Thank you, Nirav. Good evening, everyone, and welcome to Avenues AI Limited earnings call to discuss the Q3 FY26 results.

We have on the call with us today Mr. Vishal Mehta, Chairman and Managing Director; Mr.

Vishwas Patel, Managing Director and Chief Executive Officer; and Mr. Sunil Bhagat, Chief Financial Officer.

Also joining us on the call today is Mr. B.

Ravi, who is advising Avenues AI on corporate and financial strategy as an independent consultant. We must remind you that the discussion on today's call may include certain forward-looking statements and must be therefore viewed in conjunction with the risks that the company faces.

I now request Mr. Vishal Mehta to take us through the company's business outlook and financial highlights, subsequent to which we'll open the floor for Q&A.

Thank you, and over to you, sir.

Thank you, Rajat. Good afternoon, everyone, and thank you for joining us.

Today I'd like to cover three important themes. First, our quarterly performance and how we think through growth and profitability.

Second, I will explain the strategic, structural evolution of Avenues AI into an AI-native transaction infrastructure platform and why this shift materially expands our long-term growth opportunity. Third, I'd like to outline how our integrated ecosystem, which spans across Rediff, Rediff1, CCAvenue, and Phronetic AI, how they all work together as a compounding flywheel, and what investors should expect from us over the next two years.

Let me begin. This quarter represents a genuine inflection point for the company, not just because of strong financial performance, but because the architecture that we have been building over the past several years is now operating at a unified, scalable system.

We delivered one of our strongest quarters to date, with sustained growth in payment volumes, expanded profitability, and accelerating platform integration. More importantly, the strategic pillars we committed to—scale, regulatory depth, and AI integration, as well as our ecosystem design—are now converging into a visible competitive advantage.

During the quarter, we completed our corporate rebranding to Avenues AI Limited. This was not a cosmetic change; it's a structural evolution.

We are no longer a payment gateway company. We're building an AI-native transaction infrastructure platform which is full stack, regulated, data-rich, and compounding.

As you all know, artificial intelligence is no longer an overlay for any system. For us, it is embedded across our routing, fraud detection, risk management, reconciliation, orchestration, customer engagement, and decision automation.

The Avenues AI identity represents what we have already become, an intelligence-driven financial infrastructure company anchored by CCAvenue and increasingly relevant in global markets. I want to share how we see our self-reinforcing flywheel.

Avenues AI operates as a closed-loop ecosystem where consumers will generate intent, businesses will transact and operate, AI orchestrates the decision, payments monetize every interaction, and data will compound into an advantage for us. Every transaction makes the system smarter, and every data point improves automation.

Every improvement strengthens our retention, and this is just not a linear revenue model; this is a learning system. At the center of this flywheel sits Rediff1 and CCAvenue, the operating system of our ecosystem.

Let me talk about our integrated architecture here. On the consumer side, Rediff integrates identity, intent, engagement, and increasingly financial participation.

With a very large registered user base, UPI entry through a financial wellness positioning, and an evolving AI-native content layer, Rediff is becoming a meaningful consumer gateway for us. On the business side, Rediff1 becomes the convergence layer, the AI-native back office for SMEs, enterprises, and institutions.

It integrates communication, workflows, compliances, and operational intelligence into a single environment. At the monetization core sits CCAvenue, the economic engine of our system.

Its role is evolving from payment gateway to an AI-optimized financial execution layer combining compliance, scale, orchestration, and intelligence. Powering all of this is Phronetic AI, the invisible orchestrator in our system.

We think of Phronetic not just as a standalone AI company. It is the brains of our ecosystem.

It converts every data into an action, whether to approve, reject, whether to route or retry, whether to predict or prevent, or whether to notify. It deploys agentic workflows across payments, operations, risk and compliance, and customer communication.

The outcome is somewhat measurable, which is lower customer acquisition cost, higher processing, improved approval rates, reduced operational friction, and over time structural margin expansion. I want to touch upon how the system compounds over time.

Consider a small business operating on Rediff1. It acquires customers using Rediff, it runs operations on Rediff1, it accepts payments via CCAvenue, it uses Phronetic agents to optimize pricing, reduce payment failures, and automate reconciliation.

