Oct 23, 2026 · 1 hr · 12 segments
Many network engineers don’t trust AI, in part because the tech is non-deterministic and can behave in unanticipated ways. Simultaneously, the networking industry is baking AI functions deeply into…
Joby RudolphGuestEthan BanksHost
Drew Conry-MurrayHostThat helps, and I wanted us to set that up because it really gets into applying this technology to network operations and how we get to this point of trustworthiness.
How do we trust agents knowing that historically LLMs have hallucinated or become creative in ways that, from an engineering use case standpoint, we don't want it to be creative.
Why do we even want to go this route? What do AI agents bring to network operations that we didn't have before?
So, I think network operators and all the NOC, the network planners, everyone had a lot of expertise, a lot of tools at their disposal, and they were being able to use this to make decisions using all their experience, right? I'm smiling as you're giving this answer because
I got 15 tools, all of which I need to query depending on what it is I'm trying to solve.
You take a decision by looking at 15 different tools and coming up with a decision.
Maybe for planning purposes or when you're troubleshooting a major network outage, where do you even start? Because even though you have only 15 tools, you might see the issues coming in 15,000 directions.
Because when an outage happens, you can see complaints from everywhere in the network.
But then how do you tie all of that together? That's where you need something that thinks like you, but is able to do it at a machine speed.
So what we are really telling now is, I mean, the role or what the AI agents are currently trying to help or trying to complement users is to help with this thought process at scale.
Now you are able to, as the networks are evolving, expanding, the core problems still remain the same.
And this is where the AI part comes in, where it is able to understand the data and present you in a way that, okay, now you can go take action on it.
The meantime to resolution is what really is probably what AI is really helping as networks become more evolved and complicated and scale up.
So I feel that's the biggest value that AI is bringing to network operations.
Is it that the AI agents can query the 15 tools and consolidate that data in an interesting way? Or is it that I don't need 15 tools anymore because agents can go to the source or go to a data lake of some kind and find answers for me?
That helps, and I wanted us to set that up because it really gets into applying this technology to network operations and how we get to this point of trustworthiness.
How do we trust agents knowing that historically LLMs have hallucinated or become creative in ways that, from an engineering use case standpoint, we don't want it to be creative.
Why do we even want to go this route? What do AI agents bring to network operations that we didn't have before?
So, I think network operators and all the NOC, the network planners, everyone had a lot of expertise, a lot of tools at their disposal, and they were being able to use this to make decisions using all their experience, right? I'm smiling as you're giving this answer because
I got 15 tools, all of which I need to query depending on what it is I'm trying to solve.
You take a decision by looking at 15 different tools and coming up with a decision.
Maybe for planning purposes or when you're troubleshooting a major network outage, where do you even start? Because even though you have only 15 tools, you might see the issues coming in 15,000 directions.
Because when an outage happens, you can see complaints from everywhere in the network.
But then how do you tie all of that together? That's where you need something that thinks like you, but is able to do it at a machine speed.
So what we are really telling now is, I mean, the role or what the AI agents are currently trying to help or trying to complement users is to help with this thought process at scale.
Now you are able to, as the networks are evolving, expanding, the core problems still remain the same.
And this is where the AI part comes in, where it is able to understand the data and present you in a way that, okay, now you can go take action on it.
The meantime to resolution is what really is probably what AI is really helping as networks become more evolved and complicated and scale up.
So I feel that's the biggest value that AI is bringing to network operations.
Is it that the AI agents can query the 15 tools and consolidate that data in an interesting way? Or is it that I don't need 15 tools anymore because agents can go to the source or go to a data lake of some kind and find answers for me?
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
Search every transcript — by keyword, by phrase, or by meaning, across every show Radar indexes
Trends — what is surging across podcasts, measured against its own baseline
Alerts — when a name you follow appears in a newly indexed episode
No account is needed to search Radar.
