Our team adopted a new strategy last year in an effort to strengthen our alignment to Adobe’s business, by embedding cloud-like characteristics into our DNA. This meant not only reconsidering what we do and how we run applications and deliver services, but also determining how we could apply characteristics of the cloud – like its ease of use, reliability, and resiliency – to everything we do in IT.
We wanted to become the easiest group to work with, which meant eliminating IT as the point of contact for some things – but not everything. This set us on the path to exploring and experimenting with IT automation.
[ Want automation lessons learned from your peers? See our comprehensive resource Automation: The IT leader's guide. ]
When you mention the word “automation,” everyone’s visceral reaction is to wonder what will happen to their job. But once our team saw the outcomes – that they could participate in the future of thought, experiment with new technologies, and focus on honing new skills – IT automation became a real eye-opener.
Practical applications of AI and machine learning
Machine learning and artificial intelligence are both core to how Adobe is delivering experiences to our customers. In the IT organization, we’re working to take those same ML/AI principles and capabilities and inject them into what we do in IT to make us easy to work with, reliable, and resilient.
When we first started experimenting with machine learning and artificial intelligence, we looked for problem patterns on the operational side for which automation could provide a solution.
One place we found this opportunity was in help desk tickets, which we discovered through ticket analysis. By using artificial intelligence and machine learning to comb our base of information, we discovered that every year, in July and December, our help desk received an influx of messages, calls, and tickets to assist with password resets. This coincided with our company shutdown weeks, which also happened to be when certain employee passwords expired.
Once we discovered these correlations, we sent automated messages to staff a few weeks before the shutdown to remind them that their passwords were about to expire, and to change them before we closed. Now we no longer have that spike in tickets, and users are happier because they’re not waiting in the queue to reset their passwords.
Artificial intelligence and machine learning give us the ability to identify patterns that we can automate and self-heal. When something new doesn’t fall into the pattern, it’s added to the library so it can be learned and automatically fix itself.
We’ve used these technologies to find and self-heal other patterns, too. When customers visit our website and want to purchase a product, they need to complete a payment, registration, and download process. In looking at our back-end, we found some delays in those workflows, which translated into a delay for customers in purchasing Adobe software.
Our self-healing platform looks for those abnormal events in the customer journey, and once detected, they instantly self-heal without any delay for the customer. In both of these examples, IT automation has freed up time and resources so our team can focus on bigger and better things.
Building new team expertise
To support our IT automation efforts, we’ve built up a team with a variety of necessary areas of expertise. We have data and data analytics teams, with data scientists built into them. Within our infrastructure and operations team, we have people with particular areas of expertise who are working on advancing our operations within Adobe IT.
These teams are responsible for things like working with a data-driven operating model to improve internal decision making; looking for ways to apply data science to how we run operations within IT; building intelligence around a self-healing platform; and developing intelligence around testing.
The operations team, for example, built out a big data automation framework, which we call the Hadoop Test Factory. It leverages open source to drive efficiencies for testing and validating use cases. The Hadoop Test Factory resulted in 90 percent reduction in overall testing and effort time, which allows us to scale and be much more efficient. As a result, our quality improves and our time to market for delivering capabilities is reduced.
All of these improvements, however, mean that some people’s responsibilities are changing. In the ticket analysis example, some team members have shifted to working on solutions versus dealing with the actual ticket. This has opened up time and resources to really consider how we can make the experience better, versus focusing on solving issues. IT automation frees up the ability to think about the future – not just what you’re doing in your day-to-day.
Want more wisdom like this, IT leaders? Sign up for our weekly email newsletter.