Survey says AIOps gaining ground with CSPs
Despite AIOps early phase, nearly half (46%) of our survey respondents report that they have at least an evolving AI business case, while another 40% have none at this stage.
21 Dec 2020
Survey says AIOps gaining ground with CSPs
TM Forum surveyed a panel of operations experts from more than 20 CSPs, consultancies and supplier organizations around the world to gain a picture of AIOps relative to adoption, maturity, direction and real-world implementation.
74% of the respondents to our survey work in operations for CSPs in hands-on managerial roles in IT and operations. 64% of respondents work for CSPs with more than 5 million subscribers, while 41% work for CSPs with more than 25 million subscribers.
Despite AIOps early phase, nearly half (46%) of our survey respondents report that they have at least an evolving AI business case, while another 40% have none at this stage.
Compared with data collected in TM Forum’s 2018 AI survey, CSPs are slightly more focused today – 53% – on using AI to automate a single process within a single domain than two years ago –44%. This may reflect the practical challenges faced when automating processes across organizational silos and boundaries.
Yet AIOps is firmly in the early part of the adoption curve among CSP operations teams. 54% of our survey respondents are using AI today to automate a single operations process within a single domain. More than 33% of respondents are using AI to automate one or more processes across multiple domains, which aligns with our findings from 2018 – 36%.
There is not yet a widely adopted management approach to AI. 0% of our survey respondents report a C-level owner and only 14% report having a VP-level owner for AI. 33% report multiple owners, because AI is cross-functional. Another 32% say it is unclear who owns what, especially with AI.
Similarly, IT spending related to AIOps may yet be modest. Figure 12 shows 47% of our survey respondents say AIOps will occupy less than 10% of their IT spend in operations during the next 3 years. For 35% of respondents, however, AIOps will command one fourth of their budget. A small portion of respondents – 12% - say because AI will be part of everything, all spend will go to AIOps.
More than half of our survey respondents say they have no AI activity in network planning and management today, though nearly 25% are building AI into products and services in this category. Almost half have no AI activity in service creation and management, though roughly 35% have completed proofs of concept in this area.
By 2022, our survey respondents expect to see more forward momentum in AI deployments. Nearly 35% of respondents see AI being used in network planning and management. Service creation and management may see only small gains in adoption, however, with less than 20% of respondents saying it will be implemented by 2022.
CSPs still face maturity hurdles with AI as a consumable enterprise software product. Not only are AI tools immature, but the environment around them may not have transformed enough yet to feed them. For example, network automation is a killer app for AI. But 69% of our survey respondents say the biggest barrier to network automation is a lack of standards for exposing network operations data.
While there’s great excitement around what AI can do, there will be a learning period for CSPs to be able to unleash its full potential. AIOps is gaining ground because it provides a basis for finding ways to improve operations using both automation and AI. 60% of our survey respondents say automation is a primary objective for their network and service operations.
For a closer look at how CSPs are using AIOps to automate processes and AI to being the move toward autonomous networks, download the latest report from TM Forum: AIOps: From automation to autonomous networks.
74% of the respondents to our survey work in operations for CSPs in hands-on managerial roles in IT and operations. 64% of respondents work for CSPs with more than 5 million subscribers, while 41% work for CSPs with more than 25 million subscribers.
Despite AIOps early phase, nearly half (46%) of our survey respondents report that they have at least an evolving AI business case, while another 40% have none at this stage.
AIOps remains a very new term for most CSPs, as only 30% of respondents report hearing the term used in their companies at least weekly, while more than half (54%) say they never hear it.
Compared with data collected in TM Forum’s 2018 AI survey, CSPs are slightly more focused today – 53% – on using AI to automate a single process within a single domain than two years ago –44%. This may reflect the practical challenges faced when automating processes across organizational silos and boundaries.
Yet AIOps is firmly in the early part of the adoption curve among CSP operations teams. 54% of our survey respondents are using AI today to automate a single operations process within a single domain. More than 33% of respondents are using AI to automate one or more processes across multiple domains, which aligns with our findings from 2018 – 36%.
AI market still taking shape
There is not yet a widely adopted management approach to AI. 0% of our survey respondents report a C-level owner and only 14% report having a VP-level owner for AI. 33% report multiple owners, because AI is cross-functional. Another 32% say it is unclear who owns what, especially with AI.
Similarly, IT spending related to AIOps may yet be modest. Figure 12 shows 47% of our survey respondents say AIOps will occupy less than 10% of their IT spend in operations during the next 3 years. For 35% of respondents, however, AIOps will command one fourth of their budget. A small portion of respondents – 12% - say because AI will be part of everything, all spend will go to AIOps.
CSP AI activity present but modest
More than half of our survey respondents say they have no AI activity in network planning and management today, though nearly 25% are building AI into products and services in this category. Almost half have no AI activity in service creation and management, though roughly 35% have completed proofs of concept in this area.
By 2022, our survey respondents expect to see more forward momentum in AI deployments. Nearly 35% of respondents see AI being used in network planning and management. Service creation and management may see only small gains in adoption, however, with less than 20% of respondents saying it will be implemented by 2022.
AI tools still maturing
CSPs still face maturity hurdles with AI as a consumable enterprise software product. Not only are AI tools immature, but the environment around them may not have transformed enough yet to feed them. For example, network automation is a killer app for AI. But 69% of our survey respondents say the biggest barrier to network automation is a lack of standards for exposing network operations data.
While there’s great excitement around what AI can do, there will be a learning period for CSPs to be able to unleash its full potential. AIOps is gaining ground because it provides a basis for finding ways to improve operations using both automation and AI. 60% of our survey respondents say automation is a primary objective for their network and service operations.
For a closer look at how CSPs are using AIOps to automate processes and AI to being the move toward autonomous networks, download the latest report from TM Forum: AIOps: From automation to autonomous networks.