Vikul Gupta Podcast Transcript
Vikul Gupta joins host Brian Thomas on The Digital Executive Podcast.
Brian Thomas: Welcome to The Digital Executive. Today’s guest is Vikul Gupta. Chief Technology Officer Vikul Gupta is a visionary technology leader driving innovation in digital engineering and AI transformation at QualityAI. As CTO at Quality AI, he has led the creation of next-generation AI and automation platforms that enhance software delivery and accelerate digital modernization for global clients.
Under his leadership, Quality AI’s next gen center of excellence has grown into a 150-member global innovation hub, delivering cutting-edge solutions across cloud, DevOps, and gen AI. Vikul’s initiatives have earned industry recognition, including the CEO Award 2025 for exceptional impact on company growth and innovation.
Well, good afternoon, Vikul. Welcome to the show.
Vikul Gupta: Thanks, Brian.
Brian Thomas: Thank you, my friend. I appreciate it, you making the time today. You’re hailing out of Raleigh, Durham, North Carolina. I’m in Kansas City, so just an hour difference today. I appreciate that. I know sometimes it’s hard to traverse time zones, so let’s jump into it.
Vikul, you’ve spent over two decades in enterprise technology with leadership roles at HP Software and Cog- Cognizant before becoming the chief technology officer at Quality AI. And along the way, you’ve secured patents in AI-driven quality engineering. What’s that through line across those chapters in your career, and what drew you specifically towards quality engineering as a place to apply AI at scale?
Vikul Gupta: Brian, I’ve I’ve been a hardcore engineer, right, from, from the get-go. Now quality happened by accident to me. So I was, I was with HPE. I was the Global CTO for DevOps, and I was with the engineering division and with also their data center and cloud automation division. And one of the large SI wanted me to come and establish the DevOps practice.
When I joined them, I realized that their DevOps practice was with the quality organization. So, I took that as a challenge. I’ve never done quality before that. And one of the things which, which the head of quality quality organization asked me to do was apply– bring engineering discipline, convert quality assurance to quality engineering.
And that’s what I did. And at that time, one of the key levers which I thought will make that transformation really impactful and quick was AI. And from the day I started there, I, I, I was intrigued by what all is possible. The art of possibilities in quality engineering were just immense. Coming from engineering background, I realized that AI was in some form always there for, for software development, whether it was code completion and things like those.
But when it came to quality, quality as a discipline, there was some bit of automation, but almost zero AI. And then as digital became more and more prominent it was found– Like, we were churning code much faster than we could test it. And AI became not just an option then. AI became a way for us to do quality at speed.
So that’s how it started. And after my spit stint with the large service provider, I was tapped by Quality AI, previously known as Qualitas, to join the organization to do something similar. Quality AI has been in the quality business for almost twenty-eight plus years. Now, what they were looking was to, to bring in the, the next gen aspect of quality, take that journey from quality assurance to quality engineering, and that’s what I’ve been doing within Quality AI for the last five years.
Patents were an outcome of me trying to solve problems, problems like how much time does it take to design tests? How– Then after you design the test, what– how much time does it take to automate them? And after you automate them, how much time does it take to identify why your test case failed? So, as, as we were looking at all those solving those problems, we developed these solutions, and these solutions became patents.
And then once we were comfortable with something very basic, then we started looking into how do we push quality left and quality right, and AI became an important lever. Just to give you an example, pushing it left is rather than focusing on how better to write test cases worrying about their coverage, worrying about their automation, we started worrying about how to prevent defects.
And that’s where we started looking at in quality AI, how do you look at the requirements, identify if they’re ambiguous, if they’re conflicting, if they’re structured or not structured, if they have acceptance criteria as defined, and we used AI for that. That’s one of the meetings you, you, you were referencing, right?
So, thirty-seven percent of defects we find during testing cycle are because of bad requirements. So the ability to, to shift left and prevent those requirements to, to flow through the cycle and lead to defects later on, preventing that is, is, is, is very, very impactful. And same thing on the right-hand side, right?
