Gerson Benker provides his insight on the importance of AI and automation in the healthcare field.
Gerson Benker is a believer in AI for the enterprise.
As Vice President of IT Operations at Carestream Health, a global provider of medical imaging systems and IT solutions, Benker oversees an IT environment that spans 100 countries, with internal operations servicing 7,000 users worldwide (in multiple languages) including field service and sales representatives. Carestream maintains an internal cloud-based platform for radiology and diagnostics information, as well as a dental practice management system that provides lifecycle services to dental offices, from making appointments through running diagnostics.
Benker previously implemented IPcenter to maximize Carestream’s IT operations. As an Amelia AIOps (formerly 1Desk) Beta customer, he’s now preparing to leverage the end-to-end automation and cognitive capabilities in AIOps with Amelia, the industry’s leading intelligent agent, for shared enterprise services across Carestream. Last week (Nov. 7), Benker appeared alongside IPsoft CEO Chetan Dube as he unveiled AIOps at Gartner Symposium/ITxpo in Barcelona. We spoke to Benker about the importance of AI and automation in healthcare, and his insights on how AIOps will impact his organization.
Why are AI and automation particularly relevant to healthcare?
In the healthcare sector, there is very little margin for error. It is a very heavily regulated environment, an environment that needs fast action. It needs action that is spot-on the first time and you don’t get a lot of chances to get things right, especially when you deal with patients and patient diagnostics. It’s important to automate as much as you can to take human error out of the equation.
As an AIOps Beta customer, what kind of operations will AIOps support?
Things like end-to-end user provisioning. For example, a new user comes onboard, like a new salesperson in a different time zone [from IT support operations], and needs to be onboarded fast without any mistakes. Another example is permissions in our cloud for nurses and for doctors to look at the right kinds of [medical] images, and permissions for access to the right kinds of data, which is very important when it comes to data privacy and HIPAA laws around the world. You cannot make a mistake when providing access to doctors and hospital staff. Imagine a mistake that gives access to patient data to the wrong hospital. These mistakes won’t happen in [AIOps].
Where will the speed of AIOps help the business?
For example, a doctor in a hospital who needs access to a system could put in a [AIOps] request in the middle of the night and within minutes this could be handled without the involvement of any service desk personnel. We’ve centralized our IT Operations into a single location in the US, so with everything in one country, this can cause delays if you’re in another time zone. Those delays are automatically out of the picture with [AIOps].
How do you manage the organizational change required to absorb automation and cognitive technology into your operations?
Often, people don’t want to change what they have done for past 30 years and resist automation. You need to convince them their skills are better used for higher-quality tasks and projects. For their own careers, they need to get away from repetitive tasks and reskill and learn new things for the future. Then there are younger millennials who come into our company who actually don’t want to do all of these administrative tasks. As the workforce shifts, in order to keep up with the younger workers, you actually have to automate repetitive tasks, because those workers don’t want to do these things anymore. It’s going to be hard to attract and retain younger workers unless you automate, and the pressure to introduce AI technology will only increase.
The first step with IPcenter was a big step, because it was a new interface. You have to do [AI] in phases, and pick people who are advocates and who are open-minded, who can be champions for these technologies within their own divisions. With Amelia and [AIOps], we think the technology will be so good that users won’t really be aware of the implementation, they’ll just notice that it works and delivers benefits.
What kind of skills are you looking for in order to take your AI/innovation strategy forward?
We’re looking for skills like requirements gathering, understanding business operations, and helping business partners translate business needs into an IT solution. It’s not just about people executing in the back room. The days of executing tasks in the back room are long gone. No-one should be planning to run a datacenter — the datacenter should run itself. People who want to create innovative solutions, who are good at managing projects and development work: those are the people we are looking for.
In your opinion, what makes AIOps different form other technologies?
The end-to-end automation and running entire processes at scale. That the system is teachable, and that it can learn. You’re not stuck at one level, there’s no specific customization required. If you teach [AIOps] to solve your problems, it will retain that knowledge and get smarter as time goes on, and the automation will automatically increase over time. No other system that I’ve looked at or that I’m aware of in the industry can do that. [With other technologies] you can program certain customizations, but they do not have the self-learning capabilities or the end-to-end automation like [AIOps].
Ultimately I just want people to see the benefits. Business users shouldn’t have to know too much about the technology — it should be so good and so smooth they just don’t notice, and systems simply work.