As that business grows, payment volumes increase. As volume increases, our data intelligence improves.

As intelligence improves, automation increases. We believe as automation increases, margins also expand.

The switching costs move beyond just contractual dependency, and they become operationally existential. This we believe is the power of integration.

Finally, we see a clear two-year structural outcome: consumer scale expanding via Rediff, business lock-in deepening through Rediff1, predictable monetization compounding through CCAvenue, and margin expansion over time through AI. This is not an incremental feature growth; this is a platform-level operating leverage.

We don't want to build a quarterly story; it's a compounding story. We believe very few companies operate simultaneously across consumers, merchants, regulated payment infrastructure, and AI-driven orchestration, and even fewer still have them integrated into a closed-loop learning architecture.

That is the opportunity in front of us. This quarter, for the first time, we are seeing the flywheel move as one system.

I will now hand over the call to Vishwas for the operating details. Vishwas, over to you.

Thank you, Vishal. Good afternoon, everyone.

This quarter has been an execution-driven quarter for our payments and platform businesses. While the strategic direction has been clearly articulated earlier, the focus over the last few months has been on converting that strategy into operating scale, regulatory readiness, and monetizable infrastructure.

Let me walk you through the key business developments. At the core, CCAvenue continues to scale steadily across merchant segments, supported by strong demand from large enterprises, utilities, telecom, travel, hospitality, etc.

We are seeing higher share of credit card and large-ticket transactions, increased participation from enterprise and institutional merchants, consistent onboarding across SMEs and cooperative banking networks. As discussed earlier, we are consciously prioritizing volume-led growth and absolute profitability over headline take rates.

This approach is already delivering stronger cash flows and operating leverage even in a competitive pricing environment. This quarter was a particularly important from a regulatory capability perspective.

We received our in-principle RBI license for issuing prepaid payment instruments, enabling us to launch wallets, stored-value products, and prepaid solutions in the near future under the CCAvenue ecosystem. We secured the IFSCA approval at GIFT City, which will allow us to participate directly in cross-border payments, escrow services, and international merchant acquiring.

We also received the RBI authorization for offline payment aggregation, extending our presence across physical merchant locations through POS-based networks. Importantly, most of these approvals are close to monetization and not long-gestation initiatives.

Our international payment business continue to gain traction. In the UAE and Saudi Arabia, transaction volumes remain strong, supported by enterprise clients and platform integrations.

In Oman, we are now working with multiple leading banks, expanding our role as a technology and payment infrastructure partner. GIFT City will provide us with a strategic base for US-denominated settlements and cross-border flows, which we expect to scale meaningfully over the coming quarters.

We remain confident that the international markets will contribute a growing share of net payment revenue over the medium term. A key focus this quarter was operationalizing our AI-led payment architecture.

With the launch of CCAvenue Commerce AI and paycentral.ai, we are enabling autonomous transactions initiated by verified AI agents and compliance-ready, auditable payment execution. This is not positioned as a consumer feature rollout, but as a foundational infrastructure for enterprises, platforms, and financial institutions that are preparing for AI-native workflows.

We believe that the architecture will evolve gradually, but has the potential to boost a structural revenue stream over time alongside traditional payment processing. On the consumer side, RediffPay will continue to progress through regulatory and technical readiness.

Our approach here is deliberately phased. UPI and wallet capabilities will coexist.

Focus is on financial wellness, low-cost distribution, and brand-led engagement. We are avoiding high-burn acquisition strategies seen elsewhere in the ecosystem.

The advantage we have in our existing user reach and integrated payment infrastructure, which allows us to scale responsibly. To summarize, this quarter reflects a strong execution in our core payment business, the completion of a broad, defensible regulatory stack, the measured progress in AI-driven commerce and consumer platforms, and a continued discipline around scale, profitability, and risk management.

Our BillAvenue platform continues to gain ground, particularly in semi-urban and rural markets. We processed millions of utilities, mobile recharges, and bill payment transactions this quarter, with agent-led adoption driving deeper financial inclusion.