Like when something fails in production being able to identify the root cause and then find out why this defect happened and how it could have been avoided, using AI for that is, is another lever we are pushing testing to-towards the right
Brian Thomas: Awesome. Thank you, Vikul. I appreciate that. Loved your backstory.
You’re obviously, as you said, a hardcore engineer, started out in your career, and through your tenure that’s what is your focus. But you… Quality kinda happened by accident, as you mentioned, but you embraced it as a challenge. And what I really liked is you, you were really intrigued by the quality and of course leveraging AI, right?
Doing quality at speed. And you talked about preventing those defects, failures, identifying root causes, and I think that’s awesome that we’re trying to make things better by really getting down to the fundamentals. So thank you. I appreciate those insights. Vikul, the company just rebranded from Qualitest to Quality AI, positioning itself as an AI-first quality engineering partner.
You’ve described the shift as moving from an era about intelligence to one about trust. What does that rebrand actually signal about where the business and industry are heading, and why is trust the defining word for this next chapter?
Vikul Gupta: Brian, the reason we rebranded was because it, it, it, we wanted the new name to represent, the new brand to represent we, what we were actually doing.
Yes, we started twenty-eight plus years back, but the world has evolved, applications have evolved, technologies have evolved. In my discussion with a lot of analysts, right, we struggle to find a sector where there is no AI. And with AI, quality has to evolve, too. In a traditional application, when you test, you are testing a deterministic output.
You put seven, you always get fourteen, no matter how you, how, how you do it. But think of a shopping recommendation app You, you– what the shopping recommendation app gives you in the morning versus what it gives you in the evening is all dependent on what all your wife has done, your kids have done, or your family has done, and it’s a non-deterministic output.
How do you test that? So we were, we were helping a lot of companies, a lot of tech companies on, on the– in the Bay Area, a lot of large insurance, financial companies, retail companies, and we were helping them with their AI journey. This rebranding is reflecting our key focus. Our key focus is on two sides.
On the left-hand side, you have the AI life cycle, and on the right, right-hand side you have AI for SDLC. The left-hand side, AI life cycle or AI for engineering, is more focused on customers who fall in the bucket of builders. Think the large search engine, think the large social company. They’re building new models, and as they’re building models, they need consulting services, data services model design services, model assurance services, model monitoring services.
We provide all of that. And on the right-hand side are customers who are mostly consumers. These are companies, customers who are using AI for SDLC, AI for development, AI for smarter testing, AI for operations. We help– We pro– We’re helping customers build new AI-infused apps by using AI. We’re helping customers do smart testing by injecting AI through our smart platform called Qoco for complete software test life cycle optimization using AI.
And then we have customers who are using AI for customer experience you know, customer support. We’re helping customers for that. Now, you talked about trust. I think one of the things we are seeing is quality as a, as a, as a discipline is going to evolve. There are so many reports, so many, so many articles about how a lot of AI dies in the pilot phase.
That’s true. We have an era we are going through which is what is the art of possibility of AI, and that’s where a lot of pilots come in. But when you go and talk to the CIOs and the CEOs, they are saying, like a lot of our customers are saying that we do a lot of pilots, but m-many of… a lot of them don’t make it to production.
And the reason they don’t make it to production is the leadership doesn’t know whether they can trust the AI-infused application, whether they can trust the AI models. They, they don’t have an idea whether there is bias, there is hallucination, there’s toxicity. There are a lot of different things which cause that mistrust.
So the first era was about the art of possibility. I feel the next era is about the trust for taking these pilots to production, and that is where we focus on. We, we focus on providing trust, specifically in the AI side, so that you can operationalize your AI, so that it can go from pilot to production with specific services on both the sides.
On the left-hand side, when I talked about AI life cycle, we provide AI assurance services, whether it is traditional AI, whether it is gen AI, or whether you’re building agentic-based solutions, we can help assure all flavors of AI. And on the right-hand side, where you, you want to use AI, we have capabilities and platforms and services to help customers use AI for dev and QA and so on.
