Exploring operational artificial intelligence: Insights from Networking for AI Summit
In this insightful session from the Networking for AI Summit, distinguished guests Shailesh Manjrekar of Fabrix.ai and Bob Laliberte of theCUBE Research engage in a thought-provoking discussion about operationalizing AI strategy within enterprises.
Shailesh Manjrekar brings extensive knowledge in AI and marketing from Fabrix.ai, showcasing expertise in the field. The conversation, led by Bob Laliberte, covers various critical topics such as the emergence of an agentic approach to operationalize AI, the integration of networking and AI operations, and the modernization of enterprise strategies. TheCUBE Research's video hosts offer analysis and insights to provide deeper understanding.
A key focus of the dialogue is the practical application of AI within enterprises. Manjrekar elaborates on the challenges and solutions in data collection, visibility, and automation, emphasizing the role of data fabric to bring intelligence to disparate data sources. According to Manjrekar, this innovative approach aids enterprises in adapting to the experience economy, enhancing brand awareness, and ultimately driving business value.
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
The Networking for AI Summit. If you don’t think you received an email check your
spam folder.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open the link to automatically sign into the site.
Register for Networking for AI Summit
Please fill out the information below. You will receive an email with a verification link confirming your registration. Click the link to automatically sign into the site.
You’re almost there!
We just sent you a verification email. Please click the verification button in the email. Once your email address is verified, you will have full access to all event content for Networking for AI Summit.
I want my badge and interests to be visible to all attendees.
Checking this box will display your presense on the attendees list, view your profile and allow other attendees to contact you via 1-1 chat. Read the Privacy Policy. At any time, you can choose to disable this preference.
Select your Interests!
add
Upload your photo
Uploading..
OR
Connect via Twitter
Connect via Linkedin
EDIT PASSWORD
Share
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
The Networking for AI Summit. If you don’t think you received an email check your
spam folder.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open the link to automatically sign into the site.
Sign in to gain access to The Networking for AI Summit
Please sign in with LinkedIn to continue to The Networking for AI Summit. Signing in with LinkedIn ensures a professional environment.
Are you sure you want to remove access rights for this user?
Details
Manage Access
email address
Community Invitation
Shailesh Manjrekar, Fabrix.ai
Exploring operational artificial intelligence: Insights from Networking for AI Summit
In this insightful session from the Networking for AI Summit, distinguished guests Shailesh Manjrekar of Fabrix.ai and Bob Laliberte of theCUBE Research engage in a thought-provoking discussion about operationalizing AI strategy within enterprises.
Shailesh Manjrekar brings extensive knowledge in AI and marketing from Fabrix.ai, showcasing expertise in the field. The conversation, led by Bob Laliberte, covers various critical topics such as the emergence of an agentic approach to operationalize AI, the integration of networking and AI operations, and the modernization of enterprise strategies. TheCUBE Research's video hosts offer analysis and insights to provide deeper understanding.
A key focus of the dialogue is the practical application of AI within enterprises. Manjrekar elaborates on the challenges and solutions in data collection, visibility, and automation, emphasizing the role of data fabric to bring intelligence to disparate data sources. According to Manjrekar, this innovative approach aids enterprises in adapting to the experience economy, enhancing brand awareness, and ultimately driving business value.
>> Hello. I'm Bob Laliberte, Principal Analyst with theCUBE Research. And I'm excited to give you a preview of one of our upcoming Networking for AI Summit sessions. I'm joined by Shailesh Manjrekar, Chief AI and Marketing Officer at Fabrix.ai. Welcome, Shailesh.
Shailesh Manjrekar
>> Well, thank you, Bob. Super excited about this topic and Networking for AI and how we play a role. And happy to talk to you. Thanks for including us as part of this session.
Bob Laliberte
>> Yeah. Absolutely. And in this video, we're going to discuss little bit about how enterprises can operationalize AI with an agentic approach. We'll explore why networking, data collection, visibility, and automation are critical to powering these next generation AI-driven ops. But before we dig into that, Shailesh, I don't know if everyone knows who Fabrix.ai is, so why don't you tell us a little bit about Fabrix.ai to start?