In parallel, ResAvenue is leveraging our AI-based tools to help hotels implement smart pricing and optimized occupancy rates. This quarter, our ResAvenue platform enabled hotel room bookings of 768,107 room nights at an average of 8,534 room nights a day, crossing over ₹6,050 million for our 3,500-plus hotel properties.

This not only strengthens our position in the hospitality sector, but also demonstrates how AI is beginning to create tangible value across our merchant ecosystem. Early pilots in fraud detection, risk scoring, and dynamic pricing are demonstrating how AI can improve merchant efficiency, reduce costs, and unlock incremental revenues.

We remain focused on building trusted, compliant, and intelligent payment infrastructure at scale while selectively investing in platforms that extend our long-term growth runway. With this, now I invite our CFO, Mr.

Sunil Bhagat, to take you through the financial performance for this quarter. Over to you, Sunil bhai.

Thank you, Vishwas sir. Good afternoon to everyone on the call.

This quarter reflects consistent execution across our core payments business, supported by improving scale, stable operating costs, and continued discipline on capital allocation. Our consolidated gross revenue for the quarter stood at ₹2,381 crore, a growth of 122% year-over-year, driven primarily by higher payment volumes, a stronger contribution from enterprise merchants, and continuous expansion across domestic and international markets.

The consolidated net revenue stood at ₹149 crore, a growth of 6% year-over-year, reflecting steady monetization despite a competitive pricing environment and a higher mix of large-ticket transactions. Our focus remains on absolute profitability and cash generation rather than headline take rate metrics.

Adjusted EBITDA for the quarter came in at ₹98 crore, an absolute growth of 25% year-over-year, reflecting improved operating leverages as volume scaled up. The EBITDA as a percentage of net revenue also jumped from 56% to 66% on year-over-year basis.

Our adjusted PAT stood at ₹86 crore, a growth of 59% year-over-year, supported by cost control, stable credit cost, and higher contribution from value-added services. Our PAT margin has also jumped from 39% to 58%.

Even as we continue to invest in technology, regulatory readiness, and platform development, our EBITDA and PAT margins remain healthy, demonstrating the resilience of our underlying business model. As discussed earlier, the quarter saw a higher share of credit card and enterprise-led volumes, which typically carry lower take rates, but significantly higher absolute contribution.

Our strategy is very clear: that is prioritize volume-led customer acquisition and operating cost. We believe this approach creates a more durable earnings profile, particularly as scale continues to build across merchant and institutional segments.

We continue to maintain a strong balance sheet with healthy cash flows from operations and low leverage. This financial flexibility allows us to fund investments in AI and platform infrastructure and selectively invest in growth initiatives such as Rediff without stressing returns on or liquidity.

Our capital allocation remains measured and returns-driven. We have revised our full-year guidance for financial year 2026, as mentioned in slide number 17.

The upward revision in FY26 guidance reflects a materially improved business outlook and stronger-than-anticipated operating momentum. Based on current trends in the transaction volumes, enterprise onboarding, and international expansion, we remain confident in meeting and potentially exceeding our stated financial objectives while continuing to invest prudently in long-term growth engines.

To conclude, this quarter demonstrated that scale and profitability are growing together, margins remain resilient despite mix changes, and our balance sheet strength positions us well for the next phase of growth. We remain focused on execution, efficiency, and reinforcing the financial foundations of the business as we build for the future.

With this, we'll now open the floor for questions.

Hi, team. Thank you for taking my question.

My first question is on PayCentral's opportunity. Sir, I tried to size PayCentral's opportunity by looking at India's retail digital payment, which is somewhere around ₹900 lakh crore.

I tried to narrow it down to PG addressable segments like cards, UPI, net banking, B2B vendor, etc., which gives me a substantial market of around ₹50 lakh crore. Within Infibeam's own ₹8 to ₹10 lakh crore TPV, I am estimating 50% of 50% of it being automatable, and assuming just a 5% adoption rate, I'm getting a figure of ₹30,000 to ₹40,000 crore in terms of market size for PayCentral TPV.

Is this the right way to think about PayCentral addressable market and adoption potential? Over to you, sir.

Thanks. Basically, this is Vishal.

I'll take this question. You're right.