Brian Thomas: Thank you. Really appreciate that. The insights were amazing. You talked about the company rebranding, Vikul. That rebrand you… company, you, you felt that it really needed to reflect the work you were doing in that AI space, and I thought that was brilliant. You talked about the AI lifecycle includes many facets in this SDLC process.
But you’re helping your customers across the SDLC lifecycle, of course. But what I really highlighted and took away here, you talked about that the art of possibility, that’s where it all started. But you said in the future, quality as a discipline will most likely be a service, and I thought that was interesting, so thank you.
Vikul, one of the standout capabilities is using AI to identify test coverage gaps from real production incidents and customer journey heat maps. Heat maps, I’m sorry. Going, as Nelson Hall put it, beyond the common use cases, walk us through how that feedback loop works and why learning from what actually breaks in production is such a powerful way to improve quality.
Vikul Gupta: So, we call this customer insight. Customer journey maps is a way for us to understand what journey the customer went through, and there are various ways to do that before he, the, he or she hit a bug. So they click here, then here, then here, and then they did this, they added this. Being able to understand that process…
And the way we do that is we, we, we, we look at logs, we look at traces, and we build that customer journey map. Once we have the customer journey map, we, we use AI to analyze two things. One, why did, why did we miss this testing scenario? Second- We missed this testing scenario, but is this how the app was designed in the first place?
And we provide this as a feedback to the product engineering team to look at feature enhancement or feature clarification from a product lifecycle standpoint. This, this whole solution, without getting into too much IP details, is a combination of traditional AI and your NLP in general to identify those test coverage gaps and those product feature set requirement gaps.
So that’s at a high level, Brian, what, what we do. We u- we, we, we work with a lot of third-party products. Most of the customers have products we u- if it’s a web-based application, we use Google frameworks. If, if, if it’s a native application, there are the, there’s, there are multiple products out in the market which provide you the, the trace, the logs analysis to– for us to now apply AI on top to create that user journey.
And once we have the cus- user journey, we use AI to compare that with our test scenarios, our test cases, and then that comparison leads to us identifying where the test coverage gap is. And once we know where the test coverage gap is, we, we not only use that to increase the coverage so that next time this scenario doesn’t, doesn’t leak through, but like I said, we also provide it as, as an input.
A, a simple example is, without naming a bank they created a mobile check deposit feature. Now looking at the user requirements, the feature definitions, we created test scenarios and we tested, everything looked fine. But then we, we, we, we found out that for a particular user journey, it was going beyond the typical timeline defined for someone to complete that process.
Once we started looking at that, we found out that the age group of eighteen to twenty-six young adults, when they get their check, they… what they were doing was they deposited the check and automatically went in back to their account details to find out is the balance was reflecting or not.
And we found out that that whole process was not something which which, which the product had thought through. We were not covering it from a scenario perspective, like the timeframe allowed for that, and that became a feature enhancement. Now, when you deposit a check, it automatically shows your account balance reflected, right?
The pending status and when it’ll be when will the deposit be updated, and so on. So that’s a very early example. This was back, I think, in twenty- tw-twenty nineteen, twenty twenty-one or, or so.
Brian Thomas: Thank you. Appreciate that, Vikul. Amazing. I like how you really dive in from a quality perspective and understand what that customer, customer journey is like, their experience along the way.
And so you all develop that customer journey by… help them by re- obviously re- digging into the logs of traces, et cetera. And you ask those tough questions about those gaps and misses and foundationally, what was missed early on in the, the business requirements, et cetera, of, of, of an app build.
And of course, the thing I highlighted here was that test coverage gap that you really focus on to improve that customer experience and that customer journey. So thank you. And Vikul, the last question of the day, as agentic AI, autonomous systems, and AI-generated code all accelerate together, where do you see quality engineering heading over the next five years?
And what will separate the enterprises that can deploy AI with confidence from those that can’t?
Vikul Gupta: To be honest, Brian, I, I love this question. I’ve been part of a lot of conferences, AI for SAP and others. And a lot of time people think with, with AI coming in, now just because AI is being used for test creation and test automation and test execution, this and so, like, the, the quality as a discipline is going to disappear.
