The Transformative Impact of AI on the Services Economy with Certinia’s Raju Malhotra

Hello, the latest edition of the Text on AI video series. I'm your host, Mike Bazar. Today we're with Raju Mara, who's the CTO for Satia, and we're talking about the impact AI is gonna have on the services economy.

Hey, Raju, welcome to show. Thank you, Mike. When you think about it, we've been kind of relying on services to drive the economy, at least here in the US and maybe in Western Europe and other places for a long time now.

I mean, it's, it is a huge portion of the GDP. Um, now we have AI coming. How do you imagine that AI is gonna change the way we think about services, and what should we be maybe getting ready for?

Yeah, I think it's a, it's a good question. It's a multifaceted question, uh, because on, um, the very surface of it, any new technology, you know, requires a lot of help, uh, that the professional services provides. You know, it happened with the internet, it happened with digital transformation.

So it has a positive impact on the humans to provide the professional services to accelerate the adoption of, uh, you know, ai, gen, ai agent ai, et cetera. So I think the first impact is, uh, the professional services itself as a delivery organization is going to help deploy AI more for their customers, particularly in mid-market enterprise customers, et cetera. There are a lot of benefits to those customers, and I think over time, um, we are moving towards a hybrid workforce where you have the humans delivering some professional services, and you also have some, uh, incarnations of digital workers that are the agentic ai, uh, you know, uh, incarnations that help, uh, in conjunction with humans to actually deliver.

So I think the, the impact is going to be profound. Uh, it's, uh, somewhat clear. I think in the short term, it's going to be accelerating, uh, the adoption productivity, uh, improvements and professional services is growing in that sense.

Uh, but over time, I think it remains to be seen, you know, how that equation works out. To your point, I don't think most organizations have the skills necessary to deploy AI and AI agents, so they're gonna rely heavily on professional services. But elsewhere, it also seems like the AI will, uh, automate a lot of the manual tasks that required to implement something.

And the, the length of the engagement may be shorter as more of those things get automated using some sort of AI agent. Is that a, is that likely? I, I, I think that's likely, and I, I also think, uh, it's important to peel the onion on when they say ai, what do we actually mean?

Because I think, right, right now, uh, it is a very overloaded term, um, from a C'S perspective, we think about AI in three categories, three types of ai, predictive ai, generative ai, and agent ai. And I think a lot of focus these days is about agent ai, which really helps you automate some tasks and actually have some level of autonomy and decision making on behalf of you within the guardrails to accomplish, uh, those tasks. But let's not forget, I think the predictive AI has been there for a long time, for many decades, and that continues to play in very clear automation, productivity benefits.

Those are very well understood. Generative AI is becoming very prevalent, very useful, uh, for summarization, simplification, creation of content, creation of images, et cetera, et cetera. I think that is increasingly getting adopted.

And agen AI is exactly, I think to your point about what the, where the puck is moving, where there's, there's a lot of, uh, potential benefits and potential risks. So I think the agen AI is probably more about, you know, how that productivity, uh, really affects, uh, the workforce. But we should really think about, I think productive ai, generative AI are here to stay.

They are actually having already a lot of profound impact. Mm-hmm. As you think there's, through a little bit, I might argue that the number of professional services engagements has been limited because the cost has been higher.

But if I have, uh, a shorter engagement periods, might I not have more projects as we go along? Because there was plenty of things that we never get around to doing simply because the Total cost Was too high. So maybe we'll have more engagements.

Yeah. I think overall, really it comes down to, um, what exactly is the role of professional services, you know, short engagement, uh, specific engagement, time bound, or, uh, you know, exactly what the, what the kind of benefit does. I think there's a big change in that because it is moving from a time and material based model of hiring some consultants to get some jobs done for professional services to more of an outcome based, you know, model also.

But to your point about, I think, uh, would it open up the aperture to do more with professional services? Absolutely. I think, uh, is it actually increasing the revenue for professional services already?

I think that is also true because in conjunction with the, the new tools that the professional services organizations have, they are actually getting more and more outcome focused also. So they're delivering better value and they're able to charge a better rate, better, you know, revenue out of the engagements, even if those engagements are, uh, shorter in some cases. Mm-hmm.