India is the largest real-time payment market, and the next value layer is intelligence on top of the rails. What traditional aggregator would do is somewhat route transactions, settle payments, and manage reconciliations.

This AI payment agent, which is PayCentral, it cuts across all different payment aggregators. It integrates within the MCPs of all different payment aggregators and not just CCAvenue.

It auto-routes transactions for success probability, it predicts the fraud, it disputes, it optimizes all the checkout flows, and it negotiates routing and settlement logic, and automates all this merchant reconciliation and collections. Theoretically, the way to think about it is this payment is actually moving to a decision infrastructure, decision layer.

In some ways, it becomes like an operating system for all money flows, and it's just not a pipe. Now, with that said, the way we think about it internally is that through UPI alone, imagine the number of transactions that India processes, maybe ₹300 lakh crores, whatever that numbers is, and there's a lot more through other mechanism.

In some ways that intelligence layer will command a few bps of that flow, which is enormous. Theoretically, what we have seen is that the global pattern in this is somewhere this AI agents monetization, which is the PayCentral piece, is somewhat that share of that bps that you would perhaps generate through this particular intelligence layers.

For that to work and the TAM to play out, you need to have open rails, which UPI does, which is completely interoperable and API-driven. You need to have massive scale with some margins, because that is what will force this innovation to come through.

Then this whole merchant digitization wave, which already has started, and it's hugely playing out. In our thesis, where PayCentral sits is merchants don't manage payments.

AI agents manage money flows end-to-end, and whoever owns this layer will control the merchant financial workflow, their whole credit insights, the transaction intelligence, and the ecosystem control. That is a far bigger prize than a gateway margin.

For us, it's competing to become that intelligence layer. Of course, that payment agents have to be adapted across.

We believe it's only a matter of time. The adoption in the AI wave is significantly picking up, and that's what we'd like to play out.

To your point, I think the TAM that you're talking about just on the CCAvenue ecosystem that we think about is also pretty significant. The other good thing, just so that is that when you use agents, traditionally you will take a few minutes to close a transaction from intent to closure.

Using agents, you can do it in seconds. Theoretically, when you think about closing a transaction, you will have to look at the product, look at the checkout, figure out how you essentially make that payment, then the payment routing comes back, and then the confirmation comes through.

Generally, what you would see is that, and most companies they optimize on how many less clicks would you need to make as a human to be able to close a transaction. As an agent, you can do it in a few seconds.

It's far more sophisticated and much more controlled, and the determinism also is very very large in that. A lot of protocols are being set out to be able to handle such a thing.

I think to your point, I think the TAM is very large, but it'll also be a function of adoption by many of the merchants. We believe that consumer personal agents are anyways coming up.

I think transferring that agent-to-agent protocol and using PayCentral as a payment agent becomes very fertile, if you come to think of it.

Okay. Thanks for the detailed answer, sir.

Just a follow-up on this. I understand that PayCentral is not just about the transaction, but it's about controlling the entire merchant financial workflow, which you just mentioned.

Again, I know it is initial stage, but I just want to understand how do we monetize it? Is it possible for you to tell the blended take rate which we can get from this opportunity itself, or it's very initial stage and it's very hard to tell this?

Over to you, sir.

No, it's not hard to tell. I will tell you.

Generally, payment gateway as a business is not going to go away. You still require the regulatory approvals and so on and so forth to come in.

When you actually give it to an agent, agent will also start earning bps on transactions. Theoretically, I think one can imagine a very low single-digit bps across all the transaction volume is what the opportunity holds out to become.

We believe that that's significant, if you come to think of it. It'll just add up into the margins and the bps.

Okay. Just to clarify this, low single-digit, or whatever it may be, it's a add-on to the earlier bps rate which we were collecting from standard.

Correct? That is correct.

Absolutely.

Thank you. That answers my question.

Thanks for the detailed one, sir. All the best.

Hi, can you hear me? Hello?

Yeah. I just wanted to understand one thing.

With the offline PA license now in place, you cover both online and physical payment flows. Should we view offline as a meaningful accelerator to your AI flywheel or through this higher transaction density, richer merchant-level database, we'll see something better?

If the question is that is offline going to help your AI strategy, can you kindly repeat your question?

Yeah.

