Chandan Mahajan Podcast Transcript
Chandan Mahajan joins host Brian Thomas on The Digital Executive Podcast.
Brian Thomas: Welcome to The Digital Executive. Today’s guest is Chandan Mahajan. Chandan Mahajan is the co-founder and chief growth officer at DotKonnekt, an enterprise AI execution company helping retail, consumer manufacturing, and distribution enterprises move AI from ambition to meaningful business impact. Over the last 20-plus years in enterprise technology, Chandan has worked globally and led large transformation programs across commerce, merchandising, customer experience automation, and platform modernization for mid to large enterprises globally.
Somewhere along the way, the work stopped being about technology for its own sake and became about whether any of it actually changes how a business runs. Well, good afternoon, Shandan. Welcome to the show.
Chandan Mahajan: Thanks for having me, Brian. Good to be here.
Brian Thomas: Absolutely, my friend. I appreciate it. You’re generally hailing out of New Jersey.
I’m in Kansas City, and I know you’re doing some travel today, so I appreciate you making the time, being flexible with this. So Shandan, let’s just jump right into your first question. You’ve spent 20-plus years in enterprise technology, leading large transformation programs across commerce, merchandising, customer experience, and platform modernization globally.
You’ve said that somewhere along the way, the work stopped being about technology for its own sake and became about whether any of it actually changes how a business runs. What was the moment or realization that shifted your thinking and led you to co-found DotConnect?
Chandan Mahajan: Sure. So Brian, for the last twenty years that I’ve been in the business, I watched the tech industry work in silos.
I noticed consultants whose job is only to give strategy techs. We came across software vendors who only focus on selling licenses, and then the tech services companies, which are all about how many butts in seats you can fill. And the, the thought was that every one of these players would only care about their part, and this will leave the customer on their own to stitch it all together.
And then in the last few years, we saw the arrival of AI in our world, and I feel it’s like the electricity moment of our lifetime. It is rewriting every industry, every function, every task possible. But also let’s think about what electricity actually did to everyone. The factories that just swapped steam engines for electric motors did not really gain much.
The winners of the electricity era were the companies that redesigned their plants and kind of reimagined how those plants ran. I feel most of the enterprises today are like swapping motors when it comes to AI. And this is kind of, this was the realization that we had before starting DotConnect, where we felt that if the very definition of work is changing, the way we need to deliver work also has to change.
So we decided to break down the walls between consulting software and engineering services and just create one team, one platform, and one outcome for enterprises to win in the AI era. So that’s what’s kind of the, I would say, the genesis moment for DotConnect. We actually started the company– So Dheeraj, who’s my co-founder, he and I started DotConnect in the middle of COVID which is when every single company was sort of rethinking how they will run in the future.
And true to our name we help companies connect the dots between AI, their business and their growth. And over the last two two or so years, as enterprises have really began scaling their en– AI adoption across every imaginable function, we saw the same old silo problem now come back with AI.
Now, what’s happening is that the enterprise’s data, their context, and their know-how is all ending up on a vendor’s platform, right? I, I feel this is like e-enterprises are ri-renting their own intelligence. So we decided– So with DotConnect, we decided to make a promise that the old model could never do, that enterprises will own what we build.
And in a world where platforms are just selling consumption and consultants are selling engagements, we sell ownership
Brian Thomas: That’s amazing. Really appreciate that. I, I took that away right away. You sell ownership, and I think that’s brilliant. You talked about that, last 20 years you and I both been in this business.
Yeah. We see tech industry, they absolutely work in silos. But your really grasp of this right now is AI is really like that period where in- electricity was invented, and it’s really a moment of our lifetime. I, I, I believe that as well. But basically with DotConnect, you said, “Hey, let’s rethink technology.
Let’s rethink AI deployment.” So you thought one team, one platform, one outcome, and I thought that was pretty cool. So thank you. And Chandan, DotConnect is deliberately a vertical AI company-
Chandan Mahajan: Mm-hmm …
Brian Thomas: enterprise AI company focused on retail, consumer, manufacturing, and distribution rather than being horizontal and industry agnostic.
Why did you bet on going deep in specific industries, and what can a retail-trained agent do inside a brand’s ERP that a general purpose AI simply can’t?
Chandan Mahajan: Yeah, we, we get this a lot, especially from investors. See, if you look at what’s happening today, every enterprise has access to the same models, right?
And these models are getting cheaper and better with every release, and it’s pretty much happening every other w- day these days, right? So now you know that the model is no longer the advantage, but your context is. The company’s context is, right? And we feel that models are commoditizing, but the context is what compounds.
