AI Deployments Need Less Speed, More Velocity, with Raju Malhotra

Speaker A: By the way, it is becoming relatively easy for pretty much anyone to create AI agents, right? Which sort of, uh, also presents a risk of agent, uh, slop. We talk about AI slop, but potentially agent slop. And the, uh, agent slob can be more risky because once you give them access to your enterprise data and your Slack and your email, and then it can cause havoc very, very fast. Right. So, stepping back, I think, um, the way we think about the elements that make an agent a really good agent versus not so good agent, there are five elements, say five layers of a cake that I kind of want to share with you, uh, if that's okay.

Speaker B: This is SaaS scaled, the podcast where data meets action with host Arman Shrakhi. Each week Arman will be sitting down with CEOs and industry leaders from the technology sector, giving you the insight to innovate without reinventing the wheel. They'll discuss challenges, best practices, and how to identify the right metrics. So if you want to get to market faster and in a way that matters, then subscribe and join us every week as we discuss SaaS scale. This episode is brought to you by Curve, A the modern no code analytics solution. The tools you need to take action with your data on a platform built for maximum scalability, security and cost efficiencies. If you're ready to reduce complexity and dramatically lower costs, then contact us today@crave.com that's QR V E Y.com. Foreign

Speaker C: welcome to another episode of SAS Scaled. Uh, this time around, uh, we are going to have a discussion, uh, again with a lot of topics around AI, uh, with regard to what SaaS and AI have in common and what they can, how they can complement each other. Uh, I have a special guest, uh, Raiju with me and he will tell you about himself and the company. And welcome to this podcast.

Speaker A: Raj, great to be here. Uh, Arman,

Speaker C: uh, can you tell us a little bit about your background, maybe some stories, how did you end up with this company? And I know that you have worked with so many great companies in the past, so you have a very rich history of, you know, knowing software industry very well. We are very excited to really hear, uh, about your insights and vision in different topics you discussed today. But first, let's start about the company you are with right now and what you guys do.

Speaker A: Yeah, great. Um, uh, thanks for having me. Thanks for having, ah, ah, us on this podcast. Um, I am chief product and Technology Officer at Certinia, um, Certinia, uh, works with professional, uh, services organizations, uh, to provide Them um, AI native, uh software as a service. So think of the professional services uh companies in couple of sort of ways. One we work with companies that are pure play professional services like IBM, PwC, KPMG, et cetera. Their whole business is really delivering uh, change, innovation for their customers. But we also work with uh, uh companies where professional services is part of their value proposition. So companies like you know, Google, Cisco, hpe, Salesforce, um, where there are professional services organizations that actually bring their you know, technology solutions in this case to market. In both of those cases, um, we help our customers sell, deliver and manage those services. And what I mean is you know selling services as you can imagine is unique because you can only sell if you have the right skill set, right individuals. So we help them sell those, we help them manage those you know, projects and programs over a long period of time across geographies, across you know, tens of thousands to hundreds of thousands um of uh individuals and then we help them deliver them. Uh so it can be tied back to the pnl.

Speaker C: Um so ah, in order for you guys to provide such a platform and it's a full cycle and taking clients and you know managing all of these processes. Uh what is the history behind it? When you guys started, is it the same thing that the company started with or it started from m a different kind of software and then evolved or did you guys go through any merger acquisitions to complete the solution or everything is built by you guys inside the company.

Speaker A: So Certinia has been around for 17 years. Um and uh, Certinia was actually one of the very first companies as native ERP on the Salesforce platform. In fact it used to be called uh Financial Force. Um but over time um, we've actually kind of really transformed the company. Really thought about where we can help our customers in the best way. And over the years the value proposition of helping the professional services organizations really end to end not just on the financial management but also on the professional services management, resource management, project management and also more recently as uh customer success management. So that evolution uh really has been happening. Uh some of this is organic, most of this is organic. Uh some of this is also through uh, uh some acquisitions. But really over time um, the company changed enough that we uh, just a couple of years ago uh changed the name to Certinia. And the whole idea is that in this world of uh, so much risk and uncertainty and turbulence really uh, we help uh our customers be more certain by again helping them sell in the right way, sell in the certain way, deliver in the certain way and also manage what they have sold in a certain way. So that's really, I think uh, just a quick history. And of course there's a ton of change that has always been going on but this is the time where the change, the pace of change has accelerated. Uh, with AI, Absolutely.