I think, yeah, we have the online PA license, and now we also have the offline PA license. I don't want to use the word just a pipe, because theoretically offline payments is just a pipe that most offline merchants still have that basic pipe of QR scan, payment done, less intelligence in pricing, routing, credit.

Even if you come to think of it, all these collections and automations and cash flow optimization, all of that is somewhat the missing layer. Where the agent comes in in the offline payments is that rather than just the merchant receiving money, the agent can actually handle the context of the transaction, which means that it can decide what option to use as a payment routing, it can maximize the success rate, detect fraud, identify all the unusual transaction sizes, personalize the transaction by triggering certain workflows and somewhat automate collections as well.

We think that if you come to think of it, India has more than 60 million-plus small merchants. They digitize QR and one can create a massive opportunity for a system that sits between the merchant and the payment rail.

It can learn from every offline transaction and improve the future decision automatically. That's the opportunity.

We think that if you come to think of this offline payment market, you can think of any influenced transaction by the agent times the value which is unlocked is the opportunity somewhere. That's far bigger than just the share of just the payment fees.

We think that this whole AI payment agent in some ways turns every offline payment transaction into a decision engine, effectively becoming somewhat a merchant financial operating system rather than just a pipe that moves money. Does that answer your question or?

This does. It does.

Hello. Hi.

Am I audible?

Yes. Hi.

Thank you for the opportunity. One question was regarding Rediff1.

How many merchants have we onboarded on Rediff1, shifted to that platform, and how is the activity in terms of active transactions and adoption? Some color on that would be great.

Sure. Of course, we've publicly disclosed that somewhere Rediff1 is the whole productivity suite of all these enterprise platforms, and they combine communication, commerce, business productivity software suites.

In some ways we target sovereignty by in some ways local data hosting advantages and air-gapping it out so that you don't have to use any cloud system, if you come to think of it. The strategic performance that we think is what you talk about the install base is that there are more than 20,000-plus merchants already on the Rediff1 system.

The number of users of that is magnitude of that number, much much larger. We think that yeah, it's growing fairly well for us, while we've not divulged exact numbers, but the way to think of Rediff1's performance is not just as a standalone, it is essentially just a foundational identity and a productivity layer in our ecosystem.

The strength lies in activating all this existing enterprises and consumer base and deepening the engagement and enabling the entire ecosystem with payments and AI services. For us, Rediff1 is less about the short-term revenue and more about building the structural control layer that somewhat underpins our long-term monetization strategy, if that makes sense.

Sure. Of course, we have almost a 10 million-plus merchant base and I'm sure most of them are mid to small scale as well.

Typically, what's the typical timeline that you might be looking at in terms of onboarding more larger number of customers and what kind of opportunity are you looking at in terms of onboarding more customers?

We think a reasonable number to go after for the next 12 to 18 months is at least high single-digit to double-digit percentage of our base. That's how we look at the opportunity so far.

We believe that given the touchpoints and given that they trust us with so many of their other transactions, I think that's a fair estimate to go after. It'd be cool to have a million merchants using Rediff1.

That

would be great. Just a few bookkeeping if I may.

What led to this guidance upgrade? Is there some one-off revenue or something that we are expecting in a segment?

No, I think for us since payments becomes the intelligence layer and intelligence-driven area, we think that we've deliberately made investments in scale and data density. Payments are increasingly a, in some ways intelligence service and maximizing transaction flow.

If you look at our performance, we've reached the guidance that we had given for this year in FY26. The upgrade is really a combination of our capacity to actually monetize and somewhat we've prioritized revenue growth and also the depth in our ecosystem.

We're positioning the platform to generally significantly give better value. We don't see this as a one-off.

We think that that's how we'll potentially look like. It somewhat reflects in terms of if you look at the last three quarters and our performance, and if you look at the extrapolation of what it means for this quarter, you'd perhaps come down to our guidance numbers.

Sure. Just last one on mine.

We usually used to share transaction processing value for various of our platforms. Just on the CCAvenue's TPV, please, if that's possible.

If you look at the slide number 15, you would see that the value.

Okay. Thank you.

Thank you so much.

Hi, am I audible?