Well, the nature of the services being delivered change as well. Historically, um, professional services team would come in and engage, and they'd probably move in for anywhere from six months to two years sometimes, depending on what their project is. I wonder, though, if we're gonna rely more on agents to deliver those services, is the, is the professional service is gonna be more continuous, and it will be something that we're kind proactively managing on behalf of some customer somewhere.

Uh, but it's just always on. Yeah, and I think the, if you really think about what is happening in, in, in those few months, few quarters, few years of that professional services engagement, if you peel the onion and, and, and what's happening is there is some type of integrations that are happening amongst diff, you know, different systems that the enterprises have. There might be some, uh, custom code development that is happening that is actually either a glue code or a new kind of, you know, application.

There might be some cleaning of the data. So I think if you peel the onion, a few LA layers deep, then you really, you know, can start thinking about the application of agent AI in those cases. We know Agen AI actually works very well for a lot of development type of use cases.

Cursor, Devon, GitHub, uh, copilot, uh, anthropic cloud, et cetera, have already have had a lot of impact on reducing the time it takes to produce some code. And it could be in conjunction with a smaller human team for a shorter amount of time, but the output of, uh, uh, that, uh, project can be significantly increased. So I, I think, uh, you are right that a, I think it gives a lot more control in the hands of end customers who are actually buying these professional services, uh, engagements, but it also improves the level of predictability and the value that they can receive by using agent AI in conjunction with the human teams.

Mm-hmm. One of the things that we rely on for professional services a lot historically, has been these kind of application modernization projects where I had to go in and reverse engineer something and refactor. And I wonder if that's gonna become a lot simpler because, uh, the amount of time it might take to, uh, convert something from, uh, uh, one programming language to another, and these are all where those lock-ins have been.

So is the cost of switching platforms gonna drop? Well, I think, uh, in, in the case you're talking about, uh, the data is locked up in different applications and not just one platform, but multiple different platforms and quote unquote legacy, which might be, some of them might be on-prem, some of them might be in cloud one, cloud two, et cetera. So it, it can be a very onerous sort of, you know, exercise, uh, to do that through the old way of doing human, uh, the professional services approach.

I actually absolutely agree with you. I think it, it actually changes, uh, the realization of value from new technology because you really think about, instead of piecing together different silos, you really think about how do I get to the core of my organization's data? It could be employee data, it could be customer data, it could be business data, it could be different types of transactions.

And then instead of relying on the, uh, different software logics that, those appli custom applications, or even the off the, uh, you know, uh, uh, standard sort of applications, how do I actually apply the agent AI to connect those data points and create that right fabric to achieve my outcome? So for example, if I have to take an action for, uh, closing a customer deal, I can actually use my agenta and the MCP connectivity or a two, a agent to agent type of connectivity to actually much more quickly get into my employee data, get into my skills, and, uh, consultant profile if it's a professional services organization, and move much more rapidly, uh, instead of being, uh, encumbered by, uh, the, uh, integrations that I might have to do otherwise. So I think, uh, yes, it actually opens up, uh, that possibility of agility, uh, and really comes down to, uh, leapfrogging, uh, in many ways, uh, the state of, uh, affairs that might be in some enterprises.

I don't think we've used this term in a while, but it sounds like maybe this is a massive exercise in business process re-engineering, and this time we may have to go in and actually think about a lot of the processes that we have. Sometimes I feel like there's more exceptions than there are rules. So are we gonna have to go in and make it a little more structured for these AI agents to automate it?

Uh, I think you're right. I think it, it, it would require re-engineering of the processes. One of the caveats, however, is that even when business process re-engineering was done well, and I say e when, when, because there were a lot of sort of exceptions to that rule.

Those processes were designed for humans, and humans have a lot of qualities as as, as obviously, you know, we're defending our sort of human part of it, but there are a lot of inefficiencies that do not apply for agents. So I think the business process reengineering would be very different if you think for, for the agent to agent communication and, uh, the machine type of communication. And that's more of a connectivity with the right guardrails and permissions.