Shailesh Manjrekar
>> Yeah. Absolutely, Bob. So Fabrix.ai is essentially an agentic AI operational intelligence platform company. We've been around for about 10 years or so. We started as an AIOps company, and we have morphed into an agentic AI platform in this agentic era. And we partner with lot of big names, some of who are your participants as well, like Cisco, IBM, and we also have several enterprise, networking, telco, and service provider customers. So we think this is an ideal forum for us to talk about the value we add in Networking for AI.
Bob Laliberte
>> Excellent. I think that's great. And I wanted to maybe set some context first. Clear you guys have been around for about 10 years. You've seen the changes and the evolution that's happening. So as the enterprises are accelerating their AI adoption, what are the biggest challenge you're seeing around networking, data collection, and visibility?
Shailesh Manjrekar
>> Yeah. So that's a good question, Bob. And just kind of stepping back just to see what exactly is happening, we're really at the paradigm shift with this agentic era, right? And we call it the experience economy, where we have really evolved from just the availability of the network to the performance of the network to really the experience on how the end user views this, right? And it's really about hyper-personalization, it's about value creation, brand awareness, and so on and so forth. So networking becomes an very important aspect, I would say, in fact, like a nervous system for the entire agentic end-to-end play, right? And when it comes to data collection, obviously there is a rise of agentic AI applications now, both at what you call the front-end networks as well as back-end networks and the WAN as well. So the complexity involved is really ... There is disparate data sources, right? Somewhere you have wireless, like in Cisco's portfolio terms, say you have Meraki, you have DNAC, you have SEI, you have SD-WAN, right? You have Nexus Fabric, what have you, right? And these are different disparate data sources across , across data center, across optical, across mobility. And each of them have different data types. How do you bring in all of this data together, coalesce it, enrich it, and then make it ready for AI, is the primary challenge of what we see enterprises facing along with it, right? And that's where this construct of what we call data fabric emerges, right? So fabric is in our name. That a lot of companies not talking about data fabric, right? ServiceNow has been using it. Microsoft in the BI world. But essentially, the idea there is being able to bring intelligence to the source, being able to integrate with this data sources, take them, normalize them, enrich them, and make them ready such that you are able to now run AI on top of that, right? So that's what we're seeing. The challenges and some of the solutions which we foresee coming along.
Bob Laliberte
>> Yeah. I think that makes a lot of sense, especially given the fact that these environments are getting so much more distributed. That, as you mentioned, the additional data source, everything is being ... There's a sensor on everything. There's data being collected. Telemetry data is being collected and it's growing. So organizations really need to be able to collect everything if they're going to be able to provide the appropriate context.
Shailesh Manjrekar
>> Yeah. Exactly. And it's not just a telemetry. It's really AI needs context, right? So what we do in terms of real-time topology discovery and building the service map is extremely crucial when it comes to the AI agents. And more importantly, you also need to have a semantic layer on top of that such that the AI agents understand those edges and nodes in the network, right? So that's another crucial part of this.
Bob Laliberte
>> Yeah. No, that makes sense. So describe how you're helping organizations move beyond just the traditional automation and bring more intelligence into their operations.
Shailesh Manjrekar
>> Yeah. So I think you have framed the Networking for AI Summit extremely well in terms of networking. So you have the front-end networks, you have the back-end networks, and you have WAN, right? And we play smack in between all of this. So the value we add is threefold. So first and foremost, we provide you that cross-domain visibility. Being able to play with that data and normalize that data and provide you that end-to-end context across the stack is the first value we provide. The second is around the intelligence, right? Now you have all of this data, you have the curated data sets. What's the intelligence you're able to gather from this, right? In terms of performance, in terms of fault management, in terms of asset, operational insights, and so on. And the third step, which was lacking all along in the AIOps paradigm, is what I call the last mile problem, right? So now you know, you have determined that there is an issue. What do you do about it? Right? And that's where the automation part comes in, where you're able to actually take, perform an action to remediate that issue, or at least being able to partially solve that issue with a human in a loop or human on a loop, right? So those are three different aspects where we add value. And it goes back to a platform where we have that data fabric doing the data collection part, the AI fabric doing the AI fabric aspects, and the automation fabric doing the automation around it.
