Advancing Intelligent Network Automation with Gluware’s Jeff Gray

This is Textron tv. Hey guys, thanks for the throwaway here with Jeff Graves, the CEO for Glue Air, and they just got an investment from a, uh, what they call a venture capital firm these days still. And, uh, we're gonna be talking here about, well, what is going on with intelligent network automation, because I feel like we're on the cusp of something different here in the age of ai, but we're jumping into that in a minute here.

Jeff, welcome to show. Thanks Mike. Appreciate it.

Great to, great to be here. All right, give us the backstory on the investment. Where did it come from and what are you guys thinking about using it for Sure.

2 billion serviceable addressable market. The, the, the reason for the investment is because of our platform and our team and going out to attack and be able to serve customers in what is now becoming a mature market. We have been paddling a well ahead of the wave for years, and now there are actually conferences, network Automation, forum Auto Con 0 1, 2, and three.

Uh, the next one will be in, in product. So it's actually a conference directly around network automation. And the thesis is that customers are currently self building scripts and playbooks and cobbling together their own software.

And the, this has happened before in the market. So if you think of data storage, backup, and recovery, uh, this happened before. If you think of compute, this has happened.

If you think of traditional RPA, this has happened. So instead of global 2000 customers needing to write their own scripts, if there is a platform that is commercialized, that is secure, that is safe and predictable, that actually allows customers to take the scripts and playbooks that they've been writing for years, and to be able to leverage what works and reuse their self-build, uh, scripts, but also institutionalize that knowledge, uh, bring self-operating and safe and predictable change to enterprises that this can reduce cost, increase security, and really help customers and enterprises become much more agile. So there's a meeting of the minds between Quadri Capital and Glu Air, and I would say one of the, the, the key things is that their operating advisor, Saed Zui, who was a founder of Ansible and the former chairman and CEO, is now on our board.

And if there's anyone that knows about DIY self building scripts and playbooks, it's sayed. And the meaning of the mind is that, hey, this is the future that customers need a commercialized platform to really get to the next level and bring that value. This is what what is ahead for, for us.

I can go into roadmap and, and whatnot momentarily, but let me, let me pause there because that's the thesis for the investment. So where are we on the journey here? Because I feel like a lot of the times, the people with the expertise to automate the runbook, they're too busy holding things together to create the playbook.

So, um, have we gotten to some point where maybe it's becoming easier to create those playbooks? And are we on the cusp of maybe democratizing network automation? I, I would say that, that through a platform and we have what is called dial, uh, device interaction automation layer, that brings something new to the table.

Uh, I would argue that if, if customers continue to script and write playbooks that it's actually script and playbook sprawl, and they actually can't just magically automagically figure it out, everything starts working. That new humans need to come in to relearn, uh, scripts and playbooks when, when those developers or or scripters leave a company that it's very task oriented, that, uh, there's only a handful of companies that have been able to do this well at scale. Almost every customer, every enterprise has an automation team, and it's become mainstream in terms of looking to tackle this problem.

Yet, one of the things, and the key takeaway at Network Automation Forum, which is in in Denver and November, is that throughout the industry, customers are really scratching the surface. They're looking for help. In fact, one of the taglines of Network Automation Forum is why haven't, and I, I, I may not get it perfect, but you know, why hasn't this taken off?

And so you have a whole bunch of like-minded individuals that have gotten together to try to start to share ideas and recipes and solutions. And meanwhile, we have a platform that has over a million, uh, hours of, of death, uh, 301 releases strong. We're in major customers that are public, uh, in terms of references, MasterCard, ey, Merck, uh, a variety of customers that that have, have been able to put their brands associated with Glue Air, and they're running glue air throughout their entire enterprise.

So with through a platform approach where you bring best of breed called scripting and ideas and task oriented development, and if, if platforms such as Glue Air can go and call those scripts for some of the more simple things, but then for more of the complex, uh, things where it comes to like, say an EVP and VXLAN fabric with 30,000 switch ports and being able to iterate on a model and constantly change it and tune it for the enterprise need, you have to have safe and predictable automation. It has to be intelligent with pre-checks and post checks, declarative item potent, et cetera. That's what gluer brings to the table.