Speaker C: And then you mentioned AI and the impact that it has. What is the impact of AI directly in what you guys do? Is it about really changing processes in your customer side that you have to really now consider it or is it something, some kind of not many changes on their side, but it's more, your software can get smarter and add uh, more intelligence by leveraging AI or maybe both, I don't know.

Speaker A: Oh it's actually everything. So it reminds me of that movie, you know, a few years ago there was an Oscar nominee or Oscar winning movie, Everything everywhere, all at once. I mean think about uh, you know, professional services organizations traditionally has been very human heavy, consultant heavy businesses, right? Because the whole uh, value prop is to deliver bespoke tailor made solutions that work for a given customer. And that's why their customers are willing to pay a premium for that personalized, uh, specialized sort of service. So with AI, our customers are probably changing the most compared to any other enterprise segment. So professional services is right in front of our eyes going through a major transformation, disruption. Right. So that's really the market uh, that we are operating in for our customers. So as a result of that we have been on a journey of transformation, innovation and disruption ourselves. And that has actually really helped us uh, accelerate the level of you know, the products and the services and the AI agents and the you know, different types of uh, uh, capabilities that we deliver to our customers who are professional services organizations as I mentioned. So they can actually not only adapt to this change but really lead this change.

Speaker C: And of course you know, you guys are a SaaS company and then your clients are professional services or maybe not necessarily professional services as the whole. Maybe as you said, Google has a professional services department and still counts as a customer probably or some other companies like that. So when you are a SaaS company yourself, you are serving professional services. These are two sectors that have been very much, I, uh, would say impacted positively, negatively in any way you can think of, but very highly impacted by AI. Right? So AI is changing them now. Many of them may emerge stronger as a result of embracing AI and on both sides, on the SaaS part, on the professional services part. But how do you see the impact on SaaS? How do you think AI, uh is going to change SaaS, uh, based on your experience now you look at the sector that is professional services, you live in the SaaS sector, so you have a very unique kind of position. And based on uh, the experience that you have, uh, definitely that's the insight, uh, we would love to gain from you.

Speaker A: By the way, this is really critical question that you're talking about. So we are an ISV and we work with the SIS. So that intersection, SaaS, professional services, really, if there's the eye of the storm, right, we feel ourselves to be right in the middle of it. Um, so if I step back, even before we think about SaaS, even think about software, package software, professional services, really it comes down to customers pay for outcomes, they Pay for results. SaaS, software, different sort of means. Now AI, those are solutions that provide the right outcomes. So to your question about what is changing in the world of SaaS, absolutely. I think the quote unquote old way of thinking about software as a service, which is based on some workflow around a system of record delivered in a cloud infrastructure, fundamentally a lot of those pieces are there, but many other pieces have actually changed and in fact transformed the, with a AI, uh, sort of, you know, focus. So what I mean by this is what we see the world going towards is a combination of capabilities that are either consumed by human users. So you need the workflow. You may be able to click through certain things and be able to accelerate your, you know, delivery, uh, of certain, you know, services that you're trying to do. But now everything we do has also to be ready and interoperable with agents. So there are situations where, um, we are opening up, uh, not only capabilities for humans to work with agents, but also agents that we provide to work with other agents. So I think that, I'm not quite sure if there is a right word for it, other than the fact it's really SaaS with uh, the agentic capabilities. But really the idea behind this is we over the years have really, uh, taken time to understand what problems our customers are solving, have been solving and what they are going to be solving. And we are very closely connected to that. So we can translate that into what are the best ways for us to deliver those outcomes. Right. And that's where how we deliver those outcomes really comes down to assuming a world which is fundamentally a hybrid world. It's no longer a delivery by humans, it is a delivery by humans and digital workers or agents. And that requires a very different way to think about product strategy, thinking about what type of solutions we deliver. And that's really what we are focused on. Resource management is different, project management is different, selling is different. All of those things are different in this hybrid world. So clearly it's not same as SaaS, but it's actually a evolution from, you know, what we know to really where uh, our customers demand us to be.