Bob Laliberte
>> Excellent. Well, that certainly makes a lot of sense. One of the other areas I wanted to explore with you is you've introduced the idea of AgentOps, the ability to use agents, AI agents to operationalize these AI environments. Can you share what that means in practice and why that's going to make a difference for the IT teams and also even potentially business leaders?
Shailesh Manjrekar
>> Yeah. Absolutely. I think that's a very important question, right? So clearly, the advent of LLMs have been extremely transformational, right? But LLMs, in our mind, it just to an means to an end. Otherwise, every single Tom, Dick, and Harry would've just used the LLMs, and they would've been as effective as anybody else, right? But there are a lot of other building blocks which go around the LLMs in terms of MCP tool sets, being able to provide you prompt templates, being able to provide the context engineering aspect, right? And this is what is needed for an enterprise-grade agentic platform. We have this concept of what we call least-privilege and least-agency, just like we had zero trust in the networking world. And what that means is you got to make sure that these AI agents are, A, not hallucinating. And B, if they hallucinate, there is only so little they can do in terms of damage, right? So when it really comes back to the value we provide with agentic, it's really about being able to provide that visibility, intelligence, and automation, right? And what I mean by AgentOps really is operationalizing this platform. So you take really an prompt, make sure that you have access to the right data sets and the MCP tool sets, you are able to provide the right guardrails for that particular prompt and that particular persona. And then you experiment to your satisfaction to ensure that now there is an agent ready to actually be put in production, right? And then it goes back to essentially kind of productizing that agent, right? You are evaluating that agent constantly. You're making that LLM learn as you go along. You are retiring that, as Jensen calls, the IT department now soon becoming the HR department for this AI agents, right? So that is really what we mean by AgentOps, that, hey, now you have an agentic operational platform, how do you operationalize this platform and consequently operationalize the networking for AI, which is a theme for this particular topic.
Bob Laliberte
>> Right. And that's clearly also going to, by doing that, it's going to give ... Think you mentioned a couple of times, you're going to be able to mitigate the risk of running those agent environments because you have visibility into what they're doing. And it's obviously going to drive a lot of productivity in terms of these, as you said, that transition from IT to HR, managing the agents and ensuring that they're doing what they're supposed to be.
Shailesh Manjrekar
>> Absolutely. That's rule number one, that do no harm, right? Because see, these LLMs, at the end of the day, they are non-deterministic processes, what we call stochastic processes. So an enterprise-grade platform really kind of works around those limitations and kind of gets you to that intent and goal, what you're trying to kind of establish through that prompt.
Bob Laliberte
>> I think that sounds good. And I guess a question for you, for everyone watching as well, ultimately, if you're able to put this in place, what type of outcomes should both the practitioners and the business leaders expect if they're able to embrace this type of agentic approach?
Shailesh Manjrekar
>> Yeah. So clearly, as I said, the low-hanging fruit is productivity enhancements. And these productivity enhancements go from basically shrinking the POV cycles to be able to do your mean time to resolution, remedial issues, and so on and so forth. So that's kind of the first low-hanging fruit. The second over a period of time is you are able to consolidate your number of tools, right? Because as you gain trust and you see, glean lot more insights, you don't need that many domain-specific tools. And lastly, we see these operational silos around networking, security, and IT operations that converging around that data. So we see these operational silos converging over a period of time, right? But have very realistic goals when you start your POVs in terms of productivity enhancement. The second point I would like to bring to the audiences, everybody in the networking and the telco and the MSP world wants to leverage AI, but this has not been traditionally the domains of the networking folks, right? And everybody wants to leverage AI, but where do I start? And what do I do? Right? Do I have to think about GPUs? Do I have to think about the AI stack I'm going to use? What are the outcomes I should be looking around? What we have done is we have launched a program called AAA AI Acceleration Program, where we make it lot more easy for enterprise, networking, and telcos, MSP customers to kind of operationalize this, put everything together so you don't have to worry about, hey, what GPU should be using, what LLM should be using, what are the tokens, right? So we, in fact, provide you a category or a tier of tokens free of charge when you start so that we are able to prove your value, right? So it's really about building trust, right? And then you get into showcasing the value. And then you get to the monetary part, right? So it's been a very different transformational journey for us. And we're very excited. I will provide you more details at our main session as well, Bob.
