Like a lot of emerging technologies, the rise of AIOps is one of those IT trends that is about to sharply cut both ways for managed service providers (MSPs).
While there’s some dispute as to where the line between automation and IT management platforms infused with machine learning algorithms might lie, no one at this point disputes the fact that machine learning algorithms will play a large role in managing IT.
A survey of 200 IT managers that have already implemented AIOps at organizations with more than 500 employees was recently published by OpsRamp. The survey from OpsRamp, a provider of an AIOps platform, suggests that IT organizations are much further down the machine learning path than most MSPs would give them credit for. The survey finds that 87 percent of respondents are deriving value from their AIOps investments.
Specific areas where those investments are starting to pay off include more intelligent alerting (69 percent), root cause analysis (61 percent), anomaly threat detection (55 percent), capacity optimization (54 percent), and auto-incident remediation (51 percent).
The tools and processes currently used by these internal IT teams span domain specific tools (72 percent), AIOps-based event correlation (60 percent), rules-based event filtering (52 percent), incident management performed by a third-party MSP (49 percent), and manual processes (48 percent).
It’s too early to say to what degree AIOps might supplant any one of these functions, but the survey finds operational benefits derived from AIOps include automation of tedious tasks (85 percent), suppression and deduplication of alerts (80 percent), and a reduction in the average number of open incident tickets (77 percent).
A full 47 percent say they are now resolving IT issues at least 50 percent faster, while another 40 percent are resolving issues 25 to 50 percent faster.
The survey also notes that 45 percent of the respondents built their own AIOps platform, compared to 55 percent that relied on a vendor. There’s no doubt some IT services firms played a role in helping these organizations embrace AIOps.
MSPs must lead the way for their customers
IT organizations don’t appear to have fully made up their mind to what degree they trust the recommendations being made by AIOps platforms. Data accuracy was listed as the biggest obstacle to adoption (67 percent), followed by skills (64 percent), errors/loss of control (52 percent), length implementation times (46 percent), and jobs elimination (39 percent). Another challenge many organizations will face is simple getting enough data into the AIOps platform to enable it to learn the IT environment in the first place.
Many MSPs are naturally at the forefront of AIOps, as part of their ongoing efforts to drive costs out of their business models, by ruthlessly embracing automation. However, the OpsRamp survey also makes it clear that many IT organizations may embrace AIOps and start to move on without their MSP. MSPs should take advantage of AIOps platforms long before their customers figure out how to do it without them.
MSPs should take advantage of #AIOps platforms long before their customers figure out how to do it without them.
In the meantime, MSPs will need to find ways to add value beyond automating routine IT tasks. It’s not likely AI is going to replace the need for professional IT administrators any time soon. It is likely each IT administrator will be able to manage larger amounts of IT infrastructure at scale, thereby reducing their dependency on an MSP. While it may be a few years before AIOps becomes commonplace, savvy MSPs should take note of the fact that the level of enthusiasm for AIOps is starting to accelerate with or without them.
Photo: Lightspring / Shutterstock