Six years ago, I spent the months of October and November preparing for my first conference presentation as a Gartner analyst. As a rookie analyst on a stage as big as the Gartner Data Center Conference offered, I wasn’t so much as nervous about being in front of a crowd as I was about the content I would be presenting. I was thrilled that not one, but two of my proposals were accepted, giving me the opportunity to talk about two topics that to this day remain near and dear to me: social media’s impact on IT operations and the future of the IT service desk.
I bring up this time frame because one of the driving forces of my second presentation was the role artificial intelligence (AI) would play in service management. Think back to 2011, the year that Watson competed on Jeopardy! It was impossible not to think that if IBM could develop a system so capable of answering questions posed in natural language that it could beat Ken Jennings and Brad Rutter, then surely it would only be a matter of time before IT organizations would find ways for AI to address the majority of annual service desk contacts. As I was developing the slide on Watson and outlining AI use cases, my manager (and ultimate peer reviewer) stopped me before I went too far off the rails.
“It’s not artificial intelligence,” he said. “The machines aren’t doing the learning. We haven’t arrived at our SkyNet moment. Tone it down a notch.” I removed the words “Artificial Intelligence” from the presentation, but did build a slide that included an image so I could talk to the concept with the Watson anecdote in play. (#PeerReviewHacks))
Market Forces Driving AI Adoption
I share that story with you because here we are at the tail end of 2017, and artificial intelligence is all the rage. AI is the number one theme for digital successful transformation, articulated by Gartner as “technology that can learn, adapt and improve itself without humans programming it (and) promises to help you accomplish more with humans and machines than either can do on their own.” Gartner, along with other research firms, certainly play a role in both raising awareness and shaping the development and direction of AI. There’s also Elon Musk’s take on AI, which, if nothing else, gets your attention for a little longer than a typical headline does. And the pursuit to monetize this transformative technology has been relentless, as seen in the financial commitments of tech giants that recognize AI is poised to unleash the next wave of digital disruption.
As business users increasingly expect a workplace experience that mirrors their consumer experience (think: Amazon Alexa and Apple’s Siri), the pursuit to monetize this trend within the business-to-business space has continued to accelerate. Tech giants, as evidenced by their financial commitments and product roadmaps, are betting big that AI will be the next transformative technology, poised to unleash the next wave of digital disruption.
Yet it’s still very early in the AI adoption cycle, and vendor claims to deliver enterprise-ready technology that will substantially impact your business today may be a bit ambitious. In the not-so-distant future, however, AI will be a natural element of our lives, much the way personal computers are today. I’m sure my baby boomer parents couldn’t imagine a world where almost two-thirds of the worldwide population has at least one personal computer in their pocket.
We’ve long seen such benefits delivered via automation, but with the promise of machine learning and virtual agents, the conversations have shifted from using dumb-machines that smart people program into really smart machines that program themselves. This has real ramifications for individuals who stand to be augmented by AI, and like Dorothy Vaughn in Hidden Figures, they’d be wise to understand how to get in front of disruptive technology before it gets in front of them.
Moving Beyond AI Hype to Practical Use Cases
For IT service management (ITSM) leaders, there is much discussion about how much AI will tangibly impact service support and delivery. There’s no shortage of content on the topic, and it can be difficult to move past the hype to understand what can be done to pragmatically apply AI technologies in the foreseeable future.
The promise of AI in ITSM is a beautiful concept. Natural language processing (NLP) will make it easier for your users and technicians to find the content they need to resolve issues faster, making them more effective. Machine learning (ML) will extract certain types of knowledge and pattern recognition based on observations. Virtual agents will use NLP and ML to assist your users and technicians by automating tasks for them. AI will effectively apply the rule-based calculation and sequential thinking aptitudes mastered by your best service desk analysts, technical engagers and programmers and enable machines to replicate what they do, so that humans can work more efficiently and effectively.