Bob Laliberte
>> Absolutely. No, I think that sounds great. I'm very excited to hear about the convergence and taking more of that platform approach. I do recognize the ... So I'm glad you said it's going to take time because there's people, process, and technology, right? And that people and process sometimes take a lot longer than the technology. But the great part about, I think, what we're seeing with AI is because there was this consumerization of it, it's really accelerating the adoption in the enterprise space as well and in the IT space. So looking forward to seeing how that evolves. So thank you very much for joining me today, Shailesh.
Shailesh Manjrekar
>> Thanks for having me, Bob. And very excited about actually the summit and sharing a lot more details around these topics. I think this is a very pertinent topic. And the role we play is absolutely pertinent in operationalizing networking for AI.
Bob Laliberte
>> Absolutely. And just a reminder, this is just a glimpse of what we're going to cover in our full summit session with Shailesh and Rached. So please join us as we take a deeper dive into how AgentOps is transforming IT operations and why agentic environments are redefining the future of networking in AI at our Networking in AI Summit. You're not going to want to miss it.
>> Hello. I'm Bob Laliberte, Principal Analyst with theCUBE Research. And I'm excited to give you a preview of one of our upcoming Networking for AI Summit sessions. I'm joined by Shailesh Manjrekar, Chief AI and Marketing Officer at Fabrix.ai. Welcome, Shailesh.
Shailesh Manjrekar
>> Well, thank you, Bob. Super excited about this topic and Networking for AI and how we play a role. And happy to talk to you. Thanks for including us as part of this session.
Bob Laliberte
>> Yeah. Absolutely. And in this video, we're going to discuss little bit about how enterprises can operationalize AI with an agentic approach. We'll explore why networking, data collection, visibility, and automation are critical to powering these next generation AI-driven ops. But before we dig into that, Shailesh, I don't know if everyone knows who Fabrix.ai is, so why don't you tell us a little bit about Fabrix.ai to start?
Shailesh Manjrekar
>> Yeah. Absolutely, Bob. So Fabrix.ai is essentially an agentic AI operational intelligence platform company. We've been around for about 10 years or so. We started as an AIOps company, and we have morphed into an agentic AI platform in this agentic era. And we partner with lot of big names, some of who are your participants as well, like Cisco, IBM, and we also have several enterprise, networking, telco, and service provider customers. So we think this is an ideal forum for us to talk about the value we add in Networking for AI.
Bob Laliberte
>> Excellent. I think that's great. And I wanted to maybe set some context first. Clear you guys have been around for about 10 years. You've seen the changes and the evolution that's happening. So as the enterprises are accelerating their AI adoption, what are the biggest challenge you're seeing around networking, data collection, and visibility?
Shailesh Manjrekar
>> Yeah. So that's a good question, Bob. And just kind of stepping back just to see what exactly is happening, we're really at the paradigm shift with this agentic era, right? And we call it the experience economy, where we have really evolved from just the availability of the network to the performance of the network to really the experience on how the end user views this, right? And it's really about hyper-personalization, it's about value creation, brand awareness, and so on and so forth. So networking becomes an very important aspect, I would say, in fact, like a nervous system for the entire agentic end-to-end play, right? And when it comes to data collection, obviously there is a rise of agentic AI applications now, both at what you call the front-end networks as well as back-end networks and the WAN as well. So the complexity involved is really ... There is disparate data sources, right? Somewhere you have wireless, like in Cisco's portfolio terms, say you have Meraki, you have DNAC, you have SEI, you have SD-WAN, right? You have Nexus Fabric, what have you, right? And these are different disparate data sources across , across data center, across optical, across mobility. And each of them have different data types. How do you bring in all of this data together, coalesce it, enrich it, and then make it ready for AI, is the primary challenge of what we see enterprises facing along with it, right? And that's where this construct of what we call data fabric emerges, right? So fabric is in our name. That a lot of companies not talking about data fabric, right? ServiceNow has been using it. Microsoft in the BI world. But essentially, the idea there is being able to bring intelligence to the source, being able to integrate with this data sources, take them, normalize them, enrich them, and make them ready such that you are able to now run AI on top of that, right? So that's what we're seeing. The challenges and some of the solutions which we foresee coming along.