And think of context as, like the brain of your business, right? And o- one analogy I like to give is, when you think of hiring a general purpose AI is like hiring a s- very smart new graduate. It, it’ll know, like in the context of retail business, it’ll know what a SKU is. It’ll know a vendor, it’ll know a markdown.
But a vertical AI, which is your domain-specific AI, is like your 20-year tenured merchant in the business. Now, he knows that if you mark down too early, you lose margins, and they also know that if you mark down too late, you’re stuck with unsold inventory. And I think that’s kind of, for me, the, is the, is the big difference in terms of how it plays out between horizontal and industry specific agents and app– AI.
And you kind of talked about ERP as well. So if you look at… And, in the context of ERP, a general AI would know an invoice, and it’ll be able to explain the invoice to you, but a domain-trained agent or AI will notice that the price is not matching with the purchase order. It’ll check the terms and conditions of the vendor contract, and it’ll, it’ll be able to analyze and understand that it needs to send it to the right human for exception approvals, right?
So while the general AI will be able to answer questions, but industry-specific AI agents will actually get the work done. And this is where our team comes in because we have spent over the last twenty years, working very closely with retailers, with consumer goods companies, with manufacturers and distributions and tech companies over the globe.
So we know their world inside out. And that’s why we decided to go deep instead of being very wide. And every engagement that we do in this industry makes the next one better and faster because the domain k- knowledge and the context we have, compounds and carries over. And we have seen the, the test of that or the benefit of that in the adoption that our customers have seen.
So, for example, one of our clients, David’s Bridal- They have now over 100,000 brides use an AI wedding planner that was built with DotConnect. And the idea was to take the stress out of the wedding planning the cycles, right? We have another client which is an industrial services company.
They built an AI safety recommendation assistant which has seen over ninety-six point two percent adoption over a thousand you know, workers in, in the field. And these kind of use cases are scaling because the AI understands their context, and it actually helps improve how these companies operate.
And that’s why we see the adoption, getting much higher than general purpose agents.
Brian Thomas: That’s awesome. Thank you. I appreciate that. You talked about obviously now AI has advanced so far just this past year. Every enterprise have access to those same models, but what makes it different?
So you talked about that, that the model is no longer the advantage. It’s that context, what specifically about that domain or company which I thought was interesting. And you said in that space, right, vertical AI will be more efficient, knowledgeable, predictable because of that industry context it has access to.
And I thought that was interesting. But, but the example with David’s Bridal is, is amazing, and it’s just hard to believe now that AI is really becoming our true personal assistant in really just about any industry. So thank you. Chandan, DotConnect promises to take enterprises from AI ambition to production-grade impact in weeks, not years, using a mix of consulting, your Compass platform, and forward-deployed engineers who embed with client teams.
How does that embedded engineering model actually compress the timeline, and what makes it fundamentally different from how a traditional systems integrator would approach the same problem?
Chandan Mahajan: To, to understand this, let’s let’s kind of understand or let’s start with how incentives work, in the old model.
When you look at traditional system integrators they are paid for effort. So, so the time that they spend on the customer is their revenue. And you have platforms that are paid for consumptions where the usage, how many licenses you have, how much tokens you consume is their revenue.
So in the old model unless these players are breaking down their own business model and their own economic model, they have no incentive to help their clients move faster and become cost efficient. So– And if you look at our case, we are neither incentivized for effort or for consumption. We are paid for outcomes and for handing over the ownership.
And that is the fundamental difference on– in terms of how we have built the company and how we create value for our customers when I talk, talk of ownership. So that’s, that’s how we operate. And talking about embedding engineering our domain experts and our engineers, they sit with the client team work on their systems in their cloud, completely in their private environments.
So, their, their context, their intelligence, their know-how is never at risk. And we don’t do any six-month discoveries or multi-year build cycles. Most of our projects are, are, are very fast. Literally in days and weeks is how we work. And we don’t start any of the projects from zero unlike many other services companies.
We built our platform, we call it Compass Studio. It comes pre-built with the domain context layer of these industries. It comes pre-built with agent orchestration, with pre-built templates across marketing, merchandising, supply chain. So when any of these customers comes to us Not only we un– we mostly would understand how their business operates, we would already have a a pre-built solution that can be very quickly customized and configured for their environment.
And hence, we are able to move much faster compared to the traditional cycles, right? And, and then it’s just about the, the discipline that is followed in execution. We call it the one-one-one model, where we take one week to align with the customers on their approach, on their objectives, then one month to build a working prototype of the solution that we decide on, but using clients’ own data, so that they can actually see it come to life with their own data, with their own workflows and systems.
And then within one quarter, we, we take it to production. Like we build a enterprise-grade, production-grade solution that they can use to launch and scale in their business, right? And in our world, that is what ownership looks like. In terms of, we like to think big, start small, and scale very fast.