Okay. You mentioned AI reduced payment failures by 9% in Q3.

How much of this improvement is structural and embedded versus a model-specific gain that could plateau?

Go ahead, Vishwas.

Basically, AI is intelligent routing. When we get a particular failure from a particular bank, right, and then we have the ability to route it to multiple other banks that we're working at in the backend.

Intelligent AI-based system that we have implemented which gauges on the success rate going down to as back as just the last 5 minutes, which calculates the best success rates. It can be intelligently routed through this thing.

This is not across just one or two options, but this is across 200-plus options that is there, right? Even it's just the different cards that we have with eight different banks and different systems.

This kind of this thing really helps in boosting success rates. Also, not only on the success rate, a lot of transactions were also being dropped because of the earlier negative databases and frauds and dropship locations and many other things.

Now with AI, we have been able to do a very dynamic and real kind of a thing where we have been able to drop certain transactions where we see a fraud emerging in real time, so that it does not go only for the real transactions. The success rates automatically get improved when these guys are trying through multiple different modes and using different other fraudulent numbers and cards to stop it.

These all helped in boosting real transactions and other things as well. Does this answer?

To answer your question, I'll tell you. Structurally, a lot of gains have already been accounted for in the system.

For example, all this bank uptimes and all of that, that's already there, and I'm sure that it's improving by the day. My expectation is that a lot of what you would see going forward is really this model gains that come through all this, the whole retry logics and the routings that Vishwas talked about in all this, how do you handle all this timeouts and how do you essentially improve your somewhat scoring algorithms.

It's a good question. We haven't quantified that yet, because we've always assumed that structural gains are not going to be the largest contributing factor given that India is to the finest payment rails out there.

Whatever that structural gain would happen, they're somewhat slow to appear. It's very hard to attribute.

It's a good question.

We won't be able to tell what is the direct EBITDA or a net take impact of this reduction?

No, we can. If the question was that what I understood your question to be is that are you able to get a better success rate based on structural changes or the model changes?

What I was communicating was that yes, there is an improvement, we can compute the EBITDA, but to quantify how much of that success rate improvement is from structural changes versus model changes is not an exercise we have undertaken.

Okay.

What I mean by that is that did that come because we have a better bank uptime or some enhancements that maybe the payment rails has gone through or some latency issues in the network that have been identified and cleaned up or whether it was a better authentication flow for the clients. Those are somewhat more structural, but we think it's somewhat model-driven.

That's what I meant. Because what Vishwas said is right, which is we've configured our framework to include, with all these AI and optimizations that we've put in, that it's much smarter routing and retry logics and all this time out handling and other capabilities that improve and all the fraud scoring improvements that we've made that would have contributed to the largest portion of that improvement in success rate.

That's what we've attributed to that improvement.

You state that switching costs become existential, not contractual. What percentage of revenue today comes from customers where payment, compliance, and workflows are all embedded inside Rediff1, making switching operational infeasible?

Come to think of it, Rediff1 has many productivity suites, and if you come to think of it, once you offer more than one service, the switching cost somewhat exponentially goes higher. The moment you are handling a lot more data and data patterns, and then if you're also figuring out how to, in some ways, that cost of actually moving all that data and being able to serve becomes slightly higher.

Once you have a flywheel framework where you're able to offer to a particular client more than one service, we think that and especially in the world of AI where data becomes the moat, and I'm sure one can argue synthetic data can also be produced, but we think data is the moat. If we've been able to utilize that data and offer those services to that merchant, the switching costs exponentially increases, because it won't do everything that any other system.

It may be very good, but it won't be able to provide the same level of efficiency that a system that utilizes data and provides all these AI prompts to be able to offer to you with all the learnings.

Great. Okay.

I'll fall back in line for the next questions. Thank you.

Hi, am I audible?

Hello, am I audible now?

Yeah. Hi, you've added multiple layers like agentic payments, offline PA, PPIs, cross-border, UPI, and AI orchestration.

How do you prevent organization and product complexities from diluting focus before the flywheel economics fully kick in? Hello?

Yeah. The question is that of course, the two brands that we have right now is CCAvenue and Rediff, and the core for us is payments.