Speaker C: Yeah, I mean there are some examples there that for example you may think of price. Dynamic pricing is one of those examples that some companies use the algorithm, the software to really say this is the dynamic pricing for these particular products in the market. And there were some companies increasingly using AI, not necessarily the software, but AI, uh, to really determine the pricing and learn from that process and be more autonomous. Of course each one has pros, each approach has pros and cons in the product that you offer to these companies. Do you think about some of the capabilities that maybe you used algorithms in the past and the software and now you think that maybe in the future or maybe right now you may want to try to really offer it through AI, uh rather than through software?

Speaker A: Yeah. So it's very interesting for a couple of reasons. One, we have always used uh, some type of AI. We used to use predictive AI, machine learning, uh, heuristics, uh, which is by fundamentally is very deterministic. And over last few years we have started using generative AI, agentic AI, which is a combination of probabilistic, uh, and deterministic with the reasoning layer. Um, so first I think, uh, interesting part is that absolutely we think about helping our customers price the projects not only in a dynamic way, but also in a way that actually creates the most value for them and also for their end customers. Right. And that value can be created by adjusting how much of this you're using agentic AI for delivery versus humans for delivery. Which portions are you using? And oh, by the way, which locations, which time zones, which uh, skills that you're using to deliver that. So I think that is you know, perfect use case for how we use uh, AI. In fact we launched uh, uh, our offering for Veda, our offering for AI. We call it Veda V E D A which you know, means ancient uh, wisdom really the intelligence that that is how, that is one way we deliver that type of uh, uh, competitiveness on pricing, resource management, project management. But there is a very interesting other point that I think somewhere below the surface that you actually refer to. We are being very deliberate in where to use what type of AI. So as I mentioned earlier, we are in a business where um, financial certainty, predictability is Very important. So we help our customers pretty much close the books on a daily basis, weekly basis, very frequently. What that means is when you are in an account receivable, account payable type of business, it's not okay to be 99.9% right. You have to be 100% right. So to bring that level of certainty, we combine the determinism of, uh, machine learning to um, our other AI methods, plus the scale and creativity and um, the optionality of the generative and agentic AI with a reasoning layer that actually provides that level of guardrails. So I think the point is very, it's about nuanced solution that we want to offer and we deliberately choose how we actually deliver those solutions as part of our product to the customers. That really balances the need for compliance, certainty regulations which require 100% determinism compared to, you know, if you're summarizing a spreadsheet or summarizing some notes. We are absolutely using a lot of generative AI and agentic AI.

Speaker C: Um, I read something that you wrote that sounds very interesting to me, but I just wanted you to tell us what did, what you meant. I cannot, you know, I can't guess. But definitely you can explain it better. You said AI deployments need less speed, more velocity. Uh, first of all, uh, you know, many people may use their speed and velocity, you know, interchangeably, but in this case you wanted to really just emphasize that what AI projects need is not necessarily a speed to really be deployed, be implemented, maybe, or something like that, but what they need is really velocity. What do you mean by that?

Speaker A: Yeah, so, uh, if I go back to sort of like a high school physics type of thing, uh, speed is a scalar quantity, velocity is a vector quantity, and the difference is direction. This is the time where you and I, your audience listeners, everybody you know, hears this from their teams, their management, their investors and their customers. You're not moving fast enough. Go fast, go fast, go fast. And I'm totally all for understanding that because things are changing so, so rapidly. You know, OpenAI to Anthropic, to Gemini, to everything. There's like, you know, 50 different, you know, changes in just one week. And frankly, many of them are monumental changes. These are not even incremental changes. So when the world is changing so fast, I kind of, you know, the point I was trying to make is it's even more important to know which direction we want to go fast in as much as the need for speed. So I'm not saying that we don't need Speed, of course, we need speed. We need to adapt quickly. But it also, you know, is about a risk that in our effort to go speed, speed, speed, because everybody's expecting us to go speedy, we have to choose the right direction, otherwise we can create chaos. All the random entropy in the system, we can go in the wrong direction. And I think that's where, uh, the point is. It's important actually even more to know what is the vision, what is the strategy, what are the problems that we're trying to solve for our customers? Are, uh, those the problems that customers want us to solve and what outcomes that really matter? I know those are very boring and standard questions and. But to me, I think that is really part of the direction that I meant as a difference between speed and velocity.

