New Relic adds generative AI features to enhance access to observability platform

[This transcript was auto-generated.] Keith Shaw 0:00
Welcome to DEMO, the show where we have companies come in and provide us with demonstrations of their products or services or really cool features. Today I'm joined by Peter Pezaris. He is the chief strategy and Design Officer at New Relic. Welcome to demo. Thanks. Thanks for having me. All right. So tell me about a little bit about New Relic and what you guys do, I believe you're an observability platform. And then what you know, tell us about the new generative AI features that are the you know, the reason you're here?
Peter Pezaris 0:25
Sure, sure of you having to. So an observability company means that what we do is we help companies who use software to serve their customers keep their production systems up and running. So let's say you're a bank, you've got a lot of transactions flowing through your system every minute of the day. And if your transactions are flowing more slowly, or gosh forbid, if there's a problem with your production system, and your services are down, that can cost you real money. Yeah. So if you install New Relic, as part of your production infrastructure will monitor the services that make up your production system. And when there are problems we alert you to when problems hopefully before things go catastrophically wrong, and you can go and debug those problems and get your system back up and running quickly.
Keith Shaw 1:07
Okay, so what are the new generative AI features that you've recently introduced to this platform? Because this came before? Gen AI took off, right? Yeah,
Peter Pezaris 1:17
you were like, is actually a 14 year old company. And we've been serving this market for a long time. But with the advent of generative AI, we've enhanced our customer experience to make it easier to get insights from the data we collect from your production environment. Okay.
Keith Shaw 1:32
And so who is this product and mainly for what kind of job roles companies big, small, medium size? Well,
Peter Pezaris 1:39
that's the really exciting thing about general AI as applied to observability. Because we estimate there are about 50 million software engineers in the world, but only about one to 2 million of them who are trained on observability tools. Now, what that means is there are a lot of application developers out there that sort of push and pray, which means they write the code, they commit their code, they push it to production, and they hope it works, because they don't have the training or the tools to really see what's happening. So with our Gen AI product, we open up the aperture, we allow more and more people to get the benefit of observability by making it easier to use. And I'll show you some examples.
Keith Shaw 2:17
Let's get right, yeah, let's get right into the
Peter Pezaris 2:19
demo. Awesome. Well, let's first take a look at what New Relic is. And this is the main page will you see an overview a bird's eye view of your entire production state. So you can see here, each row in this table represents one of the things that New Relic has observed. So we've got your APM services, these are applications that are running, we support open telemetry, we've got hosts, which is sort of the infrastructure on which your software is running containers, mobile application browser application. So literally, anything you could imagine can be observed by New Relic. And that also includes things that may be unique to your business. Let's say for example, one of our biggest customers is a popular restaurant chain. And I'm sure you've been in restaurants where you can order via a kiosk, well that run some custom hardware and software to take your order. And this restaurant chain is observing all of those instances of all those order machines. And they've got, they want those to remain up and running, of course, but they also have custom metrics that are tracking, like, what's the average order size. And so you can gather all that information here and New Relic and have it under one umbrella. Okay. So now let's take a look at what you can do with Gen AI. So in the last example, like one of the ways that New Relic, really, really works is by alerting you when something's wrong. So you have this alerts and AI capability that's built right in. So you can see these are the problems in our system. So these are all the active incidents that are happening in production. Of course, we've got this simulated so the system isn't really this problematic. But you can see a description of all the issues. And once you find one that you want to try to debug, you can dig in a little deeper. Now it will say that most of the time, the way that engineers enter the service is through an alert that we push to them like an email or push to their mobile application. And so they can jump right into this particular issue. And we can see that the title is apt x is less than point nine for at least 20 minutes on a service called Aura composure. Now, I don't know about you, but I don't really remember what app x is. So this is the first way that Gen AI can help. I can say, what is apt X. And you can see here I'm typing in the lower right hand corner in this sort of sidekick experience, where we've got our Gen AI chat experience where he can describe to you exactly what you want and tell you what is showing up in the user interface. So that's the first use case, super helpful for people to learn how to use the software. But we can do things that are much more sophisticated than that, such as show me response time over the last day or twice testing my typing skills. Because now that we've learned that apdex is a measure of application performance So we want to understand like, Well, why is the application slow? Why is it not giving a good user experience? So we can ask specific questions. And what the A I will do in this case is translate my English question. And by the way, we support any languages, Spanish, German, French, whatever, your native languages, and you can ask your questions in English. And New Relic AI will convert this into a query and execute the query for you. So you can see a visualization of the response time over the last day, okay, and you can see here, response time bigger numbers are worse. Yeah, so that seems bad. So this seems bad. There's something going on. And so what we can do is try to troubleshoot further. And we could click through the UI and and find out this information. But now that we have Jenai, built right in, I can say things like, search, there may be some errors. So I want to say search for error in the logs, okay. And our generic platform will know exactly what to do with that it understands the context of which logs are searching because I'm looking at a particular service. And it's going to convert, again, my English language question into a query that will be executed for me. And one of the things that's nice about this interface is because you see here that will tell you what the query is that was executed, you can learn how to write your own queries and execute them directly. So now here, we see some of the some of the results from the logs that contain errors. So again, we can either do this through a Gen AI interface, or if we want, we can click right through to the place where the logs are kept. And look into the detail of logs by just clicking on one of the log lines that has an error in it. And here's where things get even more insightful is because we can use New Relic API to explain what happens when there are errors. So in this case, we've identified a particular log line that has an error in it. And we can feed that to our LLM. And take in information from the production telemetry from the service from the logs. And even the code that was used to build the service in order to form a more complete picture of what might going might be going wrong. And so it's a really cool innovation, because it gives you insights more quickly than you might be able to do on your own doing research with a Google search. And before you
Keith Shaw 7:21
add it out, yeah, and before you added all of this, the generative AI component, it would have to be done through people that knew the platform and specific language queries and specific kind of like, you'd have to know where you're going to find the explanation of what the error was correct? Absolutely.