If the question is that are you focusing on too many areas, I think these are all related areas. That somewhat completes our flywheel, the one that we have set out to architect.

We think the focus will continue being that. We have orchestration layer using a lot of AI frameworks that can somewhat compound that opportunity for us, we believe.

These are nomenclatures, but at the core, it's, in some ways productivity suites and payments, and that will always continue being our focus.

Okay. There are no set frameworks to prevent the complexities from diluting the focus?

No. If we think that all of these are somewhat naturally interweaved with each other, and I think that in some ways it's convenience.

The focus for us is actually to make the ecosystem flywheel work,

and

these are important components of that.

All right. My second question is what internal guardrails exist on the ROI?

In terms of our investments?

Yeah.

That we've been appropriately conservative all the way from the beginning, and we forward-invest, and we want to know what that investment will end up looking like for us. I think we'll continue with the same philosophy, which is how do we make sure that what we're spending for, that ROI is measured and managed, and we're able to drive it.

See, our core principle, if you ask us, is that our guardrails is not about stopping investments. We just want to make sure that whatever we spend will help us in terms of advantage, a certain platform advantage, if not today, tomorrow.

Our guardrails for ROI is not short-term thinking. We want to be an AI-first payments business.

We want to evaluate what happens in the next 12 months, because a lot of things are changing. We'd want to track what is the margin on a transaction and how do we, much like the success rate and the loss reductions and time reductions and we won't take up an initiative which won't improve our unit economics, if that makes sense.

That's the very short-term thought process, because a lot of things are changing. From a long-term perspective, we have to think about merchant stickiness, pretty much what we talked about in terms of how do we essentially get the switching cost rather than contractual to make it even more existential, and how do we use data as a moat, which is how do we ensure that you are somewhat reducing the amount of effort.

Then the long-term ROI thinking is really about where do we see the non-linear scale. How do we ensure that we are bringing intelligence, because AI is really intelligence, and if we don't have that monetization, and it won't work.

For us, that future monetization is really the strategic ROI test for us. We want to make sure that the ecosystem pull should always increase for us, which means that every new feature, everything that we do should increase the dependency of the merchant or the customer, or increase the engagement that they have with us.

That's been our philosophy of thinking through ROI. I hope this answers your question.

Thank you so much. Yeah.

Hi, sir. Congrats on a good quarter on profitability terms.

Sir, this quarter, we have seen quite a bit of interesting gross revenue growth versus a very muted net revenue growth, and the net take rate is kind of halved from 11 bps to 6 bps. I understand you've been saying since last few quarters that we're chasing absolute profitability and not the NPR.

Net revenue growth of 6% interestingly in this quarter has driven our EBITDA by 25% and profitability at 60%. The question is one, sir, if I even simply difference the net revenue, it's grown by only ₹8 to ₹9 crore versus EBITDA growing at ₹20 ₹20 crore.

One, if you can explain how this quarter has panned out in terms of operating leverage, and even our EBITDA is actually grown much more than the revenue in absolute terms. If you can give us some sense of how is that we could extract more profitability at EBITDA level more than even net revenue.

If you can explain that, because it definitely, in this quarter, focusing on absolute profitability is borne out. What kind of expense management that we have done, or is there any one-off that was there last quarter which is not this quarter, stuff like that, if you can explain that will be helpful.

In that context, sir, pardon me for my knowledge of payment system not being foolproof. I wanted to understand is if and when this UPI payment starts to become chargeable, does it help us, or it kind of affects us at a structural level?

If you can give us some sense of UPI payment being not charged as of now. There are lots of noise, it will become chargeable at some stage, and some payment providers are saying you can't keep on subsidizing these infrastructure.

If you can give us some sense of our business model versus this change if at all that happens, and that will be really helpful. Thanks.

Sure. Vishwas, you want to take the UPI part, and then Sunil and I'll talk about the net take rate and the EBITDA piece.

Yeah.

Look, if and when UPI becomes chargeable, it's going to be beneficial to all the players, including us. Right now, MDR is at zero right now for UPI, right?

Our whole focus area is on all the other options that we have, other 199 options that we have to get the margins and to increase business on that part. Yes, of course, when UPI becomes chargeable, though, even our offline and online both can contribute significant margins if the MDR ever comes in.