Um, but the workflow and the processes can be significantly accelerated and faster, uh, compared to I think what the business process engineering of the olden times has been. So, so yes, I think at, at some point it is about reengineering that, but uh, really I think it comes down to the level of, uh, application to the agent scenarios, which is very different from human scenarios. Right?

So taking that to the next logical level, on the one hand, professional services firms will have their own AI agents and customers will have their own AI agents. And how will these AI agents kind of communicate with each other to create something that goes to that outcome that we're talking about? Because it's one thing to get my AI agents to talk to themselves, nevermind talking to somebody else's.

Yeah, I think it's a sort of embarrassment of riches kind of problem, because you do want to have, um, interoperability, openness and, uh, different systems will have, um, MCP endpoints just like we have the API endpoints for connectivity. But now you have much more agency, much more control, much more agility and action ability, uh, through the agent to agent communication. So I think he, first of all, I think it's important to know the embarrassment of riches comes because it would become very easy to create agents, and that doesn't mean you have high quality agents that are certifi certifiable and good sort of proxies of your digital workers.

So there's a big difference between creating an agent and actually creating a relevant, reliable, good, useful, productive, ramped up agent. And I think that's one point. And the second point you're bringing up is, uh, even when you have the right agents for different tasks and different areas, just like kind of, you know, digital workers in different departments and across different companies, then I think the, the collaboration and communication protocols that are emerging right now, to be fair, I think we have the model context protocol that, uh, anthropic announced and, you know, pretty much, uh, all the other major players are beginning to adopt it.

We also have the agent to agent, um, you know, um, uh, communication. I think that interoperability would evolve and, uh, would help orchestrate the agents, uh, in a better way. Uh, but I think the fundamental part is still going to be, be your core value is still the data that you have, the clean, actionable information about either your customers, about your business, about your employees, and how do you actually use that to deliver value.

I think that's how, uh, the, the communication is one thing and I think will evolve that communication. But still, uh, in that communication, the unique value that you bring as your agent or agents is still going to be very important. Hmm.

How will the role of the people who bring the professional services and the customers they engage, how will that evolve? Where, how do you envision people being involved in this? Or is it all just gonna be agent agent?

No, no. I do think, I mean, I, I personally, and I think as a company, we do believe in, uh, uh, agent and human, uh, collaboration in many ways, peer to peer in many ways. You know, uh, agents working for humans, in fact, in some cases humans working, you know, for an agent.

But I think it would be a collaboration, uh, uh, uh, type of situation. But, uh, I do think the evolution of this, uh, is in service of accelerating a lot of charter that we have defined as human beings for the organization. So for an enterprise use case, it's still is important about vision.

What is the endpoint that you wanna have for your business? What does success look like, uh, for, for you as a customer, how you are creating value? What is the strategy?

What are the choices? What are the, you know, specific steps you are taking there? So the agents actually come in to accelerate, to automate, to, uh, collaborate with the human beings to, um, provide some of the kind of the tasks.

So it could be as part of an employee team, you have, you know, maybe an onboarding agent that actually helps the new employees just get up to speed much more quickly and discover the content that they need to do. Because it may not be as exciting for human HR manager to be spending a lot of time on something that is very tactical. Um, but I think in the same time, you have the, a lot of the, uh, delivery agents that work with your, uh, software engineering team to produce higher quality code that you actually wanna deploy and use.

But that code could actually be really at a 10 x level to solve some way more complex problems. So I think it's a situation where, uh, agents, frankly, our sort of point of view is that it becomes a very much a tool in service of, uh, solving the problems that we wanna solve, uh, as a humanity versus a little bit of a dystopian view of agents taking over. Alright.

Hey folks, you heard it here. We all have been a little obsessed with the mechanics of ai, large language models and how AI are gonna be built, but maybe the time has come to start thinking about well, just how are we gonna use all this stuff? Hey Raju, thanks for being on the show.

Thank you, Mike. All right. And thank you all for watching the latest episode of the Techstrong AI Leadership series.

You can find this episode and others on our website. We invite you to check them all out. Until then, we'll see you next time.