Speaker C: Yeah. And it goes back to the point that you raised before, that the outcome matters. And, uh, it matters less, you know, which tool you are using or if you are using AI or you are using not AI. At the end of the day, really, customers care about the outcome if you are solving the problem they have. And the secondary question is, what is that technology behind the scene that you are using? If the outcome is there and it's impactful and helpful, then that's what matters most. And that hasn't changed that philosophy. That kind of, you know, we need to provide the right outcome to the customers that will stay the same. Everything on the surface may change, but nothing underlying is changing from that aspect.

Speaker A: You hit the nail on its head. I think that's exactly right.

Speaker C: And, uh, when we are talking about these AI agents that you mentioned, uh, we are getting closer, closer for these AI agents to be real and, uh, to be productive. Of course, at the very earliest stage that we are right now, uh, you see more of these AI agents be experimental still. I think the trust is a problem that we may not trust them to really, you know, it's like you hire a new employee and during the first few months you may not want it to really for that new employee to really take a lot of actions autonomously without really coming back to you and checking with you before really sign that or change that. So, and I think AI agents is more like that. And you mentioned that from your perspective, AI agents actually there are. They are like employees. They are like, know you are adding more employees. And these AI agents, uh, are, you know, a great one, is like a great employee even, you know, something that can work autonomously, work all the time, work intelligently, be extremely fast. But what are the other aspects? What are the other Attributes that makes you feel like AI agents are similar or like AI, like employees. And what are the aspects that you would see that these are also different from these aspects?

Speaker A: Yeah, so by the way, it is becoming relatively easy for pretty much anyone to create AI agents. Right. Which sort of uh, also presents a risk of agent, uh, slop. We talk about AI slop, but potentially agent slop. And the uh, agent slop can be more risky because once you give them access to uh, your enterprise data and your slack and your email and then it can cause havoc very, very fast. Right. So stepping back, I think um, the way we think about the elements that make an agent a really good agent versus not so good agent, there are five elements, say five layers of a cake that I kind of want to share with you, uh, if that's okay, um, you know, very quickly. The idea by the way, is that our average agent that we ship as part of Veda, uh, is a reflection of your best employee. And just think about it, if average agents can reflect your best employees, then your productivity, your impact, your outcomes would be so much better. But how you get that, I think there are, there are five sort of, you know, layers to that cake. At the foundation, think about this as your data, structured data, which could be your CRM, erp, customer data, sales data, all the kind of your structured data. It also includes unstructured data, the zoom meetings, the teams, the slack, the emails. And if you can actually pull that together into build a context that is a foundation for making those agents be very effective on top of it. The second layer is about telemetry and metadata and that helps the agents to be self learning to create the loop that is more self sustaining. And we believe in um, uh, human in the loop and semi autonomous agents, you know, but they are becoming more capable to more autonomy. But really the first layer provides the structured unstructured data. The second layer provides the telemetry and metadata. And then think about the next layer as the business logic. Every organization has ways of working, you know, this is what it means to be the best consultant for our organization. PwC has, you know, hundreds of thousands of consultants that uh, you know, we support, we, they use our software. But PWC has its own ways of actually, you know, running that. We work with 1400 plus uh, enterprise customers. So we also have figured out what are the best ways to actually sell, deliver, manage work in the professional services. And what we've done over the years is, you know, we have the business logic that we deliver on top of the two layers of Data. Think of the fourth layer as a reasoning layer which is picking up which actions to drive in which uh, particular scenario. And then the top is the user experience. So a lot of times we get hung up on the user experience of the agent, uh, which by the way in the headless world doesn't even have to be a agentic user experience. It could be, you know, wherever you're running your business, you could be in cloud, you could be Gemini, could be OpenAI and we can actually interact that via uh, headless. Because the foundation of structured and m, uh, structured, unstructured and structured data, telemetry and metadata, business logic, reasoning, those four layers is where the magic happens and then that gets expressed wherever the user is. Right? And that's the UI that could be Certinia app, that could be UIs like Slack or uh, other sort of, you know, chatbots, et cetera. But we basically provide that capability as a digital worker type of capability to uh, those customers. Because then the application capabilities come to you, agenti capabilities come to you versus you going and logging into an application. So that really is how we think about agentic um, impact for our customers. And it is far more than being able to spin up an agent because you have to be tethered in the right harness in the right context. And I think that's interestingly for a human, it might take one month, two months, three months to onboard. If we do our setup in the right way, we can spin up uh, the agents. The average agent that is a reflection of your best employee in a matter of minutes, in a matter of second, because you are starting with a repository of uh, so much great context, reasoning, business logic that kind of makes it uh, way more productive and contextual to you.

