It’s a brave new world. Self-driving cars are on the roads, computers trade stocks at lightning speed, and drones may soon deliver the mail. But where is artificial intelligence showing up in the property business?
“Automation is…using technology, machines, etcetera to automate processes that would have normally been done by humans,” said Brendan Canny, American Managing Director for property software and automation firm Open Box, at a recent Propmodo Live event. The next level beyond simple Excel macros is called “robotic process automation”, or RPA – the application of straightforward automated processes at the firm-wide, or enterprise level. The final step in the automation pyramid introduces artificial intelligence, or AI, in order to give those simple processes a spark of smarts – the building blocks of decision-making.
“…I guess maybe I’m coining a new phrase…IPA, or “intelligent process automation”, Canny said, referring to this final step in automation development. “And so really what that is doing is trying to take the concept of this enterprise automation, or RPA, and applying some level of computing intelligence or thinking to kind of come up with the best of both worlds.”
Canny explained this sort of automation development with the typical task of lease abstracting. With an IPA, the abstracter would first use an RPA system to automatically store and label a given lease. After that, the real intelligence kicks in: a computer program reads the scanned document and actually determines for itself what the different parts of the lease are – who the tenant is, who the landlord is, what the rent is, and so on. This is in marked contrast to a simple or “dumb” RPA system by itself, where the raw data must be parsed and analyzed by human eyes as usual. From there, IPA systems would be able to bridge the gap from the lease to the other software platforms that rely on the document’s data to function.
“So in theory, in this whole process the only time a human is intervening is just to review what has been abstracted by artificial intelligence,” Canny said.
After discussing the role of automation in the real estate field, Canny shifted gears to offer counter-arguments to a number of common automation fears and misconceptions. No, Canny said – robots aren’t yet ready to take away all the world’s jobs. Instead, he explained that automation processes today are performing those mundane tasks that waste time and hurt productivity, leaving professionals with more bandwidth to solve problems and increase value.
Additionally, Canny said that automation need not be limited to the realm of cost savings. The greater workloads of business expansion – more leases to abstract, more properties to manage, etc – would be prime targets for process automation. And while it may seem that automation is limited to the purview of company IT professionals, Canny added that modern automation system providers are able to work directly with business leaders themselves, and cut out the corporate IT intermediary.
However, Canny also cautioned against being over-optimistic about the role of automation.
“[In RPA systems] you as the human still need to define what are the rule-based, repetitive process that we are trying to automate. You can’t just automate everything,” Canny said, adding that human oversight is and will still be necessary to tweak and change systems as outside factors demand.
“And then importantly as well, they’re monitoring this process, right? So if something is stopping for some reason, there still needs to be that human applying that cognitive ability to figure out what’s happening in this scenario and this case, and how do we fix that,” Canny said.
For their part, Open Box’s own proprietary automation engine – or robot – Rob Sparke has shown “his” capabilities well, bridging gaps between software platforms without built-in integrations, cutting process lengths by tens of hours, and engaging employees more fully in ownership and standardization of their work processes.
Encouragingly, the road to automation need not be a long one, as Canny broke the implementation process down into three steps.
“The first thing is really trying to figure out…what processes could we automate?”, Canny said. After identifying the repetitive, rules-based processes that make for good automation candidates, the next step is to break these processes down into their constituent steps, one by one. Finally, the last step is actually integrating the new automated process.