I as a chair at PCI also have been working with the Finance Ministry and other things. Hopefully we are hopeful that some kind of MDR should come in, at least for large merchants.

As well as we are hoping that even the money that the Finance Ministry gives out every year to subvent all these UPI transaction also grows significantly, so that can be an additional one. Apart from that, Vishal, to you.

Sure. To your point, two things.

One is net take rate has decreased, but you're saying that EBITDA as a percentage of the slide, it shows up to be a larger number as a percentage. If I understand your question, why is that particular phenomenon, right?

Net revenue

did

not grow that much, but EBITDA has grown much more.

Exactly.

Yeah.

For us, see, listen, as far as the net take rate is concerned, in absolute terms, you're right. The net take rate has gone down.

That's a deliberate investment in scale and data density for us, because payments are increasingly a gateway to higher-value financial intelligence service. That's the way we want to architect the company.

Maximizing all these transaction flows today will strengthen what we look like in the long term, monetization. By prioritizing revenue growth and somewhat the ecosystem depth in terms of usage, we think that positioning the platform to generate significantly higher lifetime value per merchant over time will be the right metrics for us to live with.

It's less a position of our competitiveness in the industry and our efficiency, and more about a strategic decision on getting to that scale and data density. Now, if you look at the EBITDA numbers, EBITDA is basically a percentage of net revenue.

As net revenue has increased from say about ₹130 ₹140 crore to ₹148 crore, that ₹8.5 crore, the EBITDA percentage is a percentage of net revenue. Certain expenses of course have been more efficient.

That we've redeployed quite a bit in terms of how we want to scale and build out this company. It does not apply to us; it'll apply for every company.

There will be hiring freeze on many of the potential areas that can be automated using agents. A lot of manual tasks, reconciliations, onboarding, there will be, and we believe that there is a significant shift happening even today, because the kind of workforce that you would require in a AI-first framework is very different than what you would require in the past.

You will see a significant change in terms of just the workforce mix. While some of the large companies you've been hearing lately in the news that there is a change, we also have a significant opportunity to upgrade and change the kind of opportunity.

The position that we think about is that anything which is manual, which is task-driven, where you just have to keep on doing the same thing over and over again, all those activities will be taken over by task agents, and we're deploying that across all our companies. Second is the orchestration piece where, once that task has been defined and identified and you need to actually create workflows, there'll be orchestration agents.

I think that also is somewhat completely automated going forward. The real trick over here is this decision agents that will perhaps drive decisions, and you will see that over a period of time, you'll see many companies become super efficient and leaner.

We think that that's actually somewhat a margin expansion opportunity, but rather than thinking of it as reducing the workforce and we would rather essentially upgrade the entire workforce and make sure that there's a natural transition to where we are versus where we would want to think of being. You'd see some of that going on in terms of the overall size and scale of the opportunity.

I think that from a take rate perspective, yeah, it's since it's a percentage of the net revenue, that's why you see the percentages go in that direction, if that makes sense.

Sure, sir. Sir, about upgrading the guidance, sir, if we do a required run rate for fourth quarter from the nine month, you take the midpoint of the revised guidance, sir, fourth quarter, we are quite conservative here.

The question is one, I understand fourth quarter is the best one, but even if I take a midpoint of our guidance and subtract the three quarters, then the run rate for even EBITDA and PAT is not even matching the second quarter, which is a normal quarter. The question is one, are we conservative here, and should we optimistically at least assume that we should be hopefully on the higher side of the guidance, or is there some seasonality which we have to keep in mind while we model our projection?

Thanks.

I think midpoint of that guidance range should of course be achievable in our opinion. We have been appropriately conservative.

If we find that we are doing better and larger, which is reasonable to also assume, we'd be investing more in branding, marketing, and some of the other activities. It's reasonable to assume that we'll end up somewhere in the midpoint.

Sure, sir. Thanks a lot.

Yeah.

Thank you all for joining our third quarter earnings call, and we look forward to keeping in touch and looking forward to building out a terrific strategy and a moat going forward using a lot of AI frameworks. We will keep in touch and keep you updated.

Thank you.