In this exclusive Fierce Network interview, Dave Duggal, Founder & CEO of Enterprise Web, reveals how Networks.ai enables seamless, intent-based automation across multi-vendor, multi-cloud environments—without the risks of black-box AI.
🔹 Faster service deployment
🔹Vendor-agnostic automation
🔹Scalable, real-time network optimization
Unlike traditional LLM-based solutions, Networks.ai leverages a graph-based telecom ontology for accuracy, low latency, and energy efficiency—a game-changer for CSPs and enterprises alike.
Watch now to discover how Networks.ai is redefining telco-grade AI.
Alejandro Pinero:
All right. Hello everyone, and welcome to another exclusive interview right here on Fierce Network. I'm joined today by Dave Duggal. He's the founder and CEO of Enterprise Web. Dave, thank you so much for joining us and welcome.
Dave Duggal:
Oh, great to be speaking with you, Alejandro.
Alejandro Pinero:
Excellent, Dave. Well, this is very exciting. We're, of course, in a way, celebrating your success in 2024, your company was a finalist in the Fierce Network Innovation Awards, and surely it was an important milestone for you guys. So what does it mean for you to be recognized in awards such as this and what does it tell you about the journey you're on?
Dave Duggal:
So the first thing I want to say, this is actually our second time. So we actually won a Fierce Network award back in 2023 for AI automation analytics category. So it's great to be back, but let me say first thanks to the judges. I know you get lots of submissions, so actually it is a big deal to be selected. So we really appreciate that. I mean, I'd say at the first level, it's just great to get that validation from respected industry experts and as founder CEO to know that you've been compared favorably. That's a value proposition in and of itself.
Of course, the primary benefit is that the news goes far and wide to Fierce Network's audience and opens up doors that we might not have, either opportunities we might not have been aware of or people we didn't know, et cetera. So that's the great aspect of it. It gets amplified by Fierce Network.
Alejandro Pinero:
Sure. And you mentioned there we do get a lot of submissions. I have firsthand experience in looking at them, and I'm always impressed by the breadth and just the quantity of innovation and wonderful solutions that are brought forward. So let's talk specifically about this year and Networks.ai that was your submission. What do you think makes it fierce in the case for this award?
Dave Duggal:
Yeah, so Networks.ai, so Enterprise Web, we've been around in Telco for a long time. So we ran the first ETSI NFV proof of concept. We won a host of TMF TM Forum Catalyst projects and recently collaborated with Microsoft, which was the source of our last award with you to be the first people in the world to demonstrate telco-grade generative AI for intent-based orchestration. So we have a history in the industry.
Networks.ai is our AI-powered no-code platform for intent-based orchestration and autonomous networking. What makes it fierce is that it allows developers to talk to the network, to design, deploy, manage services using a natural language interface. And then behind the scenes, which I think is now more understood, when we were first talking about this a couple of years ago in the industry, it was a little more foreign to people, but now people understand the idea of agents and graphs and AI and automation.
But you can imagine 3, 4, 5 years ago, talking about those things was a little harder, but now it's easier. So developers can talk to the network and behind the scenes, our platform agents are dynamically composing and configuring tasks into complex workflows and calling functions to fulfill requests, respond to events. The agents leverage a telecom ontology, which is a harmonized, standards-based, graph-connected model of telco operations. And the agents can use that model to interpret events, evaluate conditions, automate decisions, and optimize their actions based on real-time context.
So it's a lot of the kind of things that people are looking for today with AI, and we're doing it already at a telco-grade level. And I think adding one more aspect to it, which is that these agents can work end-to-end across multi-vendor, multi-domain, multi-cloud networks to dynamically configure the networks for continuous assurance and optimization. So this is not like a light agentic AI solution that the agents might be right, they might be wrong. This is telco-grade.
Alejandro Pinero:
Brilliant. And so a lot of talk of gen AI, of course. So what makes your approach to it different and what are some of the advantages you can tell us here briefly so folks can get a better understanding of its uniqueness?
Dave Duggal:
Yeah, great question. I mean, obviously gen AI is exciting, but we don't believe that LLMs or large language models or even small language models should be the center of telco architecture. And that's what it's being set up as is like, "Oh, here's this new intelligence layer." But there are a couple of problems. No matter how much training, tuning and tokens you do with an LLM, they're simply not good enough to be trusted for mission-critical operations where you want available and reliable networks.
So issues with accuracy, issues with consistency, hallucinations, security costs, latency, et cetera, right? The resource and energy costs are exorbitant, and these have real-world implications, right? We're talking about accuracy, latency, energy costs, and all of these things get compounded when you think about running anything at the edge, because you don't have necessarily the infrastructure at the edge. You're very constrained. So you certainly want accuracy and you want low latency, et cetera.
And lastly, there's another thing too, is you don't want to make an LLM the center for the architecture because then you start tightly coupling around a single version of a single model. And the problem with that is the industry's evolving too fast. These models are changing too fast. So investing heavily in a single model when we're getting announcements like the DeepSeek announcement just last week, which shook the entire markets globally, I don't think it's a great idea.
So with Networks.ai, we make that telecom ontology, that graph-connected model. We make that the center of the architecture. And there's a reason for that because it's deterministic, right? It's a model of domain knowledge and real-time operational state and telemetry. So it serves as a central source of context for large language models. Importantly, our approach also is a low-token, low-latency, energy-efficient approach to optimizing LLM accuracy.
So instead of making the LLM the center of the architecture, we're making the graph the center, and then the graph can support one to many LLMs. It also allows you to point that graph at new LLMs or updated LLMs. As opposed to investing so heavily in every discrete LLM, it allows the telcos to evolve with the industry and to safely use those LLM inferences in their automation.
And that's the goal here, right? Our view is that telcos need a practical solution to leveraging generative AI that leverages its strengths while mitigating its weaknesses. And that's what we do. We call that a domain-oriented generative AI architecture, as opposed to what most people do is they just build scaffolding around an LLM and try to do anything possible to marginally improve its accuracy and consistency, but with limited results.
So even the DeepSeek model that everybody's talking about now is, "90% are accurate," but it's not really 90% accurate in a deterministic sense, right? It's 90% relevant in a subjective context. And the problem is, as all the research studies show, is the more complex the scenarios, the more agents you use, the less reliable and trusted those LLM responses can be. So again, telcos need a balanced solution that leverages generative AI as well as classic AI, which is AI ML, RL, analytics, and good old-fashioned rules. So a balanced solution that helps them get the benefits they need while still being telco-grade.
Alejandro Pinero:
Brilliant. Well, I think Dave, there's a lot to unpack and to think about on your brief conversation here today. So thank you so much for taking the time to speak with us, and congratulations on not only this year, but your continued success on the Fierce Awards, which is a testament to the great work that you and your team are putting together.
Dave Duggal:
Fantastic. We'll try and impress you next year as well.
Alejandro Pinero:
And we look forward to it. And to you our viewers on Fierce Network, thanks for joining us. We'll be back soon with another great interview from our tech leaders. Bye-bye.