23 AI Strategies for Making Your Life in IT Suck Less - Part 1

Unless you’re truly a tech newbie, you’ve probably heard some kind of hype around artificial intelligence—often referred to as AI—machine learning, predictive analytics or a whole host of other words, terms and phrases, all pointing toward a similar conclusion: Emerging technology can do some pretty amazing things for people, teams and entire organizations.

The strategies and plans for making the best use of technologies like AI can vary widely. It can be hard to make sense of it all, let alone pick the ones you want to investigate deeper.

So how do you get started?

We have 23 tips, tactics and strategies for how you could bring artificial intelligence into your organization...with impact. In this part, we’ll cover the first 12 tips for you.

(If you’re already well-versed in AI and want to jump ahead, we recommend checking out our "AI for IT: Preventing Outages With Predictive Analytics" guide.)

12 Tips to Bring AI into your Organization

1. Start small by automating routine practices

To win in today’s market, teams have got to be fast; IT is constantly looking for the right combinations of tools and technology to bring teams the efficiency they need. Here’s where automation comes in. (We know, we probably isn’t the big and flashy kind of thing you first imagined when you heard the word “AI,” but automation remains a powerful and pervasive feature of technology platforms, applications, and tools.)

One quick and often low effort win is to identify the routine practices that take the most time and let machines take over the work. This can streamline the practices and operations of a particular team or even the entire organization.

2. Or go big by combining automation, monitoring and service desk

While there’s certainly something to be said for starting small, sometimes the rewards are even better when you go big. A powerful way to go big with artificial intelligence is to combine automation, monitoring and service desk on a platform positioned for AIOps.

This kind of combination makes automated monitoring and detection of systems and services possible, so the right teams can know where to help and when.

3. 360-degree views across business and IT

Siloed views of business and IT prevent either team from effectively driving revenue, protecting their brand, and delivering exceptional customer experiences. Constructing a 360-degree view across both puts all of the consumable metrics you need in a single consumable place, so the entire organization can move with greater agility.

Consider platforms and partners that can bring data together from many different sources and sets, so when you apply AI, you’ve got the complete view to work with.

4. Use AI and machine learning to reduce event noise

On average, out of the 1,200+ monthly IT incidents, 5.1 will be critical, costing IT money and the business revenue. Are your teams empowered to cut through the noise, so they can respond and remediate quickly?

Use AI to ingest all of your event data and let the power of machine learning point the right teams to the right issues, right away...or even before they have a chance to happen.

5. Shift from reactive IT to proactive and predictive IT

Who doesn’t like to win? And to win, oftentimes you’ve got to be first—first to market, first to a new trend, or sometimes first to resolve a problem. But imagine if things like problems didn’t require teams racing to understand, investigate and then act, but instead, they got a signal that something was going to happen before it actually became an outage?

With artificial intelligence, your IT teams can shift from reactive to proactive or even predictive. With the right platform, you can ingest data from across all of your infrastructure and services and let the technology tell you when an anomaly in one service will eventually lead to a degradation or outage in another. With a platform for proactive and predictive IT, you can decrease the amount of time to investigate, detect and resolve, and even get to a negative MTTR (mean time to resolution) by predicting the problems before they occur.

6. It’s okay to play the field...with algorithms

Well-constructed algorithms can be the backbone of a winning AI strategy. So don’t restrict yourself to a single approach. Pick technologies, platforms and people that can work transparently, ones that make clear the kinds of algorithms available and help you select the best ones for your problems. Bonus points to the providers that help you create and use your own models.

7. Train your models

Training with AI might be a bit like training for sports—the better you train, the better your performance and results. So make sure the AI technology you’re relying on lets you experiment and train your models with your specific sets of data.

8. Detect anomalies

With the right platform—one that’s capable of ingesting all of your data—you can use things like artificial intelligence and machine learning to not only detect anomalous or outlying behaviors and events, but to also automatically alert the right teams and people.

9. Consider trends and thresholds

It’s easy for some applications of artificial intelligence and predictive analytics to call something anomalous. But what happens when those data points aren’t anomalies, but are instead just parts of how your business ebbs and and flows? Some of the best AI technology ingests diverse data, helps you identify trends and lets you define what kind of events and behaviors are outside of certain thresholds.

10. Don’t miss the forest...reduce noise by clustering

When you can identify trends in your data and understand how they compare to each other, it makes being proactive even easier. You can use artificial intelligence to identify data peer groups and correlate events, so you can reduce event noise and help your teams be efficient and proactive.

11. Forecast to enable planning

What could you do if you could predict service health, capacity planning and maintenance requirements...all before a problem even occurs? AI and predictive analytics can help you put resources where you need them before there’s an outage, so you can maintain your SLAs and keep your business partners happy.

12. Prediction is different from forecasting

Part of making the best of AI is learning just a little bit of the lingo so you can act confidently. Not all AI capabilities are the same, and the difference between forecasting and predicting is one great example. Making a forecast usually means relying on a single pattern or data set to anticipate a future value—think about it like the rising and setting of the sun. Making a prediction usually means relying on a broad and complex set of interconnected and interrelated data to understand and anticipate where and how a chance in one might affect another—think about it like a dirty oil warning light that signals a machine’s imminent breakdown unless the oil is changed immediately. Prediction can sometimes be more powerful than just forecasting, so it’s important to know the difference.

For tips 13 through 23, check out Part 2 of the series!

If you’d like to jump into deeper learning on how AI can transform your IT operations, check out our handy guide AI for IT: Preventing Outages With Predictive Analytics.

You’ll learn about:

  • The advantages of predictive and preventative IT
  • The components of predictive analytics
  • What to look for in an AIOps solution

Bryan Jennewein

Posted by