The numbers Brief Episode map Timeline Takeaways Charts Quotes Clips Glossary Go next Transcript

Episode 169 · Enterprise · 41 min

The store that still doesn't know you

Structured retail sits between 12 million kiranas that remember your face and an e-commerce stack that logs every click. RDEP's founder argues it has inherited neither. His bet after thirty years in the trade: the platform's job begins after the transaction, not at it, and most of the missing intelligence isn't AI at all, it's joined dots.

D2C & commerce Consumer India SaaS & enterprise

Pavan Panchamukhi

Founder, RDEP · with Vishal Krishna

41:00

Said in this episode

▶ 6:45

12 million

Kiranas still standing in India

The mom-and-pop store count Vishal cites while explaining the format to non-Indian viewers. The segment with all the personal touch and none of the technology.

▶ 1:56

30 years

Panchamukhi's run inside retail

From setting up distribution channels for CD-ROM titles in 1992 to checkout intelligence today, with several companies in between; he jokes that he has to keep reminding himself he is old.

▶ 25:15

~10

Large enterprise customers

Not a long list by design. RDEP sells B2B to large enterprises, with a few smaller and mid-sized retailers alongside the ten.

▶ 3:28

1 → 50

His first channel team, 1992 to 1996-97

At Kirloskar he went from a single-man army building a retail distribution network for multimedia titles to almost a 50-member team in about four years.

▶ 21:01

100+ visits

Toy-store trips that taught the store nothing

Seven years of shopping for his children produced only birthday-month discounts of 30%, the anecdote that anchors the episode's argument about missing recognition.

The brief

The argument in sixty seconds

Panchamukhi's claim is that Indian retail solved scale and never solved recognition. Sort the market into three, 12 million kiranas with personal touch and no technology, e-commerce with technology and no touch, and the structured chains in the middle that ought to have both. The middle, he argues, has neither, which is why a store in 2026 greets a shopper exactly as it did in 1994. The diagnosis is deliberately unfashionable: the fancy ERPs and CRMs are already installed, the last mile is still missing, and frankly speaking this doesn't even need AI. It needs human intelligence and joined dots. RDEP's answer is a platform that begins where the POS stops. Every point-of-sale system is defined by the transaction; RDEP defines what happens after it: a store assistant prompted with who just walked in and what they abandoned last visit, a smart receipt that becomes a personalised page carrying offers and delivery status, and rules the retailer writes rather than buys, because the SaaS products that fail are the my-way-or-the-highway ones. It is a bet with scar tissue. BluePay's largest Indian customer, Future Group, folded during COVID, the value perception for experience never materialised at home, and the company re-registered in Singapore and won Kmart Australia, where its mobile POS ended up billing shoppers in the parking lot. Ten enterprise customers and a decade later, in scale mode and raising strategically, the wager is that experience has stopped being a differentiator and become the only remaining way anything sells.

Worth your time if you are

Retail CIOs whose ERP works and whose last mile doesn't

Store operators deciding whether to buy self-checkout kiosks

B2B founders selling software into thin-margin industries

Bootstrapped operators raising a first strategic round

Episode map

Where the conversation travels

Every block is a chapter, coloured by what it's about. Click any of it to jump straight to that minute on YouTube.

D2C & commerceConsumer IndiaSaaS & enterpriseAI & machine learningFounder journeyProduct strategy

