Episode Transcript

Chrystal Taylor:

Welcome to this mini episode of SolarWinds TechPod. I’m your host, Chrystal Taylor, and with me as always is my co-host Sean Sebring, and we are back with Matai Wilson to talk about some things that we couldn’t get into in our last episode, but now that there’s been some SolarWinds Day happenings, there’s a little bit more we can talk about. So Matai, what would you like to share with us about how SolarWinds is thinking about AI and Agentics specifically with the announcements that happened in October?

Matai Wilson:

So yeah, thanks for having me back. I’m glad to be here. And I think where we left it off last time was kind of around this digital intuition concept and proactive remediation, and I’m excited to say that I couldn’t say this last time, but we are actively working on delivering that sort of thing. So we have features such as capacity planning for our K8s and some of our databases, which are going to allow us to proactively predict when capacity will be met and when we may overload actual thresholds, and that’ll allow us to then proactively mitigate and remediate those situations before they occur. So really diving into that proactive remediation, I do strongly believe that within five to 10 years, the idea of talking about MTTD and MTTR are probably going to go the way of the dodo bird, and we’re going to be talking more about how many incidents you have in a year versus how quickly you’re resolving them.

And then there are a whole bunch of other features that we’re launching. We’ll be launching RCA, which traditionally is called root cause analysis. We’re actually coining that to be root cause assist, and so that is based on generative AI that actually takes into account alerts that are coming in from our node suppression system, et cetera, across different applications. And then it correlates those alerts to figure out what may have actually caused the issue and how to resolve it, and then offers you in plain English a summary event and what you might want to do with that. Additionally, we talked about some agentic concepts last time. We are going to be launching, I think later this year, our general contractor is how I describe it, of an LLM. It doesn’t have a name as of today, but we’ll see what it ends up being called.

But yes, this general contractor, this governing orchestrator will be the first semblance of our agentic framework. That agentic framework, again, not necessarily all being AI agents, but some being tools that are orchestrated via an MCP server to allow our general contractor to actually go and deliver on actions and other things like research again, like finding out the root cause, et cetera. So things like root cause assist, which would be a tool or a contractor inside of this general contractor’s tool belt will be available to it via MCP server. So when you ask our chat-bot or our LLM, “Hey, what’s going on? I saw this alert pop up,” it will go intuitively know that you’re talking about that specific alert and it will pull the RCA tool and do a root cause analysis and provide you a summary and then recommend it or guide it and express actions.

We will also have another agent called SRE agent, which is meant to be kind of like the sidekick of an SRE that does a lot of that RCA stuff, but I would describe it as RCA on steroids. So it really dives deep to go figure out event causes and correlate those and figure out exactly what to do and exactly how to mediate that in the event of a critical situation. And all of this, again, will be available directly integrated within this general contractor LLM along with a suite of other tools. The greatest thing about this is that because of the way we’re architecting it will only just add more tools to its toolkit over time and become even more capable kind of guiding our way towards that autonomous IT operations.

Sean Sebring:

So name idea, we can just call it Mat, we’ll name it after you because you’re Mat AI.

Chrystal Taylor:

That’s good. Mat AI. That’s good. You should use that somewhere.

Matai Wilson:

I’ve never heard that before. I’ve heard that many times as far back as 2018, I think somebody-

Chrystal Taylor:

Really?

Matai Wilson:

Yeah, I won’t say my personal email address, but it involves my name and the person that I was speaking to thought like, “Oh, you’re really into AI because your email address is Mat AI.”

Chrystal Taylor:

Your name is Mat AI.

Matai Wilson:

I was like, “No, that’s just literally how you spell my name.”

Chrystal Taylor:

That’s just my name. It’s a good handle though. You work in AI, I mean… I think that all of this is really exciting. I want to say, you can correct me if I’m wrong, but I think Root Cause Assist has been in tech preview for a little bit at this point, even at the time of recording. And that’s all really exciting stuff and being able to, everything that you’ve talked about is especially exciting for me as someone who has been the manager of the monitoring system and not necessarily the person that works on the database or the network piece or whatever. I’ve been handling monitoring observability and less so being in the weeds on the actual support of the things. So being able, you’re talking about going right into readable terms, human understandable terms is very exciting. It’s like what normalization of logs did for everyone where you could actually understand what the logs are saying. That’s all super exciting. Thank you so much for sharing with us.

Matai Wilson:

There’s actually cool stuff we’re doing logs by the way, such as Log Insights and Log Assist, which basically use it as our log compression algorithms to then go figure out which logs to look at, and then based on the logs, figure out exactly what’s going on. And then in very, very plain natural language offer you a pointed summary of like, “Here’s exactly what happened at this point in time and here’s what you can do about it.” And then again, all of this is correlating, all of this is being brought together, so there’s really, really powerful tools that are going to be coming to fruition pretty soon.

Chrystal Taylor:

So much time saved.

Matai Wilson:

Yeah, I don’t know if you’ve ever had to look through logs.

Chrystal Taylor:

Yes, I have.

Matai Wilson:

And it’s absolutely miserable.

Sean Sebring:

I was going to say fun, but yeah, no, referring back to our last episode that we did, this is exactly where I was saying I’m excited to see what we’re going to do with AI now that it’s here. I was referencing reporting, which is again, kind of similar to going through logs. It’s collating data to tell a story, and now it’s not just going to be a graph, it will actually tell you a story of what’s going on.

Matai Wilson:

And this is kind of core to SolarWinds original operating philosophy, which was two things: the single pane of glass for everything in your observability; and then also very, very easy to learn and use because it solved one of the biggest problems that all IT operators have today, which is turnover. IT admins are constantly churning, and if you’re constantly churning, you’re having to constantly learn and onboarding takes a ton of time on these very complex IT observability systems. And so another feature that we’re going to launch is actually like a guided assistant that will help you learn and answer questions about how to do things within the actual platform so that you can drop that onboarding time and make it go even faster. So when you have a new employee onboarding to SolarWinds, super snappy. If they know how to do something, they don’t have to go ask the person across the office or something like that. They literally just ask our digital assistant and our digital assistant says, “Sure, this is exactly how you do this,” and it’s trained off of all of our help documentation and everything else.

Chrystal Taylor:

Thank you so much for sharing these insights. I think it’s really exciting. I was excited at the SolarWinds Day as I was part of it, so I was excited at SolarWinds Day, but hearing even more about it is genuinely always more exciting. I think, like I said, I’ve been around the SolarWinds products for 15 years and it is exciting every time we take steps forward, it is exciting to think about how that’s going to make it easier for all of these admins to do their jobs and reduce the time that it takes to get to, I know you said eventually we’ll not be measuring mean time to resolution, but getting to resolution faster is a step in the right direction, and we’ve already been kind of working on that. And so I think that, yeah, anything we can do to make people more proactive and save time is great.

Matai Wilson:

Yeah. I mean, lots of exciting stuff. I can tell from personal experience that if I had these things when I was having to do this job, that would’ve been amazing.

Chrystal Taylor:

Awesome. Well, thank you everyone for joining us for this mini episode, and thank you Matai for sharing what insights that you’re able to with us on what we’re doing with AI in our products.

Matai Wilson:

Wonderful. Thanks for having me back. See you.

Chrystal Taylor:

Join us next time on TechPod.