IBM Think this year was virtual and while the speakers were well-rehearsed, the event length was just about right, I was again reminded that we still have a lot of learning to do to virtualize conferences effectively. One of the tools that could be used to do this is IBM’s Watson, which could help customize the content and make the presentations far more interactive and exciting at the massive scale IBM needs for this kind of event.
Watson is IBM’s near-term secret weapon, and at Think this year, the most exciting iteration of this platform was called AIOps. It reflects that IBM has finally started to apply this cutting-edge AI at scale platform to what CIOs care most about, running their shops. And the magic I think will be that once CIOs get more comfortable using Watson themselves, they will be more likely to deploy Watson in the other areas where it was already focused.
Let’s talk about Watson AIOps this week.
The Big Problem With IT At Scale
Virtually every large company CIO is a standing example of the Peter Principle, regardless of their skills and capabilities. Large companies have massive amounts of vendor and solution diversity implemented over decades that often interoperate poorly, if at all, and are relatively unreliable and increasingly difficult to diagnose and fix.
IT organizations tend to be underfunded and understaffed and must spend much of their time keeping their environment up and running and disappointing other executives, because they lack the free resources to make the progress that line executives are demanding.
Technology, by its nature, is a force multiplier, and AI by design, particularly IBM’s AI, is mainly focused on increasing the capabilities of the people that use it. Much of the initial Watson work was focused on healthcare, finance, and defense (which we don’t talk about). But this left the very people needed to deploy it without the resources required to get the job done and no trust for AIs (particularly given the belief that most AI projects fail).
To advance AI in the enterprise, you need to free up IT resources to implement and train the solution, and it just made sense to apply AI to this IT problem as well. And when you use AI as a force multiplier for IT, you effectively apply that force multiplier to the entire company making this one of the most potentially powerful solutions IBM has ever brought to market.
IBM demonstrated how this would work using a web sales example (there is a similar demonstration here). You see, and I’m sure most of us have been through this, typically when a web site starts breaking you don’t get timely notice, it may take hours to days to diagnose the problem, and then hours or days more to craft and implement a solution.
What they showcased is that AIOps immediately will identify the problem, forensically determine what the cause of the problem is, and then automatically craft recommended solutions. So, you get a holistic problem report, which included collateral damage and impact on the business, recommendations on a fix, and links that help you implement the recommendations.
While this is currently implemented in Slack, future versions will use other collaborative tools. In an increasing number of cases, as the platform gains intelligence, it should also be able to implement some, if not all, the fixes.
At some future point, it might be able to operate autonomously and, even now, in many cases, it should be able to identify pending problems that customers and employees haven’t even seen yet, and as a result, may never see.
Watson AIOps is –when professionally trained and fully implemented – the closest thing to a CIO superpower I have yet seen. Potentially AIOps should dramatically reduce or eliminate most of the headaches and problems most IT shops regularly have. This elimination frees up resources to provide more timely development solutions for IT’s internal customers and massively reduce the target on any CIO’s back that deploys the solution.
While still early, this has the potential to make the CIO job far more enjoyable and far less the masochists dream it has unfortunately become. This is where Watson should have initially been focused but following the rule of “better late than never” by addressing the CIO problem, it will form a stronger foundation for AI, and Watson, to be deployed more widely in any Enterprise. If this turns out as expected, this could become the most potentially lucrative, powerful, and beneficial tool IBM has ever created.