Peter Pezaris 7:37
And there's a lot of input into this equation, because you've got to be trained on observability, you've got to understand your the software that's running very deeply. And you've got to understand what these types of errors happen in the you've got also got have to have memorized Stack Overflow. And essentially, because when this particular error happens, there are things that you can find online that will guide you towards the solution. So what's fantastic about this interface is that you can have a conversation with an AI that understands all of those things, I can get you right to the solution. Even if you're not an expert yet
Keith Shaw 8:08
does this open up the platform, the New Relic platform to new users within a company that using it, you know, maybe you had a highly specific trained staff of three or four people that were using it now you can open that up to more developers, engineers, IT staff things like that you're spot
Peter Pezaris 8:23
on. Okay, so let me jump to another one. This is cool. Yeah. So we have dashboards as part of our service. And you can ask your AI to describe the dashboard to you. And it can tell you what these numbers mean, which is phenomenal. And we can even if this is a particular service, that's having some errors. So if I click on a particular error, you can see here we've got errors inbox, I can see one instance of this error is throwing me a stack trace. And a stack trace is sort of like a map within your code base where the problem happens. So all applications do this. So now we've created this capability to open this in your IDE. And this is through our Code Stream extension, where we've got the ability to load up your error StackTrace into the IDE, and I can click to navigate through the codebase to see where the problem is happening.
Keith Shaw 9:11
Let's it's probably where you lose me as for right. But for application developers
Peter Pezaris 9:16
like this, this is the coolest thing because we've got the information about the air, you've got production telemetry, and we've got your code and we feed it all into the LLM. And it can tell you what's going on. And it's amazing because not only can I tell you what's going on, but it can suggest the fix. So this step is automatically
Keith Shaw 9:37
it's not actually fixing it yet, or suggesting
Peter Pezaris 9:40
the fix. So there's still a human in the loop shocked. I'll have to look at the code that it's suggesting. But more often than not, it gets gets it right.
Keith Shaw 9:47
Have you been able to figure out a difference in the time like before the tool after the tool? You know, how much improvement are you getting?
Peter Pezaris 9:56
Yeah, from talking to our customers. We're about a 50% hit rate on like Whether the suggestion that we have is exactly the right fix, which is a pretty phenomenal accelerator to get faster resolution to your problem.
Keith Shaw 10:07
Sure. Anything else
Peter Pezaris 10:08
that you want to know that that was the that was the finale? So
Keith Shaw 10:12
all right, yeah. So, you know, we're hearing of a lot of companies that are doing in the observability space where they're adding a lot of these giant AI features. And, and, you know, is it the same thing or is, you know, as your is your products, you know, what, what differentiates you,
Peter Pezaris 10:27
I think the last six I heard was that 96% of the Fortune 100 are adding Gen AI capability. So this is no longer an experimental thing, like it's happening. And for every one of these companies, as they add features, like I just demoed to you, it's a brand new world, there's a new technology stack, there's a new set of metrics you have to care about, you know, bias and hallucinations. And, you know, lag becomes super latency becomes super important. And all of that represents an opportunity for things to go wrong, which means that you need observability to make sure things go right. And so New Relic also offers AI monitoring. So in addition to our ability to monitor your production estate, you can now plug in our AI monitoring capability into your software. And we'll let you know when you're having those problems.
Keith Shaw 11:19
Does that mean so that means you're using AI to tell me whether AI is doing it correctly.
Peter Pezaris 11:23
That's exactly right. Wow, that's cool. That's like aI all the way down. Yeah.
Keith Shaw 11:28
Peter, this is great. Where can people go for more details on this? Obviously, this is this is just a cut down version of the entire product. But where can they go for more
Peter Pezaris 11:35
info new relic.com. There we have a free trial generous free trial so you can a free tier that you can use in perpetuity to see if it works with your software stack. Okay,
Keith Shaw 11:44
cool. Thanks for the demo.
Peter Pezaris 11:45
Awesome. Thank you.