01Cold open: scale without experience 0:00 Vishal sets the question, whether retail that has scaled enormously in India has evolved at all in customer experience, and Pavan defines RDEP as the bridge between retailer and end customer, already live in Australia, Malaysia, Indonesia and Singapore. 02Thirty years, from CD titles to checkouts 2:08 Asked why retail, he answers that he doesn't know anything else: a developer pushed into sales at Kirloskar in 1992, building a distribution channel for multimedia titles sold on CD (books on discs, ebooks before the word existed) from a one-man army to almost a 50-member team by 1996-97, and walking into Shoppers Stop's first Mumbai store in 1994. 03Three retails, and the hollow middle 4:46 RDEP is a platform rather than a product or a service, and the thesis behind it is a three-way split: mom-and-pop shops with personal touch and no technology, e-commerce with plenty of technology, and the structured retail in between that should have had both and ended up with neither. 04The same store, twenty years later 7:00 Vishal's complaint runs from 12 million surviving kiranas through Kishore Biyani's 2006 instinct for matching oil to rice to masala, to e-commerce that merely converted paper to PDF, and Pavan concedes the point while refusing to blame retailers working on very thin margins. 05COVID, and Kmart's parking-lot checkout 9:00 COVID removed the option of not changing, even as store operations pushed back on what they could absorb; RDEP's mobile POS, customised and integrated for Kmart in Australia, ended up billing customers in the parking lot, the revelation that a checkout need not be a place at all. 06A retail OS that begins after the bill 12:40 RDEP sits on an existing POS or becomes one, integrating CRM, loyalty, supply chain, inventory and warehouse data so the retailer can keep talking at every touch point, including a smart receipt that turns the bill into a personalised page carrying offers, new-store announcements and live delivery status. 07The tough question: why not just prompt it? 14:50 Vishal puts the CTO's objection, that with GenAI and agentic tooling in-house there is no reason to buy a SaaS at all, and Pavan answers that every POS is defined by the transaction while RDEP defines what happens after it, turning a cashier into a salesperson who already knows what you left in the cart last time. 08We build the platform, you write the rules 17:10 RDEP's AI and agentic modules deliberately refuse to set the retailer's rules, since the SaaS products that fail are the my-way-or-the-highway ones, so a marketing team defines its own algorithms, schemes and reports instead of asking the vendor for an eleventh report. 09The toy store that still doesn't know you 19:10 A recommendation engine and Bluetooth-level recognition set against seven years of toy-store visits that produced only birthday-month discounts. Proof, Pavan argues, that this doesn't even need AI, just joined dots, while fancy ERPs and CRMs leave the last mile where it was and UPI counts as convenience, not upgrade. 10Future Group folds; Singapore opens 24:00 The value perception for experience never arrived in India, and when COVID took down BluePay's largest customer the team registered in Singapore and won Kmart Australia, after which India sold itself, with Levi's India tracking them down without a decent website and Kingdom of White growing from one store to more than 35. 11Licences, hybrid cloud, and the audit 27:00 Six to eight months into a scale phase and adamant that it does not want money to scale and die, RDEP sells licences, runs a hybrid of its own private cloud, customer private clouds and public cloud with on-prem only where Kmart required it, and is taking ISO 27001 because every enterprise customer audits before awarding a contract. 12Instagram, privacy, and owning your data 29:20 Social commerce pushes more products rather than better-matched ones, but Pavan credits its maturing recommendation engines, calls surrendered privacy the price of personalisation, and says RDEP, which could be its own Shopify, wants to move from the offline store into online so retailers own their data. 13Experience is now existential 32:10 Retailers have understood that nothing sells without experience and that people no longer buy simply because they need something, and that self-checkout kiosks are failing in India and merely tolerated abroad for want of staff, which is why RDEP would rather put the kiosk on the shopper's own phone. 14Quick commerce noise, one smart receipt 35:10 A quick-commerce pilot exposes the absurdity of a retailer paying to send an SMS and an email at every delivery step, when the smart receipt already issued at checkout can carry the status changes itself, the pitch RDEP now takes to D2C players. 15Thirty years, cats and 'Getting Better' 36:40 Entrepreneurship's unreported half: the mid-level engineers with two or three years' experience who stayed through the worst phase and are still there a decade on, the heavy personal cost, the cats that do the unwinding, a lawyer son who summarises books like a generative AI, and the Beatles line the episode closes on.

Timeline

The arc, briefly

Thirty years at the Indian till, and the recognition the structured middle never built.

1992A developer pushed into sales. At Kirloskar he set up a distribution channel for multimedia titles sold on CD, books on discs before anybody said ebook, working the field as a one-man army.

1994Shoppers Stop opens its first Mumbai store. He visited it that year and met Nagesh. People came in for the parking and the discounts, treated it as a family outing, and went back to the kirana for daily needs.

1996-97One man becomes fifty. The channel he had started alone was close to a 50-member team in about four years.

2006Kishore Biyani changes the game. Before that there was no experience to speak of. Biyani started matching products, so oil pulled rice and rice pulled masala. India still has 12 million kiranas, the segment with all the personal touch and none of the technology.

2020COVID takes the largest customer and opens a market. Future Group, BluePay's biggest account, folded, and the value perception for experience had never arrived in India anyway. The team registered in Singapore and won Kmart in Australia, where the mobile POS they customised ended up billing shoppers in the car park so nobody had to queue. They stopped selling into India after that, and India came back inbound: Levi's India tracked them down when they had neither a decent website nor an email address.