So instead of like self-build on one side and then a platform on the other side, if you could bring these worlds together, if we can go ahead and provide an on-ramp for the developers to accelerate, uh, their, their path in the enterprise, then you can bring both of these solutions together. And so what we're doing moving forward is really opening up our platform catering to the developers, and it's a both and not an either or. We're better together with Ansible, we're better together with Python, we're better, better together with Linux operator, GitHub, NetBox, you name it.

And so what you're gonna see from us moving forward is embracing the developer while, uh, making sure that executives at, at the businesses get the value. And there's a dedicated business case for every single customer. So I would say through platform as well as some scripting and these coming together that there really is a tipping point that's happening.

You take one or just the other, and you know, it's not, it's not going to, not gonna work out throughout the enterprise. You've got some, some competing interests. Um, but this is a really, really exciting path for the future is, is my belief and our belief and really the thesis for the investment.

Mike, Of course, these days you can't walk down the street without somebody leaping out to tell you about their great new AI thing. But how will network automation and large language models and gen AI and all these things come together from your perspective? Well, I will, uh, let me speak, uh, directly from our perspective.

We announced Gluer ai and we have two co-pilots, uh, not one, uh, our CMO likes to Jo joke. Uh, what's better than one network automation co-pilot, two network automation co-pilots. So we have one that's focused on NetOps, which is being able to interact with the network to understand what is going on with the network, what are the vulnerabilities all through a, a prompt, uh, where we fully integrated LMS and fine tuned and trained models.

And what this means is that you can now say not only what is going on in network, but go ahead and make changes. And so if you were gonna do that with scripts and playbooks alone, without building all the pre-checks and post checks, it could be very, very scary. With Glue Air, you can interact with the network devices and elements and fabrics in a safe and predictable manner.

So now you can type commands and say, go ahead and upgrade all of the routers at this certain time and integrate with ServiceNow where we are vulnerable and, and Gluer and Gluer as copilot, NetOps copilot can go and do that. There's a a ton of value there. We're working with customers in early field trials, which we announced getting feedback, and this will be very exciting.

The second copilot is our net DevOps copilot. And this is targeted for those sophisticated users that know how to script, they know how to write playbooks. And now instead of our IDE, our integrated development environment, that's 10 x faster and more efficient, uh, in cranking out intelligent automations where our sophisticated customers are writing applications, apps, and abstractions on, on top of our platform.

Now with an integrated LLM, it's like a hundred times faster, and I'm not exaggerating, it's like 10 seconds, boom. You've got an automation for Cisco, you've got an automation for Arista, you can do vendor swap, you can do some really, really amazing things. So sky's the limit of where this is gonna go.

So with two copilots, one for NetOps and one for net DevOps and being able to share, have customers share in a community, which will be, you know, we're going that direction. Uh, now you've got the good guys and gals that can go out and be able to share best practices, share secure, uh, secure, uh, architectures and policies and have like-minded customers really starting to, what, what it means is they don't have to reinvent the wheel, they can put it into production, it's been tested and validated and glue will continue to stamp the architectures that we've gone through and tested. So these are real world, uh, AI value where it's not like, Hey, here's just a, an agent that doesn't, can't do anything.

The the fact that we're taking an ag agentic approach, but we have the ability to push and, and be able to, to make safe and predictable change. There's very, very exciting value ahead, Mike. My suspicion is there might be more agents in their future as well, but how do all the agents talk to each other so that they can hand off tasks back and forth?

Because the networking agent will need to talk to somebody else's server agent and somebody else's, uh, application agent at some point, right? So, so this has really, in my my view, this has not been, uh, accomplished yet in the industry. This all, it's all talk at this point.

Um, having said that, you're completely right. This is where the future is going. So Federation of Agents, now agents can, uh, gain information from each other and be able to interact and you know, it's yet to be determined exactly where the human will be in the loop.