And in, in a lot of my discussions with, with, with industry experts and analysts, I, I feel that it’s just the opposite. Code is being generated, generated even much faster. Faster to an extent that people are questioning whether agile as a methodology needs to be updated to be relevant. There are no more two-week sprints.
Like, people what used– if they, they used to take five days, now it is getting done in probably one hour. So, code is getting generated, and the, the code which is generated is not just human-generated, it is also machine-generated. That’s one problem. O-outside of it being machine-generated, it has AI infused it, so the deterministic, non-deterministic aspect is important.
And because a lot of this code has been generating, generated mostly for digitization of traditional sectors, so user interaction, direct to user is happening more and more in traditional sectors. With, with new features, with new capabilities, with AI big thing, the possible– and with machine writing the code, the possibility of something going wrong becomes much higher and much faster, and it impacting your core business because of customer sat and production issues become even higher.
Now, my belief is quality is going to become even more prominent, more critical because of that. The role of quality, the scope of quality will also evolve I already talked about shift left, shift right. I already talked about the AI assurance aspect of it. Traditional tasks in the QA world or quality engineering world of, of automating test cases using an army of people are going to probably become more and more prominent.
It’s going to be AI-led. But then it’ll also open up newer areas, newer tasks, tasks like hallucination checks, bias tests identifying coverage gaps auto-triaging. These will become more and more prominent. Initially we discussed about where we are from an era perspective. I think quality will become that AI trust gate.
I, I use that term because we believe in that. Without quality playing that critical role of, of being the guardian of that AI trust gate, pilots will die their normal death. They won’t move on to production. So QA function is that trust gatekeeper for moving AI pilots to AI in production The other aspect is, unlike traditional application, with faster releases, with AI being infused in, I think it’ll force the concept of continuous assurance, continuous app assurance, continuous AI assurance.
‘Cause now AI will learn with what is happening in production, so there is chance of drift, model degradation, bias seeping in. So you can’t just test it once and, and, and leave it like that. You’ll have to continuously monitor and provide that continuous AI assurance so that a system that works at launch and has the possibility of degrading over time doesn’t degrade with this continuous AI assurance.
It’s no longer going to be just a… QA is no longer going to be a moment in the development phase, but it’s going to be a continuous process over the life cycle of that tool or that product.
Brian Thomas: Thank you. Appreciate that, Vikul. Just to highlight some things here, you talked about, Yeah, I know people think AI is gonna replace everybody, some people even think humans, at some point.
But really, what this is gonna do is really proliferate a lot of code out there. You talked about that. But you believe quality as a discipline will actually be in high demand and more prominent. The big thing I took away here, obviously, we need guardians of the AI trust gate streamlining that whole process with quality, of course moving the AI pilots to AI production and ultimately, at the end of the day my big takeaway is continuous AI and quality assurance through the whole process.
So I really appreciate that.
Vikul Gupta: Yeah.
Brian Thomas: And,
Vikul Gupta: Brian, just to add, in one of the conferences, one topic which came up from the audience was, will QA or QE be still QE or will it be named different? What was interesting was, one of the option which was floated by panelist in, in that panel was, instead of it being called quality engineering, it might be renamed to trust engineering, with that focus on AI trust.
So just a thought, but I thought it was interesting it was an interesting concept, so I thought I’ll just mention that here.
Brian Thomas: That’s an interesting concept, for sure. A lot of stuff you hear at conferences, things are t- constantly emerging, and that’s why we focus on talking to people like you on the podcast.
So thank you. Vikul, it was such a pleasure having you on today, and I look forward to speaking with you real soon.
Vikul Gupta: Absolutely, Brian. And thanks, thanks a lot for… I think those were great questions, great insights we shared. I love this opportunity, and I look forward to working with you.
Brian Thomas: Bye for now.
Vikul Gupta Podcast Transcript. Listen to the audio on the guest’s Podcast Page.
