Sentient IT – artificial intelligence meets IT department

I was reminded about the old saying about cobblers children and their shoes, when listening to a representative of a well-known IT service management solution on stage. He was praising the new version and all the fancy features to rapidly integrate it, transfer data and overall improve the end user customer experience. Strangely, one thing was missing though. I stood up and asked him: “how are you utilizing artificial intelligence in the processing and recognition of the incidents?”. The silence that followed was palpable.

IT departments are currently requested to provide artificial intelligence based solutions to customer service, process optimization and various other service or efficiency improving tasks. Why are there no articles or thinking on improving technology support or IT operations? It can’t be about availability or money – MS Azure and Amazon offer Machine Learning studios where the algorithms can easily be accessed and utilized.

Future of IT is sentient!

The future vision of IT operations should be a sentient IT: A set of automated data collection, software robotics and machine learning, which automatically detects problems in the IT environment and resolves them either by itself, or by ordering a supplier to do it. Application portfolio optimization? No problem, let the AI just collect data from the end users and create a plan. Recurring problems? Non-existant, because the AI reacted to them already before they turned in to problems.

How I see it, IT will have three hurdles in reaching this vision:

1. Missing data
2. Availability of skills
3. Willingness to change

Missing data

Primary problem with all artificial intelligence is in providing it enough data for learning. The basic data collected by IT today (ticket data, event logs) are not even nearly enough to describe the processes of IT sufficiently for the AI to learn something other than trivial tasks. In order to fix this, IT needs to collect massive amounts of data about the technology environment, users and their interactions. This means utilizing solutions like Applixure to monitor and analyze the environment, recording IT support interactions and even recording end user interaction on their machines using keyloggers.

Availability of skills

Normal IT organizations are usually full of IT generalists, who understand the company technology and the relevant business functions, and know how to keep the suppliers in check. When new technologies are available, they are normally the domain of the CTO or a technically minded CIO. AI skills are not present in IT, due to the fact that they are mostly useful for business analytics. They are also notoriously difficult to source, as AI is currently at the top of the hype curve.

Willingness to change

Nobody in a safe corporate job wants to make their job obsolete. Selling a vision of an “unmanned IT” is nearly impossible to the people whose jobs would be on the line. Looking at the corporate perspective though, it would be a very appealing prospect. Naturally, it requires that business builds some core technology skills in to ensure AI is making sound decisions.

How to get sentient?

There are various paths available, depending on the IT maturity. One thing common to all is that you need to have a vision and a plan. Vision to describe what the future model of IT will look like, while the plan describes the broad steps that are required and the approximate effort and benefit of each step. Most companies will start with setting up their cloud-based data collection capabilities (with tools such as Applixure, ), then moving on to create a small pilot service to kickstart change management and education activities. From there on, it’s a matter of agile development according to where the biggest value for money can be found.

Does the sentient IT sound like fiction? Maybe it is today. Tomorrow, I promise you, it will be here. If you want to understand how, let’s have a chat!