Bob Laliberte
>> Yeah. I think that makes a lot of sense, especially given the fact that these environments are getting so much more distributed. That, as you mentioned, the additional data source, everything is being ... There's a sensor on everything. There's data being collected. Telemetry data is being collected and it's growing. So organizations really need to be able to collect everything if they're going to be able to provide the appropriate context.
Shailesh Manjrekar
>> Yeah. Exactly. And it's not just a telemetry. It's really AI needs context, right? So what we do in terms of real-time topology discovery and building the service map is extremely crucial when it comes to the AI agents. And more importantly, you also need to have a semantic layer on top of that such that the AI agents understand those edges and nodes in the network, right? So that's another crucial part of this.
Bob Laliberte
>> Yeah. No, that makes sense. So describe how you're helping organizations move beyond just the traditional automation and bring more intelligence into their operations.
Shailesh Manjrekar
>> Yeah. So I think you have framed the Networking for AI Summit extremely well in terms of networking. So you have the front-end networks, you have the back-end networks, and you have WAN, right? And we play smack in between all of this. So the value we add is threefold. So first and foremost, we provide you that cross-domain visibility. Being able to play with that data and normalize that data and provide you that end-to-end context across the stack is the first value we provide. The second is around the intelligence, right? Now you have all of this data, you have the curated data sets. What's the intelligence you're able to gather from this, right? In terms of performance, in terms of fault management, in terms of asset, operational insights, and so on. And the third step, which was lacking all along in the AIOps paradigm, is what I call the last mile problem, right? So now you know, you have determined that there is an issue. What do you do about it? Right? And that's where the automation part comes in, where you're able to actually take, perform an action to remediate that issue, or at least being able to partially solve that issue with a human in a loop or human on a loop, right? So those are three different aspects where we add value. And it goes back to a platform where we have that data fabric doing the data collection part, the AI fabric doing the AI fabric aspects, and the automation fabric doing the automation around it.
Bob Laliberte
>> Excellent. Well, that certainly makes a lot of sense. One of the other areas I wanted to explore with you is you've introduced the idea of AgentOps, the ability to use agents, AI agents to operationalize these AI environments. Can you share what that means in practice and why that's going to make a difference for the IT teams and also even potentially business leaders?
Shailesh Manjrekar
>> Yeah. Absolutely. I think that's a very important question, right? So clearly, the advent of LLMs have been extremely transformational, right? But LLMs, in our mind, it just to an means to an end. Otherwise, every single Tom, Dick, and Harry would've just used the LLMs, and they would've been as effective as anybody else, right? But there are a lot of other building blocks which go around the LLMs in terms of MCP tool sets, being able to provide you prompt templates, being able to provide the context engineering aspect, right? And this is what is needed for an enterprise-grade agentic platform. We have this concept of what we call least-privilege and least-agency, just like we had zero trust in the networking world. And what that means is you got to make sure that these AI agents are, A, not hallucinating. And B, if they hallucinate, there is only so little they can do in terms of damage, right? So when it really comes back to the value we provide with agentic, it's really about being able to provide that visibility, intelligence, and automation, right? And what I mean by AgentOps really is operationalizing this platform. So you take really an prompt, make sure that you have access to the right data sets and the MCP tool sets, you are able to provide the right guardrails for that particular prompt and that particular persona. And then you experiment to your satisfaction to ensure that now there is an agent ready to actually be put in production, right? And then it goes back to essentially kind of productizing that agent, right? You are evaluating that agent constantly. You're making that LLM learn as you go along. You are retiring that, as Jensen calls, the IT department now soon becoming the HR department for this AI agents, right? So that is really what we mean by AgentOps, that, hey, now you have an agentic operational platform, how do you operationalize this platform and consequently operationalize the networking for AI, which is a theme for this particular topic.
