Navigating SaaS in the AI Age with Ryan Manning

This is Techron tv. Hey guys, thanks for the throw. We're here with Ryan Manning, who's vice president of product at BMC software, leading the charge for Helix Service Management and the platform team.

And we're gonna be talking about the impact that data clouds are gonna have on SaaS applications. Hey Ryan, welcome to she, Thanks for having me. We've seen the rise of these data clouds.

There's a lot of 'em out there. And SaaS applications typically been, you know, starve for data on the, when they first get started and kind of reluctant to let data go once they do get a hold of it. So how do you think that the way we think about SaaS applications is gonna change in the era of the data cloud?

Yeah, yeah, that's a, it's a great question. One we've been looking at really closely. Um, you know, I I, I'm relatively new to BMC, but even prior to my time at BMC, uh, at Coupa, I was noticing this trend of, you know, folks being much more interested in getting data out of Coupa and getting it into a data cloud.

And I continue to see this at at, at BMC. And, um, you know, with the emergence of AI and ML and now gen ai, it's never been more important to get your data in one place. And data clouds a great way to do that.

And, uh, you know, we like to say over here that those with access to the largest amount of enterprise data and the closest to real time will, will win in ai. And so that's been interesting for us to work more closely with the data in cloud rather than build around. But more interestingly is that some of them have began building native app frameworks that allow workflow vendors like us to deploy our applications through containers, uh, onto the data cloud natively.

And we think this could be pretty disruptive to the wider community of at least, uh, our perspective enterprise software. So do you think AI is forcing that conversation? 'cause data clouds have been around for a bit, but do you think that, um, IT leaders are starting to realize that they need all that data centralized as much as possible?

Very good point. And no, I don't think, uh, you know, it's, it's been interesting. I see a lot of anxiety out there in INO teams.

You know, it's like everyone's having fun in the enterprise fooling around with these gen AI tools, but it's really happening in sales ops and marketing ops and, uh, product operations. And, uh, so we, we, we, we are starting to see an uptick now in IT workloads making their way over to data clouds. And we feel like it could increase because of, to your point, to your point, uh, ai, but there's also the data governance and security aspect of AI that makes, um, putting your data and data cloud technologies who have had to worry about this in a very deep way for a number of years, um, to ensure, uh, you know, proper ai rather, Will this drive a consolidation among the SaaS providers because a lot of them will, uh, maybe become, uh, parts of larger enterprise workflows, or we essentially, you're reinventing ERP suites here as we go along by rolling all these things up.

Yeah, it's a, it's another good point. You know, I, I, I talk about internally we have this rallying cry internally, it's the, the best UI is Jet ai. We're gonna conversationalize the entire user experience on the service management side of the house.

And, um, I have a personal goal of eliminating about 60% of the user interface in favor of these conversational experiences. So I don't need to invest in all this CUI, um, for the users, for example, portals, right? Everybody loves their portals and we can brand it, we can make it look, you know, like our company website and things like that, but end users never wanted to go to portals, right?

They want, and a lot of them, especially the younger generation, just want to go into teams and, and inter interact with a bot. So these applications are becoming very lightweight or or about to be, so they're not as heavy. So the concept of deploying through containers onto a data cloud and and running it there is more, is more realistic.

And, you know, speaking of containers, uh, you know, certainly one of the reasons why I joined BMC is because, you know, we had a large enterprise, uh, enterprise install base on premise. And so we needed to worry about, you know, getting them close to cloud-like capabilities, like, you know, updating, um, the software to the latest version, but still on premise. And so Docker and Kubernetes have, have been enabled us to do that.

And, uh, uh, we think that that's a another, uh, interesting, uh, trend going on where, you know, is certain companies they, they're gonna be able to monetize their data for the, for, they didn't even think that that would be a revenue champ for them. So I think folks are gonna get pretty guarded about where their data, where their data goes, Will these data clouds change the way we think about securing that data? Because a lot of it is kinda, uh, gonna be shared across different application use cases and then invoked via LLMs.