2026Same store, same experience. Vishal's complaint is that a shopper walks into the store he first saw in 1994 and gets exactly what he got then, in an era of smartphones and continuous communication. RDEP now sells to about ten large enterprise customers across Australia, Malaysia, Indonesia and Singapore, is profitable, is taking ISO 27001 because every enterprise audits before awarding a contract, and is raising for scale. The formulation is his: they do not want money to scale and die, they want money to scale and remain.

Takeaways

Ideas to carry out of this hour

The structured middle inherited neither touch nor tech

Panchamukhi's map of retail has three players: the mom-and-pop store, which has personal touch and no technology; e-commerce, which has a whole lot of technology; and structured retail in between, which sits next to both and has taken nothing from either. That is his explanation for why the chains scaled without ever converting scale into engagement, and why a shopper's experience in a big-box store in 2026 is indistinguishable from 1994.

▶ 5:00 ▶ 5:17 ▶ 5:45

The software is installed; the last mile is missing

It is not that technology never reached these retailers. They have the fancy ERPs and the fancy CRMs and plenty besides. The last mile is exactly where it was, because the only upgrade anybody prioritises is the point of sale itself, and in India the visible upgrade is paying by UPI instead of cash. That, Vishal notes and Pavan agrees, is a convenience, not an upgrade.

▶ 22:15 ▶ 22:30 ▶ 22:38

Most of this doesn't need AI, it needs joined dots

Confronted with a toy store that saw a customer a hundred times in seven years and never learned that his children like RC cars and dinosaurs, Panchamukhi refuses the fashionable answer. Let the AI be there, he says, but if the AI itself isn't built properly it won't work, and frankly this doesn't even need AI, it is very basic technology and human intelligence, connecting the dots. Vishal's framing for the same failure: retail solved for scale and removed the judgment.

▶ 21:17 ▶ 21:31 ▶ 21:45

A POS is defined by the transaction; RDEP starts after it

The commoditisation objection is real, since anybody can build a POS in a week, so RDEP deliberately doesn't identify itself with POS systems at all. For every POS, the activity ends at the transaction; RDEP defines what happens post the transaction, sitting on top of an existing till and prompting the person holding it. Take an identifier at the moment of billing, and the store assistant stops being a cashier: they know who walked in, what was abandoned in the cart last time, and what to ask next.

▶ 15:43 ▶ 15:58 ▶ 16:44

Sell the platform, never the rules

SaaS products fail, Panchamukhi argues, when they are my-way-or-the-highway, when the vendor ships ten reports and charges to design the eleventh. RDEP's architecture inverts that: the AI and agentic modules exist, but the retailer writes the algorithms, defines the schemes and designs the reports, because RDEP understands retail and is emphatically not the retailer. The claim is that adapting to an incumbent stack, rather than replacing it, is the only way a platform survives a thin-margin buyer.

▶ 17:12 ▶ 17:43 ▶ 18:14

Value perception, not product quality, exported the company

BluePay wanted to make it in India and worked with Future Group, then the largest of its customers, until COVID took the group down and the perception that experience is worth paying for still hadn't arrived. The team registered in Singapore, won Kmart in Australia, and describes the contrast as liberating: the customer wanted exactly what they wanted to build. India then came back inbound, with Levi's India searching them out at a time they had neither a decent website nor an email address.

▶ 23:45 ▶ 24:16 ▶ 26:14

Give the kiosk to the customer's phone

Self-checkout kiosks went in because retailers can't staff the desks, and a lot of them are failing: in India because shoppers aren't comfortable, elsewhere because the gap is understood but there is no alternative, so customers are forced to use them. Panchamukhi's counter is that people like using their own devices: put the same scan, cart, pay and receipt flow on the shopper's phone, and show the receipt at the exit.

▶ 33:58 ▶ 34:10 ▶ 34:26

Take money to scale and remain, not to scale and die

RDEP has been in scale mode for six to eight months and says it is ready, but ready is not the same as funded, and the round it wants is a strategic one with people willing to stay for the distance. The formulation is his own and unusually blunt for a founder mid-raise: they do not want money to scale and die, they want money to scale and remain and grow from there.

▶ 26:58 ▶ 27:15 ▶ 39:19

The numbers, drawn

What the episode measures

Every figure below was said on air, with timestamps included and caveats kept.