Bob Laliberte
>> Right. And that's clearly also going to, by doing that, it's going to give ... Think you mentioned a couple of times, you're going to be able to mitigate the risk of running those agent environments because you have visibility into what they're doing. And it's obviously going to drive a lot of productivity in terms of these, as you said, that transition from IT to HR, managing the agents and ensuring that they're doing what they're supposed to be.
Shailesh Manjrekar
>> Absolutely. That's rule number one, that do no harm, right? Because see, these LLMs, at the end of the day, they are non-deterministic processes, what we call stochastic processes. So an enterprise-grade platform really kind of works around those limitations and kind of gets you to that intent and goal, what you're trying to kind of establish through that prompt.
Bob Laliberte
>> I think that sounds good. And I guess a question for you, for everyone watching as well, ultimately, if you're able to put this in place, what type of outcomes should both the practitioners and the business leaders expect if they're able to embrace this type of agentic approach?
Shailesh Manjrekar
>> Yeah. So clearly, as I said, the low-hanging fruit is productivity enhancements. And these productivity enhancements go from basically shrinking the POV cycles to be able to do your mean time to resolution, remedial issues, and so on and so forth. So that's kind of the first low-hanging fruit. The second over a period of time is you are able to consolidate your number of tools, right? Because as you gain trust and you see, glean lot more insights, you don't need that many domain-specific tools. And lastly, we see these operational silos around networking, security, and IT operations that converging around that data. So we see these operational silos converging over a period of time, right? But have very realistic goals when you start your POVs in terms of productivity enhancement. The second point I would like to bring to the audiences, everybody in the networking and the telco and the MSP world wants to leverage AI, but this has not been traditionally the domains of the networking folks, right? And everybody wants to leverage AI, but where do I start? And what do I do? Right? Do I have to think about GPUs? Do I have to think about the AI stack I'm going to use? What are the outcomes I should be looking around? What we have done is we have launched a program called AAA AI Acceleration Program, where we make it lot more easy for enterprise, networking, and telcos, MSP customers to kind of operationalize this, put everything together so you don't have to worry about, hey, what GPU should be using, what LLM should be using, what are the tokens, right? So we, in fact, provide you a category or a tier of tokens free of charge when you start so that we are able to prove your value, right? So it's really about building trust, right? And then you get into showcasing the value. And then you get to the monetary part, right? So it's been a very different transformational journey for us. And we're very excited. I will provide you more details at our main session as well, Bob.
Bob Laliberte
>> Absolutely. No, I think that sounds great. I'm very excited to hear about the convergence and taking more of that platform approach. I do recognize the ... So I'm glad you said it's going to take time because there's people, process, and technology, right? And that people and process sometimes take a lot longer than the technology. But the great part about, I think, what we're seeing with AI is because there was this consumerization of it, it's really accelerating the adoption in the enterprise space as well and in the IT space. So looking forward to seeing how that evolves. So thank you very much for joining me today, Shailesh.
Shailesh Manjrekar
>> Thanks for having me, Bob. And very excited about actually the summit and sharing a lot more details around these topics. I think this is a very pertinent topic. And the role we play is absolutely pertinent in operationalizing networking for AI.
Bob Laliberte
>> Absolutely. And just a reminder, this is just a glimpse of what we're going to cover in our full summit session with Shailesh and Rached. So please join us as we take a deeper dive into how AgentOps is transforming IT operations and why agentic environments are redefining the future of networking in AI at our Networking in AI Summit. You're not going to want to miss it.
Bob Laliberte