And it's not clear to me that we've kinda worked all that out yet. No, it's, it's probably, it's, you know, when you talk to those, those companies, it's, it's, uh, the highest importance right now. Um, and you know, when you're, a lot of these teams, the InfoSec team and legal teams are starting to wise up to what's going on.

And, uh, I think there's a little bit element of like, it's a, there's a little bit of cowboy land going on over o over there and some of these tools. And I think more and more of that is gonna get looked at in the purview of InfoSec and legal. And now I don't think that we've figured it out yet.

I think the jury's out on how we're gonna solve it. But, you know, like I said, that, you know, these native data cloud companies have to worry about this. 'cause then you've been collecting data from multiple data domains and having to worry about access, uh, for some time.

So it makes sense that that team would be heavily involved. Um, the team in the enterprise that, you know, the data engineering team, um, that manages the data clouds, Where will the automation enabled by the LLM and the reasoning capabilities built in, um, leave off and the need for a separate automation framework begin? 'cause I think a lot of folks that I talk to are trying to figure that out right now.

'cause a lot of times they, the LLM has a certain amount of capability, but maybe not enough. And so what Yeah, platform's gonna do what? Yeah.

I'll give you just a quick, uh, example of, uh, some early customers we have working on this. So we have a, a large telco customer, and what they've done is they in, in, in their operations management group, they've sent data from multiple data, main data domains into a data cloud. They've got their Broadcom data from test SecOps data from Splunk, uh, their ITSM and IO data from BMC and then their dev and build data from Atlassian and Jenkins.

And they just threw it all in. Um, and they were for the last, you know, six to nine months, uh, you know, building analytics on top of that with AI and ml. But then they started sprinkling on gen ai, uh, and, and now our Helix GPT product and started having conversations with the data.

So, you know, I asked, I even get got the example of, hey, you know, we've got this homegrown sales quoting tool that gives us a bunch of headaches. Can you take a look at this? And it comes back and says, yeah, you know, you know, within 90 days after every release, you get an influx of P one and P two incidents, and you see again, like who shows a report?

It builds a report in the moment, and then you, you can dive deeper into root cause came back and said, Hey, your, your DEB team out in Estonia is not doing any unit testing. Here's the links to the Jira records. You might wanna take a look at that.

So it's, and, and for them it was like, for me it was why could, could this be the solution to break down the barriers, the culture wars that exist across all these teams, right? And for them it was like, this was ops, it was like ops, it was a tool to give them better information or have data informed conversations with development in that case. And I don't know if I'm answering your question or if I drifted off there a bit, but Well, do you also think, I mean, we're talking about in the context of SaaS, but there's still a lot of data residing in on-premise environments as well.

Can we maybe leverage gen AI to have a more federated approach the way we think about, uh, data regardless of where it happens to be residing or how much gravity is holding it in place? Yeah, yeah. I know, I, I don't know all about that, but I do know we're, we're experienced, we're really leaning into this concept of large action models at at BMC.

We're, we're, we're, um, building a fleet of robots, I like to say, um, that will act as virtual assistance to the processor knowledge workers we're, you know, we, we've gotten the end user use case, right? Chat bot, everybody's got that. But now we're looking at how it can be a partner into, into the workers that I use our technology, um, uh, day to day.

And as we look at these different agents, it's become shocking what they're capable of. It's quite scary at times, if I'm being completely honest. We just build a SQL DB agent that can, you know, through a very small set of operating procedures, can go into a SQL database and, uh, you ask it a natural language question and it will go grab the answer without having to move the data.

So that's like, I mean, we're talking like a couple weeks ago I saw this demo and this, uh, during this spiking event, it's just all this is happening so quickly. It's just wild. How will all those agents be orchestrated?