But from our customers, they do want a human in the loop, uh, to be able to make important decisions where if there are, you know, some more simple types of tasks, things that are, let's say, gold standard that needs to be remediated, the customers will allow that to just run on itself. I would say the sky's the limit of where this can go. And we're gonna take a very real world approach because the last thing I, I want to do, because I've never done this, I've never got ahead of our ourselves and our skis.

I wanna make sure that our customers bake it. I wanna make sure that it's in their networks at global scale with 30,000 to 150,000 plus devices and they're seeing value from it. But lemme give you a couple of examples.

Uh, one example is an observability agent. So most of the observability platforms out there, well actually all of them, they look at problems. They don't actually fix problems.

We're not interested in being an observability company. That's not what we were designed to be. There are great companies that are out there, you know, ala Ken Tech, uh, you know, you look at some of the things selectors doing that, you know, some of the folks that that spoke in network automation form, and they go and see issues.

So if we have their agent be able to interact with our agent and say, okay, here's an issue and we believe it's, it's down to these devices or these parameters in the network, they can actually hand that off and create an integration where Gluer can go off and make those changes, then gluer to ServiceNow and forms Ken Tech that the problem is fixed and, and life is good. In fact, two years ago at NU Open Network user group, we showed Ken Tech and Luer interacting with a live Zoom call. Where in that POC there was a Zoom call, someone made a change in the network, the Zoom quality went bad.

Uh, Ken Tech notified Glu Air Glu were often fixed it. Now, these were not two agents that were talking, these was, this was a dedicated integration. It's gonna get much more interesting when agents can interact and talk.

We're not there yet. I don't believe anyone's there yet, but that is exactly where the future is, is going. So that's my view.

To what degree will I need to be a specialist? I mean, I think we all agree that some human will be in the loop, but does that human need to be a networking specialist going forward or can they be more of a traditional IT administrator who is maybe networking savvy to a degree, but they're trying to manage all these things in concert with one another? We have, we have roles for both.

And that is the, the part of the key with the glue wear platform. In fact, one of our, one of our customers calls it the, uh, the brain drain prevention, uh, machine that, that a lot, one of the most sophisticated network engineers come and go when they go. A lot of the, the, the, the, the historical knowledge goes with them once the designs and architectures are baked into glue air and there, there is part of our platform that allows the network architects and the senior network engineers to be able to onboard their designs.

Once that's done, then that can be handed over to some of the less sophisticated folks that are more network operators and expose a handful of knobs to turn things that won't like go out and break the network at scale. But, but some of the best practices to go through and be able to remediate, uh, things that are that, that may be broken. You know, one of our, one of our advisors, Kevin Carney, uh, he was the chief network architect at, at MasterCard and he, he selected us when he was at, at MasterCard, has been on audio a whole handful of times.

So like what is, what is one of the biggest issues that happens when you're troubleshooting a network? It's like, well, uh, when you're removing security restrictions, a lot of times either network engineers forget to reapply them or they don't stick because they're in there manually troubleshooting. And so, you know, part of the value for customers is that if there is something that's detected that, that the network ops team has the core best practices that are safe and predictable, where they can go through and push a button or agree that this can go back to gold standard, even with a human in loop or without, and things will get, get, get remediated and, and put those security, uh, the, the the security policies back on, on on the devices.

And so it doesn't require everyone to be an absolute complete rock star and black belt. Um, that there has to be both. And the other thing that I would say is that there's a lot of challenges, and we've seen this throughout our customer base, you can't be the best network engineer in the world and the best developer and scripter in the world.

Now there are a handful of unicorns, but that's not, you know, everybody. And so a way, you know, as you said, democratize, uh, automation, I believe, I think you said something democratize scripting or playbooks or, or whatnot. I mean, that really is the approach that we're taking, but it's more about enabling those to what, what their strengths are and allowing them a path to be able to move to the other side more of an elegant manner.

Because if you have to build everything from scratch, you have to have really, really great network engineering. That's one person. Then you have to have really, really great, uh, development.

That's another person. Now you're taking two people out of the workforce to be able to try to self build something. So if you can give 'em a platform that allows them to go faster with the skill sets they have and meet in the middle, that's why customers are buying blue air.

