in a 28% increase in sales.
Business security improvements can have other kinds of significant positive financial impacts, such as improving employee retention rates; however, that kind of security contribution to business value has rarely been examined in the past, partly due to the difficulty of making such business correlations.
Deon Chatterton, former Senior Manager of Corporate Security Technology for Cisco Systems, calls the emerging generation of AI-enabled video systems “enterprise video,” because the information generated by such systems is of value to many functional areas of a business, not just security.
“Enterprise video requires a completely different customer engagement than what traditional security technology has utilized,” Chatterton explains. “Information technology will continue to evolve rapidly to further enable business insights, operational efficiency and effectiveness, and improved employee and customer experiences.”
For businesses, corporate digital transformation mandates that these kinds of data correlations be discovered and that business insights be brought front and center – and those kinds of discussions inside customer organizations position the security leader as the rightful custodian of computer vision and visual intelligence technology.
IT will own the business-side data integrations and applications – which are more complicated but also are types of data and business applications more familiar to them; however, whether it is in the cloud or provided on-site as a managed service, the security leader is put in the position of owning the AI computer vision use cases and being empowered to ensure their full realization.
With these demands, often the only way for security integrators to enable their security director customers to achieve this is via a managed services model.
The Impact on Integrators
Security systems technology must continually evolve in place to keep pace with these new demands of enterprise end-users; in fact, many businesses will no longer invest in technologies that simply go obsolete at the end of their lifecycle. Instead, these end-users want technology that advances with the business – which requires a managed services approach that includes a highly collaborative component whose purpose is to keep increasing the technology’s value to the business.
Security integrators and consultants (who should consider teaming up for some of these opportunities) need to help their security practitioner clients transition to using managed services, because it is the only sensible approach given the accelerating pace of technology and the emergence of high-value computer vision applications. Managed services not only enable technology deployments to stay technologically-current, they enable continuously increasing business value from the improvements.
Yunag says this puts security integrators in a very strong position. “Security integrators are knowledgeable purveyors of visual intelligence technology and have decades of experience in maintaining mission-critical visual intelligence networks,” he says. “Furthermore, their customers have been operating that visual intelligence technology and safeguarding its data throughout their careers.
“The first conversations with customers ought to be around the concepts of visual intelligence, and what their organization could do if it had a company-wide internal visual intelligence resource – for example, an AI-enabled enterprise-networked video system,” Yunag adds. “The discussions that the security practitioner customers would then have with other business leaders would be along the lines of exploring what kinds of visual intelligence would be of value. Convergint is bridging that gap, and while I am not saying such discussions are highly fruitful every time, there is always value in having them. Sometimes it is immediate, sometimes it pays off further down the road.”
Six Steps to AI Monetization
AI Readiness is a journey that security integrators will travel with customers, with the integrators providing thought leadership and guidance as well as technology services to bring their customers new kinds of security and business value.
To monetize the opportunities that physical security AI brings, integrators must take the following steps, while including customer collaborations that involve covering new ground for business value discovery. Be sure to take it one step at a time, at speeds that are harmonious to each customer. Things will speed up as stakeholder interest rises and physical security AI use-case successes become demonstrable.
1. Learn more about the managed services model. The Managed Security Service Provider (MSSP) program from PSA Security Network (www.psasecurity.com/mssp) is already helping integrators make the transition. Learn more about the program from Dan Dunkel, Managing Director of the program, in his Security Business article from 2019 (www.securityinfowatch.com/21106456).
Also, read lessons learned from Security Business coverage of other integrators who have made the transition, including Communications Engineering Corp. (www.securityinfowatch.com/21133719) and Star Asset Security (www.securityinfowatch.com/21106432).
2. Discover physical security AI use-cases. Find them and establish the security rationale and business rationale for each. There are many business benefits to security, and these should no longer be ignored. Identify the low-hanging fruit and begin moving forward on those in support of security objectives. Some of these opportunities are explored in Steve Surfaro’s article on page 36 of this issue.
Another of these opportunities, as mentioned earlier, is AI-enabled VMS applications capable of eliminating the avalanche of false/nuisance alarms – many of which typically result from the use of video motion detection. AI-based false alarm reduction is now available for any video deployment regardless of what VMS or NVR/DVR appliances are in use.
Calipsa (calipsa.io) offers a cloud-based deep-learning false alarm reduction application that typically achieves between about 80% and 95% false alarm reduction using only three images of scene motion activity spaced one second apart (read more about it in an article from the October 2020 Security Business at www.securityinfowatch.com/21154542). This is a good example of security AI “low-hanging fruit” that can help customers start on their path to broader AI adoption.
3. Get details on the progress of the customer’s digital transformation. Learn about scope, status and direction of the customer’s digital transformation efforts and their key business AI use cases. Physical security AI will be a small but important part of the company’s overall initiative and should be well-aligned with overall AI objectives as well as internal data governance and data stewardship programs to achieve a frictionless deployment.
A helpful frame of reference for understanding customer AI initiatives is STATWORX’s “A Maturity Model for AI” whitepaper, also available at the link mentioned earlier.
4. Discover business computer vision use-cases and prioritize them with customers. Use-case prioritization is a key success factor in digital transformation. Explore business-side computer vision use-cases and identify the key stakeholders. To achieve frictionless adoption of business-value computer vision use-cases, collaborate with these stakeholders on prioritization according to established criteria; in fact, high-priority use-case categories often have earmarked funding.
5. Manage progressive AI adoption. Computer vision AI adoption can start with immediate security AI use-cases, and then progress over time to include business-value use cases. This can easily be a two- to three-year timeline to align with business activity, transition to managed services (OpEx) funding or hybrid funding (CapEx and OpEx mix). If success is defined as meeting objectives on schedule, make sure that schedules are not overly aggressive, and recalibrate them as necessary for changes in business priorities and overall digital transformation progress.
6. Document and celebrate success. Business sponsors and AI implementation stakeholders all need to be acknowledged and thanked for their roles in making success possible. It is helpful to all involved, and inspiring to those involved in other AI initiatives, to hear about the success factors as well as lessons learned. Also remember that AI adoption is a long-range proposition, since AI will continue advancing in ways we cannot possibly foresee. It is a journey, not a project.
Ray Bernard, PSP CHS-III, is the principal consultant for Ray Bernard Consulting Services (www.go-rbcs.com). He is the author of the Elsevier book Security Technology Convergence Insights, available on Amazon. Follow Ray on Twitter: @RayBernardRBCS.
