Automation & IT Projects – VS Joshi, Digitate

This is texturong TV. Hey guys. Thanks for the throw.

We're here with vs. Joshi who's Global head of products and solutions for digitain and we're gonna be talking about what's going on with it projects and possibly the impact aiops is about to have but yes, welcome the show in the first question. I want to get to off the top of the bat is so many IT projects fail.

And you know, if we were a baseball team, this is our Banning average we probably against so my question to you is yeah what is wrong with the current state of it? And why are we encountering so many failed projects? I think I love the analogy that you give just now and yes, it is true that about 83% 83% of the IT projects fail.

Now, I keep projects failed because of multiple reasons. It can be lack of planning. It can be lack of resources.

It can be the objective is not very clear. Talent is not there that teams are not aligned. They're multiple multiple reasons why the IT projects can fade and there are some places where aiops helps in essentially trying to mitigate the risks that are associated with this failure of IT projects.

And how does that work is a lot of people are of course skeptical of all things AI these days but the issue is how in a machine learning algorithm or any other algorithms that matter kind of being able to know or anticipate when something was gonna likely fail Okay, so see technically speaking AI Ops is the application of AI and machine learning to it operations to business operations. And that application of AI and machine learning is essentially to essentially provide with the cross domain analysis to derive insights to predict failures and to prevent failures. That's what AI OBS does it essentially look at it this way that within an IT environment.

Also, you have multiple domains networking storage servers applications digital experience monitoring and all these various domains have their own monitoring to less such. So there isn't a single overarching tool essentially which encompasses all these various domains. So getting that cross domain analysis getting that cross domain analysis becomes an extremely difficult thing and that is what aiops provides it provides.

It is an aggregation of all the monitoring tools that are there and it provides across domain analysis. And that's the reason why it's very easy for AI of solution to derive certain insights and to take certain actions. proactively How does it learn what's going on in the environment?

Because at least as I understand it, it takes a while for the algorithm to kind of figure out what's going on and then be able to predict when something may or may not fail but yes. Yes. I have something new that I just added to the environment which is where that thing is.

Most likely to fail. How does it know about the new thing versus all the things that have already been there. Yes, I think so very good question.

And essentially if you look at the AI machine, so they're called as prediction machines. In fact, I have a book over here. I can check it later.

It's called prediction machines. And essentially the main purpose of an AI machine learning product is essentially the prediction aspect of it. Now, how does it predict it predicts by essentially a couple of reasons a couple of ways.

There is a historical data. There's a historical data that is available to various companies. And so from this historical data, there is a pattern recognition.

There is an analysis of this particular data and based upon that and Our product let's say for our products. Our product is called as the new AI Ops we essentially have a corpus of knowledge that we have derived because through our parent company wherein we have this knowledge as to what can and cannot can go wrong within an IT environment and all this knowledge through all these knowledge. We have used it to essentially create this inference engine or a prediction machine you can say and that is what we are providing to the customers.

So there is and we have almost like 10,000 or so out of the box automation libraries. So through this knowledge that is there with us number one and number two. Once it is in the customers environment.

It gets the data it gets the customers data. It understands the patterns. It understands.

So what are the what are the what are the issues and how those issues were taken care of previously based upon that based upon a certain rule-based case based and model based algorithms. We are in a position to understand what is going wrong. And what is the action to be taken?

There are so many people out there who are talking about aiops these days in your mind what differentiates one platform from the other? Again a good question. I think see if you look at AI Ops.

I mean people people limit the concept of AI Ops to essentially Gathering the data doing the analysis and predicting what we do differently and we call our AI of solution as a closed loop AI of solution we call it closed loop because it's not only about gathering this information. It's not only about another analysis. It is not about only about prediction and It is also about taking action.

It is also about taking action so that our software product It fixes the problem it fixes the problem automatically. There are many companies who will stop at the prediction aspect of it. And then they have to get another product from another company that will automate that whole thing.

But with ignou AI option with that is one of our differentiating factors is we are a closed loop AI of solution and because the closed loop yeah Absolution we take care of the issue not only in finding the root cause of the problem but also trying to find also find a fix for it and if the fix is not available, we can send that request to a manual cue wherein again we help with the triaging aspect of it has to be who is the subject matter expert who can take care of this problem. It gets routed to that particular person. So this entire thing is what we call as close loop.

Yeah, because we take action. We not only just prescribe and predict. Yeah.

One wag one said you know, it's one thing to be wrong. It's another thing to be wrong at scale. So how do I get trust in the system to know that what they're going to automatically fix will work and not make things worse?

Super okay. That's a good question again. Uh, yes, I think what we I mean again now over here.

What we are recommending our customers is and again, you're right that many times customers are not very are not completely ready to give the entire controls to a machine or to a software. They want to have that control. They want to essentially decide as to They would like to come into a picture and take an I take an action themselves.

So we provide we provide that ability. We provide that feature by which we will essentially again, we we can predict certain things and we can suggest these are the couple of actions you can take and if in case the operator would like to take a certain action by themselves, they can do it by themselves or they can wait for us to do it. And yes, once they see the couple of times that okay.

Yeah. This is completely taken care of by the by the igneous software then they that trust bills into the system but it does take certain time for people to get comfortable with it and somewhere I feel we as vendors we have to provide that capability by which you give the operator you give the customers the option of essentially deciding how much I mean They need to you know, like they have to dial as to how they have to dial the whole thing as turkey are how much automation I need and how much I automate and we provide that facility. We provide that feature by which they can do certain things so that the trust is built into this so that the trust is built outside.

When we come full circle on this conversation, do you think that perhaps we're a little too aggressive in the number of projects we launched and we should scale back our Ambitions because we know that so many of them are going to fail. So do we need to do a better job triaging up front which ones we're picking? I think it's a it's a very double edged sport you can say because We are an extremely competitive world.

I would say and if in case we don't take the button and if we don't move forward, our competitors will do that and because our competitors will do that. You will be completely out of the market because you were not there as when the situation when you know, when it was required. I mean think of it.

I mean nobody predicted covid and all those companies that did well were the ones who had already put certain things in place this put certain things in place much in advance so that they were able to write that entire covid working from home working remotely kind of a scenario as that having just no physical presence, so I do feel that somehow I think Which IT projects we get into what we do is also a function of the competitive industry of the computer domain that we are in. And because all the other competitors are moving fast, and if you are not moving fast enough, you will lose that way also, so you have to move fast you have to ensure that you mitigate all those risks that are associated with the projects failing outside. All right sounded to me like BS is arguing that experimentation is the cost to do in business these days and just have to suck it up butter and even put sir very well put very good very good application.

Yeah. Hey vs. Thanks for being on the show.

Thank you very much, Mike. Thanks a lot. Really enjoyed this.

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