AIOps is going to be the biggest change to how you manage your network for at least a decade, if not longer. The potential sounds incredible – it could give you almost infinite capacity to proactively optimise your network in real time, increasing agility, decreasing downtime, and mitigating risks.

Of course, bridging from where you are today to this ideal state is far from simple.

However, what is critical is that you start to develop that vision state, and work backwards to the actions you can take today to deliver immediate benefit, while start building towards that vision.

To help you start your journey, in this article we will begin by defining what AIOps is, why and how it could benefit you and your organisation, and the steps you can take to start delivering those benefits.

 

 

 

What is AIOps?

Let’s start at the beginning with a definition of AIOps: “The application of AI to automate, streamline, and optimise the management of complex network infrastructures”.

That definition gives us a great start point because it neatly encapsulates the AIOps journey we at TNC envisage for most organisations. No organisation should be giving total management of the network to AI today – the capabilities simply aren’t there. However, most organisations can start with simple steps to streamline ways of working, leverage AI’s data analytics capabilities to start optimising the network, and build towards a more automated future.

 

 

 

What Does AIOps Matter?

Most organisations are placing ever greater demands on IT to deliver new services to their customers and to enhance employee productivity. These greater demands on IT are in turn driving greater demands on the organisation’s network infrastructure – more workloads, more end points, needing to be delivered more quickly, with ever greater levels of change, and more security challenges than ever before.

However, most organisations are also demanding this is done within ever-tighter cost envelopes, meaning resources and headcounts are permanently constrained, and this same budget constraint exists throughout the supply chain, particularly into expenditure with telecoms suppliers.

For these reasons, organisations are being challenged to do an awful lot more within the same budget, and often lower budgets. AIOps is increasingly being seen as a critical way to square these circles.

In other words, the potential of AIOps to alleviate resource and headcount constraints, increase the productivity of the existing headcount, and enhance the benefits of existing investments, is increasingly being seen as critical to deliver the networks organisations require in the future.

 

 

 

Key Use Cases for AIOps

In the short term, the main use cases for AIOps are to leverage the capabilities of AI to crunch data to deliver greater insight into network performance and recommend ways in which the network could be optimised, particularly correlating complicated datasets to identify underlying issues, recommend proactive changes to avoid downtime etc. Crucially, in this initial phase of AIOps, the focus is largely on delivering insight, analysis, and recommendations to skilled humans who will ultimately make decisions about what actions should be taken.

TNC defines this as “Level 1 Capability” – humans in total control and ultimately executing changes, with AI providing recommendations.

TNC then identifies the “Level 2 Capability” as the AI still providing recommendations, but also asking for permission to make those recommended changes. Again, the humans remain in total control, but the AI having the potential to go beyond just providing recommendations to having the capability to implement at least some of those recommendations, once instructed to do so by a human.

The final step is “Level 3 Capability” in which the AI develops the recommendations, but has the autonomy within rules designed by humans to implement those recommendations, with the humans receiving reports about what the AI has done, but without the AI needing permission to act.

TNC’s analysis suggests that many organisations could reach Level 1 Capability very quickly with existing toolsets that may even be deployed today, or could be within 1-2 years. Furthermore, TNC would suggest that limited adoption of Level 2 Capability could take place within similar timescales.

However, TNC would say that there remain a number of challenges to reach Level 3 Capability. Partly these challenges are around building trust in the competence of the AI. However, as we will see in the next section, an equally significant challenge is building the ecosystem between the various actors within the network environment to enable AI to enact changes.

 

 

 

Who Delivers AIOps?

One of the biggest challenges for the adoption of AIOps, and the one of the biggest questions for the future is who will deliver it, and how?

As we all know, there are many actors within most organisations’ network environment. Just within the network domain, there can be the internal network team, the internal security team, telecoms suppliers, security partners, and various systems integrators. Looking more widely, we can often also see the hyperscalers, cloud providers etc.