Where do you think security will fall in this equation eventually? 'cause it does seem we're seeing a lot more convergence of networking and security, especially around security operations. Um, are, are these things going to meld somehow and what does that look like?

They already are, and that's a great observation. I believe this is happening very, very fast, Mike. So we've seen a lot of teams that, that have actually come together where you have both the network and security that have been collapsed.

And so security policy is, is is a really big deal for ware moving forward. Um, it's difficult to have a security policy platform for traditional firewall rules and then a network platform. And customers are asking us why are the two platforms?

So you'll be seeing some pretty large announcements from, from ware, uh, and what we've already done with security policy throughout, throughout networks and these roles, uh, are, I'm seeing converge because you can't just think of networking separate and have your hands in the air about security. That security has to be a key ingredient. It has to be a way of life and make sure that everything that network engineers do when their designs are secure.

And so if you can have multiple roles in an automation platform that cover both traditional networking as well as next gen networking as well as iot, NO ot, and by the way, massive security vulnerabilities in in iot NO OT throughout, you know, those networks that are not regularly patched, that are not regularly, uh, segmented with macro and microsegmentation. And it's a real difficult pain point for customers to be able to, to to solve. And so if you can have a platform, if customers can have a platform that not only handles multi-vendor, multi-domain and traditional networking as well as next gen networking, also, uh, security policy, firewall rules, et cetera, and then moves into iot, ot, this is a platform that customers are highly, highly interested in.

Outside of, you know, acquiring your platform, what are you seeing customers who were well prepared for this next wave of automation doing? Or is there a mindset here? Is there a cultural change change?

Are there, uh, is it just technical training? What's kind of the secret sauce? It's a work in progress and I would say that that it, I would say the big, the big change is doing things differently.

And it has to be driven from the top, meaning top of customers has to be exec driven. The customers that have been able to get the furthest along and have the most value, both financially from security and outage prevention. This is driven by the C-I-O-C-T-O VP of it, depending on the size of the organization, where they've bonused their teams on the transformation.

And if they do it the old way, meaning the manual way or go off and, and write a script, if it's already baked in a model with glue air, if it's not baked with glue air, if it's, if it's something that the script's been working and that's part of an RPA workflow that is tested and approved, great, but if they go off the rails and they go and do something manual that there are repercussions that like no, they can't just go off and make manual change. And so, you know, the adoption of everyone rowing the same direction, not they say something in, in, in a meeting and then they go off and, and still do it, the, the, the, the same way that doesn't work. And so this is more human behavior changing at operations more than anything else.

And the best part about it is that these network engineers and these developers, they get their weekends back. Like they're not getting the calls at four in the morning. We've had, uh, over a 99% reduction in network outages throughout our customer base, 300, uh, times faster OS upgrades.

We're seeing a 50 to one time and cost savings with customers. And so if they adopt it, if they get with the times and they look to do, they're gonna free themselves up for more of the strategic things. And instead of, you know, wrapping themselves into just like trying to keep their hands on the keyboards, that actually they'll do be doing things for the business that moves the needle more.

And, and when the IT execs recognize that behavior and they reward that behavior, then we see transformation. Now what I would say is that it's really, really hard to get that transformation if everyone is taking just a pure scripting and self building mentality. If you couple 'em together, meaning with best of breed platform plus some customization and scripting on, on the side, I mean, it's pretty amazing what what can happen.

And we've got some just amazing, amazing customers that have, have been able to move mountains and get a tr tremendous amount of value. Which by the way is why a private equity firm has now moved forward with gluer because I believe we are the first company that has gone the private equity route, which means we get to continue on the mission, we get to continue to grow, and we really get to continue to be a leader in the space instead of ending up in, you know, business of, of a strategic. So, All right, folks sharing it here.

There's an old saying, like, if you keep doing the same thing and expecting a different result, you just might be crazy. So it's time to think about it differently, right? Hey Jeff, thanks for being on the show.

Thanks, Mike, I appreciate it. All right. And back to you guys in the studio.