And I'm asking the question because a lot of folks are like saying, well, we have various copilots and maybe the copilots are gonna be extended to kinda run everything. Or the co-pilots will need interact with other agents that are specially trained for specific tasks. And it's gonna feel a lot like, I don't know, for lack of a better phrase or comparison, it's kind of like Downton Abbey.

There's one agent in charge of all the other servants and they all kind of know what to do because they talk to each other and no one knows exactly how it was accomplished, but things get done. Yeah, it's been a really fun conversation with customers on this. You know, they're just being presented.

Here's Einstein agent, here's your Coupa assist and Zoom assist and co-pilot and blah, blah, blah, blah. Who's gonna sit there and help orchestrate all this work? And we, you know, our approach, again, has been very open.

We connect to the tools that you're either already using or just about to start using. And the way that we've architected this platform for Helix GPT, we think there might be a lane here for us to help. And so we're working with one major hyperscaler, um, on, on doing this work and through some design partners right now.

It's been really fun. Someone is gonna need to do it. And interestingly enough, like we at BMC are, you know, se several business units and one of 'em is a market lead leading DBA solution and Control M which has been orchestrating data pipelines for quite some time, massive workloads.

And so we're looking at them and, and thinking there might be a, a really good opportunity to work with that, with that team to, to build this partner with them to build this solution. Do you think ultimately as we think about this, that the way our IT teams today are structured will fundamentally change? There's all these data science running around, there are, uh, developers and data engineers and security people, and maybe, you know, we're applying, uh, yesterday's model to try to fight tomorrow's battle and that never ends.

Well, It's really, it, yeah. It goes back to what I was saying earlier. I mean, I'm, uh, there, in some cases, in most cases that I've seen of the large enterprise, like they're in the same organization, the data teams and the INO teams, right?

But they don't really talk to each other. Um, and so you've got this person that, you know, I'm used to talking to, which is the IT app developer who wants to, you know, create innovative solutions on our platform to solve business problems within the business. But what is the, what is the future definition of an application?

Is it really gonna be forms and lists and workflows and DA dashboards, or is it gonna be, you know, AI infused applications or just conversation channels or agents that they need to build? So I'm trying to, to bring these teams together. Um, and it's been interesting 'cause you talk to the, the data engineers, you know, they don't really, you know, think in the, you ask them the definition of an application, you'll get a completely different answer if you asked the IT app dev folks.

So they can help each other, they need to partner up. And I, and I feel like it's, uh, our responsibility at BMC and our customer base to, to create that connection and even another reason to look at control m uh, and that's solution to bring the two teams together potentially. What's your best advice to folks then, as they start working with trying to bring together data across multiple SaaS applications and they're got these data clouds and they're have an eye on ai.

What's the one thing that they should probably be focused on now to kind of ensure a success leader? Yeah, I, again, I feel this great responsibility. 'cause I think we're, you know, we're 12 months, six months away from the CFO coming down the CIO street and asking for significant efficiencies.

Um, and so I think the natural reaction right now from our customer base, because they're not at involved in that work, is like, can you just provide this whole thing as a service to me? Right? And I get that, I get that, that, that, that it would be easier at first, but I encourage folks to get in, in involved in this conversation internally.

You don't need to use the same OpenAI account that your sales ops people are, uh, are using. You can spin up another one, um, and you can get access to that data cloud technology and we can help you do that. Um, but you know, we see a future that's, um, that's very open and that we're connecting to the tools that you are already using or about to use.

And so it's kinda like that advice you get where someone, you know, just in your day-to-Day personal lives who haven't used chat TPT yet, you're like, just go create 13 conversations, just have 13 conversations with it, and you'll start to understand what this thing is capable of. Um, get involved. Yeah, it'll, it'll be good for your career.

All right, folks. I think we said the most important thing to do right now is experiment. Because with experiment comes the gaining of experience.

And that's when you'll know what to do with all this stuff. Hey Ryan, thanks for being on the show. Thank you.

All right. And back to you guys in the studio.