The first set of use cases will be simple in nature. The majority of vendors and solution providers will package and deliver mature technologies framed as AI to capitalize on early consumer interest. An example of this would be speech recognition technology, which translates human speech to text for further processing. While this technically fits within the AI area and is familiar to consumers from their interactions with Siri, Alexa and Cortana, we’re only scratching the surface in recognizing what Natural Language Processing can do in the enterprise. So instead of just asking your ITSM solution how many tickets are in your queue and getting an answer, imagine a world where you’ll also get the answer with context of how to resolve the easiest issues faster and what can be reassigned to other analysts.
As with all transformative technologies, however, the use cases get ahead of the technology, and the tendency to think that AI can solve a multitude of workplace challenges leads to the over-promising and the subsequent under-delivering of the capabilities. Organizations that are looking to increase profit margins want to understand if AI can help reduce the cost of their largest operational expense (people) without trading off quality or the end-customer experience.
How To Proceed In Light of Uncertainty
So what’s the bottom line for IT teams? How do service management leaders and practitioners alike lay the foundation for next-generation approaches to automation, given a technology that is as of yet still very immature, but evolving at a rapid pace?
I’d like to share my (and many others’) view on the minimum requirements for successful AI adoption, all of which are explored in greater detail in our latest eBook: Ready or Not, AI is Coming: 5 Practical Steps You Can Take Now to Prepare.
- Knowledge Management. The bridge to rich information analysis starts, first and foremost, with a solid knowledge management foundation. Gartner sees a future for AI in service management and predicts that through 2020, of artificial intelligence (AI) initiatives in IT service management (ITSM) will fail due to the lack of an established knowledge management foundation. I don’t just agree with this because of my tenure there—I agree with it because it’s true. What happens when we don’t appropriately build and structure the content required to make NLP use cases function properly? Or properly document the knowledge for machine learning to extract? Or provide the virtual agent the frame of reference to automate the workflow you asked it to?
- A Culture of Self-Service. The promise of AI-enabled ITSM might sound like the same promise offered by self-service, which alleviates the burden of costly technicians by pushing resolution back to the user. But that only applies when the user is trained to use self-service and finds self-service to be a faster path to resolution than contacting IT for support. Users can’t solely control and deliver that, it falls on both IT and the business to foster a culture of self-service that trains, encourages, rewards and promotes members who participate in its delivery.
- Agile / Lean Frameworks. The notion that IT must move faster is not new, but the application of lean and agile methodologies that work to simultaneously create value for the user—and have respect for people—plays a profound role in establishing AI-enabled service management. In addition to rapid prototyping, sprints and minimum viable products, the requirement to empower people also needs to be considered. How will AI will help (or hinder) IT’s abilites to create trust and foster respect for each other—as well as with the users and customers they serve?
- Enterprise Service Management. The “IT” in “ITSM” has been dropped by many organizations who recognize that service support and delivery aren’t exclusive to IT. Automating processes and standardizing outcomes at a favorable cost are things lines of business have raised their hands for, and the ability to apply AI technologies to reduce costs and increase efficiencies across the enterprise—far beyond the borders of IT—should come as no surprise. As IT organizations extend service management capabilities into lines of business, they should be mindful of the application of these technologies to improve the ways all employees work, not just the ones in IT.
- ITSM Platform Interoperability. In understanding that no single vendor will rule the artificial intelligence technology stack, interoperability is vital requirement for organizations that will desire the flexibility to work with the tools they get the most value from and reduce the threat of vendor lock-in. This also enables the freedom to work with new, emerging technologies—such as AI—as they see fit.
Like most disruptive technologies, AI will sit at the peak of inflated expectations for the foreseeable future, for early adopters to experiment with, while the rest wait on the sidelines to learn from their mistakes and apply accordingly. I would strongly recommend that approach, but it’s important to also recognize there are things you can do today to prepare the service desk for AI—so you don’t miss the AI party that is certain to continue for the foreseeable future.