Speaker C: Amazing. Clearly I see what you uh, say about really if you put these layers in the right order and if you take a systematic approach and do it in the right way, then building these agents can be very uh, fruitful and very beneficial to the organization. It's like, you know, you can add the best employee every day. That doesn't happen in real world.

Speaker A: Right.

Speaker C: So I have a random question for you. Uh, uh, and that is if you really wanted to give me a city that you like, that you have been there and you know, you like this city might be even a small town or something anywhere in the world that you liked it so much, it was relaxing, fun, uh, and you would say, hey Armand, go and just visit this city, this place, this town, this location. What would be that, Sydney?

Speaker A: Um, I would say, I mean There are different sort of cities that come to my mind. Uh, I would say the top of my list would be, uh, Rome. And I know people actually have interesting kind of, you know, sometimes, uh, different views about Rome being crowded, a lot of, uh, other sort of challenges. I'll tell you why I like Rome, particularly in a. Because it has fantastic kind of, you know, uh, places to visit. It's very walkable, a lot of piazza. You just sit down, you know, have a glass of wine, long meals, good conversations. So that's really, you know, makes Rome Rome. I also like the idea that you have, uh, uh, a very good sort of mix, uh, of, uh, modern and the historic. And there's so much kind of, you know, behind the scenes. So. So if I were to pick one city and I would say, you know, Arman, why don't you go visit, or why don't you take me with you? Uh, or kind of my families, I think that'll be my kind of place. Um, that really jumps out for me.

Speaker C: Very, very nice. Very nice. Um, and then, uh, you have a very good taste. And obviously you like Italian food as well. I guess so that helps. I love it. Um, so my last question, uh, for you would be about a book that you may personally or professionally liked it. And you may say, hey, I would like to recommend this to others to read as well.

Speaker A: Yeah. So there is one, uh, book. So few years ago, I kind of switched over from paper books because I used to carry, you know, anytime we move, anytime we go. I would have, like, boxes and boxes of books, and some of them I would, you know, love to kind of highlight and carry around, etc. But then I kind of switched over to Kindle. Then I kind of switched over to audible, so. So I'll tell you, there is one book that I have paper copies. I have Kindle copy, and I have an audible, and I actually go back and refer to it every once in a while. It's not necessarily a professional book, even though I do believe it's one of the best professional books. Uh, it's a book, uh, called the Power of Now. Uh, it's by Eckhart Tolle. And really the gist of that book is the current moment matters. And really that's all we have. It's the now. It's not what's coming tomorrow. It's not what was yesterday. It's not even what was last hour or what's coming up next hour. It's really now. And I think it's about presence. It's about awareness. It's about just recognizing the surroundings. Uh, and, uh, I just find that book fascinating and every time I read it, it's a very short book. I think many folks know this because it's been around for 20, 30 years. Um, it's very, very short book. Um, but that's something that I would say if I were to pick my, the most favorite book, uh, that would be it.

Speaker C: That's fantastic. Thank you very much, Raju for this great discussion. I learned from it and I hope audience have the same feeling. Thank you again for joining us and for your time.

Speaker A: Thank you, Arman for having me and it was great talking to you.

Speaker B: Thank you for listening to SAS Scaled with Arman Eshragi. For show notes and any resources mentioned in today's episode, go to Saskate. If you're enjoying our show, give us a five star review and share on LinkedIn. And be sure to subscribe for any updates on future episodes. Thanks for listening. This episode is brought to you by Curve, A the modern no code analytics solution. The tools you need to take action with your data on a platform built for maximum scalability, security and cost efficiencies. If you're ready to reduce complexity and dramatically lower costs, then contact us today@crave.com that's Q R-V-E-Y.com.