In theory, all of these players could have systems and tools that need to work together to deliver AIOps – whether it’s the toolsets from the network vendors (e.g. Cisco, Palo Alto etc.), operations toolsets (e.g. ServiceNow), enterprise-wide agentic toolsets (e.g. Copilot), toolsets within your telecoms suppliers (e.g. BT, Verizon etc.).

For Level 1 Capability, it could be sufficient for the AI to have access to internal data to look at network performance. For basic Level 2 Capability, it could be sufficient for the AI to have access to CPE to make config changes based on its analysis of internal data. However, for true Level 3 Capability, the AI is going to have to be able to access and correlate data across multiple datasets, and then enact changes across multiple domains.

 

 

 

Key Challenges to AIOps Adoption

TNC considers that there are 4 major challenges for organisations in adopting AIOps:

• Tech
  First and foremost, we are in the early stages of the development of the technology necessary to deliver AIOps – whether it is the AI itself, the application of the IT into capable toolsets etc. Of course, as with all areas of AI, the pace of development is rapid

 

• Data
  Most organisations have found any attempts to automate processes have been impacted by either a lack of underlying data, or the fact that the data is siloed in different places, captured in different formats etc. AIOps is no different – one of the key requirements particularly to progress from Level 1 to Level 3 Capability will be the ability to correlate across datasets, and progress will need to be made to achieve this

 

• Trust and accountability
  As with all aspects of AI, building trust in the capabilities of the tools, developing processes to validate performance, and correctly allocating and tracking accountability across the organisation are key to de-risking adoption of AIOps, and particularly to eventually taking the humans “out of the loop”, which is where AIOps can deliver most value

 

• Interoperation
  As noted above, particularly to gain the benefits of Level 3 Capability, organisations are going to have to find ways for their AIOps tools to operate across multiple sub-systems provided by multiple actors within their network environment. As might be imagined, all actors are hoping *their* tools will become the industry leader, so it isn’t a given that all systems will embrace interoperability

 

 

What Should You Do Now?

The most important step for all organisations is to take in the coming months is to develop a defined roadmap for AIOps. Whilst this sounds simple, in practice you will need to consider multiple inputs to develop your individual plan, including:

• Your wider organisational AI strategy
  What is the overall organisational appetite for AI, what platforms are you using, what governance processes do you need to consider when developing new AI use cases etc.

 

• Your network lifecycle
  Where are you in the lifecycle of the key elements of your network infrastructure e.g. your telecoms supplier contracts, your hardware lifecycle, your tooling deployments etc. Put simply, if you’ve just signed a 5 year contract for your network, your timetable to adopt AIOps is likely to be different from an organisation that is approaching the end of its contracts and/or hardware lifecycle and has more capacity for immediate change

 

• Your application strategy
  The extent to which your applications and workloads are migrating to the cloud will be a big determinant of the type of network you have deployed, which in turn will be a big determinant of the benefits and consequent drive for AIOps adoption

 

• Wider drivers
  You may well want to consider wider drivers within your organisation – for example, if you are under significant pressure to deliver cost savings, your organisation’s appetite to adopt AIOps may well be increased; similarly if you are going through changes of ownership, M&A activity etc., the budgets and drivers for change may be greater, again increasingly the potential appetite for AIOps

Developing a good understanding of these inputs should enable you to start to consider key questions such as the potential benefits AIOps could deliver, particularly thinking across the Level 1, Level 2, Level 3 framework set out above., and the potential time horizons for adoption of each.

In summary therefore, AIOps is something that almost all organisations are going to embrace over the coming years, and which will almost certainly enable enhanced network provision to the organisation. The question isn’t whether you will adopt AIOps, it’s really just a question of when and how.

 

 

 

 

 

How Can TNC Help?

TNC is the UK’s largest independent network and telecoms strategy consultancy. We help over 320 of the UK’s leading organisations develop and execute market-leading strategies for network services, and helping them develop AIOps strategies is an increasingly important part of our portfolio.

We have the people, data, tools, and experience to accelerate your adoption of AIOps, whilst ensuring it delivers maximum value, and mitigating risk.

Contact us here to talk about how we can help you with your AIOps journey…