*Blog by Brandon Larson, SVP and GM, Cloud, AI and IMS Business Strategy, Mavenir
For years, the telecom industry has focused on technology milestones to measure progress. What “G” is deployed, what level of 3GPP release compliance, cloud adoption, automation has guided how operators assess their evolution.
Today, that focus is shifting. The biggest transition taking place is a move away from measuring technology implementation towards measuring business outcomes. In the context of autonomous networking, this shift is critical.
There is still a strong tendency to focus on levels. Operators ask whether they are at level zero, one, two, three, or four. But the more important question is often left unanswered. If an operator reaches a certain level, what does that deliver to the business?
From Technical Progress to Business Outcome
Autonomy on its own does not create value. It is the means to an end, which is a business outcome. To achieve any business outcome, it must be defined at the beginning before any technology decision is made.
This means defining the business problem to be solved and defining what success looks like in terms that can be measured.
“Autonomous” networking is a broad solution to a multitude of daily operational challenges. The key to solving them is naming them specifically, precisely and defining the solution by a measurable outcome.
The Intersection With AI
Similarly, AI on its own does not create value. AI models and tools offer powerful capability but it must be shaped to deliver the desired outcome. There are three guiding principles for shaping AI outcomes.
The first is determining on whose behalf is the AI acting. This helps define the role, scope and boundaries of an AI solution. Is it taking on the role of L1 support, filling out a CIQ or upgrading software?
The second is defining the problem the AI is tackling in precise terms. Is it detecting faults, routing tickets, producing designs, making decisions on upgrade actions?
The third is defining what success looks like in measurable terms. What is the false positive/negative and first-time ticket accuracy thresholds. What are the acceptable upgrade success rates and execution time?
By moving to an outcomes approach, the business value becomes more tangible. Example metrics would be: Operational rollout time reduced from 18 months to 8 weeks, first time ticket routing improved 40%, fault isolation reduced to under 5 minutes, average time to restore now under 5 minutes.
From Partnership to Real-World Execution
This shift is not theoretical. It is already being explored and implemented in real environments.
Mavenir is beginning to deploy its Agentic Service Assurance solution in operator environments. It is acting on behalf of L1 and L2 support engineers. It is tasked with filtering alarm noise, enriching tickets and running diagnostic workflows based on skills given to it by Mavenir experts.
It is currently assisting Mavenir L1 and L2 personnel making them more efficient in their daily operations. The Mavenir engineers are also assessing the AI outcomes and making adjustments to make it more effective. The key to shaping AI is embedding our expert operational wisdom into the AI. This can only be done with the domain knowledge we possess.
One the AI agents are performing at the level of our current human experts…or better, it can be used by our customers, giving them access to Mavenir expertise through our AI agents.