One store to thirty-five

stores

The only growth series quoted on air: Kingdom of White, an Indian customer, began as a single store and today has 'more than 35'. The second bar is the stated floor, not an exact count.▶ 25:43

Three retails, and the hollow middle

sorted his way, by what each one actually has

Panchamukhi's own three-way split, with 12 million kiranas on the left. He refuses to blame the chains for it: they work on very thin margins, and the only upgrade anyone prioritises is the point of sale itself.▶ 5:17

Where the till stops, and this starts

one transaction, and the two halves either side of it

His answer to the build-versus-buy objection. Anybody can build a point-of-sale system in a week, so RDEP does not identify itself with one at all: it sits on the existing till, takes an identifier at billing, and turns the cashier into someone with a reason to ask a second question.▶ 15:43

Seven years, one discount

what a toy store did with a hundred-plus visits

The anecdote the whole argument rests on. His reading is that a recommendation engine and Bluetooth-level recognition would have closed it, and that neither is exotic: let the AI be there, he says, but this one was never an AI problem.▶ 21:01

Worth keeping

Lines that stay

Frankly speaking, this doesn't even need AI. This is very basic technology — it's human intelligence. All you need to do is connect the dots.

▶ 21:31

For every POS system the activity ends at the transaction. The POS is defined by the transaction. For us, we define post the transaction.

▶ 15:43

They have the fancy ERPs and the fancy CRMs and so many things — but the last mile is missing even today.

▶ 22:15

People no longer go and buy a product because they need it. You have to create the need for it — and that need is going to come from your experience.

▶ 33:25

We do not want money to scale and die. We want the money to scale and remain, and grow from there on.

▶ 27:15

Clips that travel

Short on time? Start here

Retail CIOs rethinking where the till lives

Kmart's parking-lot checkout

The COVID accident that became RDEP's proof point: a mobile POS billing shoppers in a car park, and the queue problem it dissolved.

10:20 → 12:40 · 2 min ▶ Watch clip

CTOs weighing build-versus-buy against GenAI

Why can't the retailer just build this?

The episode's sharpest exchange: anybody can build a POS in a week, so the defence has to be everything that happens after the transaction.

14:50 → 19:10 · 4 min ▶ Watch clip

Anyone who has walked out of a store nobody recognised

The toy store that still doesn't know you

A hundred visits, seven years, birthday-month discounts, and the claim that the fix is joined dots, not artificial intelligence.

19:10 → 24:00 · 5 min ▶ Watch clip

Founders whose home market won't pay for the product

Future Group folds, Singapore opens

Losing the largest customer to COVID, re-registering abroad, and finding that India then came back inbound on reputation alone.

24:00 → 27:00 · 3 min ▶ Watch clip

Store operators about to buy self-checkout hardware

Kill the kiosk, use the shopper's phone

Why kiosks fail in India and are merely tolerated elsewhere, and how the same flow plus a smart receipt replaces both the kiosk and the SMS barrage.

32:10 → 36:40 · 4 min ▶ Watch clip

Glossary

The jargon, unpacked

Kirana

India's neighbourhood mom-and-pop grocery, roughly the off-licence or family store elsewhere; 12 million of them still trade, with all the personal touch and none of the technology.

Structured retail

The organised chains and big-box stores between kiranas and e-commerce. The segment Panchamukhi says has scaled without ever acquiring either the shopkeeper's memory or the internet's data.

POS / mobile POS

The point-of-sale system that rings up a transaction; the mobile version untethers it from a counter, which is how Kmart ended up billing customers in its parking lot.

Smart receipt

RDEP's post-transaction artefact: instead of a static bill, a personalised dynamic page that can carry offers, new-store announcements and live delivery status for that one customer.

Recommendation engine

The module that learns what goes with what and what a given shopper actually buys, so a marketing team can define a scheme (buy this rice, get these vegetables) that reaches the right people.

Agentic AI

AI that executes multi-step workflows rather than answering one prompt; RDEP runs an agentic framework internally but leaves the decision rules for the retailer to define.

ISO 27001

The international information-security management standard RDEP is certifying against, useful across regions, though every enterprise customer still runs its own audit before awarding a contract.

Connections

If this resonated, go here next

EP 26 · Consumer

Full transcript

The whole conversation, searchable

Auto-generated captions, lightly cleaned. Click a timestamp to open that moment on YouTube.