In fact, one of the clie- one of the clients we have is a company called Bedding Manufacturing which is a hundred and thirty-six-year-old company that is partnering with us to completely reimagine how their business runs. And their CEO recently went on record saying that the kind of work that typically would take them almost twelve months to get done is now getting done in twelve weeks, and they own the solution, right?
And that’s, that’s kind of how we operate. Where everyone is claiming that there is no lock-in we are putting transfers and ownership in our contract.
Brian Thomas: Thank you. Appreciate that. Just to highlight some things here, obviously we know this we both worked in that consulting world. Incentives in the old model was time equals revenue.
But in your case- you’re not paid on out you’re paid on outcomes, not the time licensing or token usage, which I think is really interesting. Very much incentivizes the customer to work with you. Your platform obviously is adaptable, customizable, and you can come in and, and to talk about that turnaround time, your model, the one-one-one.
With your model, the ownership looks like one week of aligning with the customer, one month to build a working prototype, and really within one quarter to take it to production. And I thought that was amazing because a, a project typically back in the day, as you know, would take anywhere from six months to eighteen months, and, and we just don’t have that time nowadays.
Right. So amazing. Thank you. And Chandan, on the last question of the day, if you could briefly share, you’ve mused about whether we’ll eventually need horizontal AI to manage all the vertically focused AI, and the whole category of agentic AI is evolving fast. So if we look out five years, how do you see enterprises actually running once agents are embedded across merchandising, operations, and customer experience?
And what will separate the companies that truly own their AI from those that got left renting it?
Chandan Mahajan: Yeah. So Brian, I wish I had the crystal ball because I feel five years is an eternity in the AI world. So while it’s very hard to predict how things will really pan out I feel there are some non-negotiables that every company should hold on to.
And see, the first one is how work gets done in this world– in this new world that we’re living in. Where every company will be AI native, but human-owned or should be AI native, but human-owned. You can have the agents run the routine workflows across merchandising, operations, growth, customer experience, and so on and so forth.
But humans should ultimately own the policy, the guardrails how to handle exceptions and and just own all the decisions. And yeah, we will probably need some horizontal AI to manage some of the vertical agents, but the lay– that layer has to belong you know, to the company. They have to own it.
The second part is a concept what we call enterprise AI sovereignty, and which basically means that a company should be able to own, control, evolve and move their AI with no dependence on any cloud, any model or any vendor. And it basically comes down to three things. It means that you should own your technology so that you can make decisions based on what makes sense for your business and not what makes sense for your vendor.
Second, you should own your context so that irrespective of whatever technology stack you use in what function, the know-how of your business, the competitive moat, it always stays with you. And third is the capability, so that as you scale with AI, you truly build your company as an AI native company or as a business, and not always be dependent on a third party to, to do that work.
So that you’re also building that that capability and that muscle in-house within your own teams. And that ownership basically is the one is one that gives the freedom for you to pick the right models, pick the right tools for each job, and, and then switch them when something better comes along and which we– which it will.
And the third part of that is the, the just the economics, right? See, I feel if your AI is always living on someone else’s platform, you talked about renting, right? I mean, if you can’t see it, if you can’t control it renters will, always find their bill going up every time the AI is working.
I mean, we, we– everyone calls it token maxing these days, right? Where basically, tokens or high tokens cannot be seen as productivity because on– when you move on consumption pricing your bill is basically your vendor’s revenue, and they have no reason to bring it down.
So I feel in five years, I don’t think anybody is gonna ask which model are you using, but they will definitely ask who owns your intelligence. So that’s why I feel every company you know they could rent the model, but they must own their intelligence.
Brian Thomas: Thank you. That’s amazing. Appreciate that. And if you don’t mind, Chandan, I’m just highlighting some things I think it’s important here.
W- w- we know in this AI age that five years is, it’s gonna feel like 50 years or more. But you talked about how work gets done in this AI age. Companies need to be AI native and, but human-owned, as you talked about. But at the end of the day, you mentioned about companies needing their AI sovereignty, and the three things in there that I took away was own your own technology, own your own context, and own that capability.
‘Cause at the end of the day, you talked about that vendor pricing and, and consumption is a big deal right now. Tokens are just going crazy, right? But at the end of the day, ownership is key, and having that flexibility of owning all that just makes you able to pivot on a dime when need to when the business landscape changes. So I appreciate that.
Chandan Mahajan: Absolutely.
Brian Thomas: Chandan, it was such a pleasure having you on today, and I look forward to speaking with you real soon.
Chandan Mahajan: Thank you, Brian. Yeah, have a great time.
Brian Thomas: Bye for now.
Chandan Mahajan Podcast Transcript. Listen to the audio on the guest’s Podcast Page.
