Lenskart is no longer ‘opening’ stores in the traditional sense. It is placing them, with precision, all with the help of AI. And that is one of the reasons why the company is able to deliver a new design within a single day of a customer walking into one of its stores.
At first glance, this may not sound extraordinary in the age of 10-minute deliveries, until you look at what sits underneath it.
For decades, offline retail expansion has relied on instinct disguised as strategy—location scouting, gut calls, brand pull, and some historical data. It worked, but it always carried one constant risk. Demand was never fully visible.
In an interaction with AIM, Peyush Bansal, Co-Founder and CEO of Lenskart, describes how that uncertainty is being steadily reduced. The signal that something deeper is happening shows up in a metric most companies would overlook.
Same Pincode Sales Growth (SPSG), a metric used by Lenskart that measures the aggregate revenue across all stores within a specific geographic area over a set period, is growing at 36%. At the same time, Same Store Sales Growth (SSSG), which measures the year-over-year revenue increase of a store, stands at 28%.
“The gap between the two is something we think about as a health check on densification. SSSG tells you how an individual store is ageing. SPSG tells you what is happening across an entire neighbourhood,” Bansal explains.
The implication is simple: if pincode-level growth is outpacing store-level growth, new stores are not eating into existing ones. They are expanding the market. That is one the hardest things to achieve in retail, and also the easiest thing to get wrong.
What’s Lenskart AI Strategy for This?
What sits behind this is GeoIQ, Lenskart’s proprietary geo-analytics engine. It ingests a wide set of variables like catchment density, income profiles, mobility patterns, competitor presence, and point of interest density. It even incorporates signals from Lenskart’s own digital layer like app engagement, virtual try-ons, wishlists, and store searches mapped to pincodes.
The output is a revenue prediction for any given coordinate. But the real strength lies in the feedback loop.
“As stores open and actual performance data comes in, the model updates. It is designed to be self correcting rather than static,” he says.
This is where Lenskart starts to resemble a system rather than a conventional retailer. Every store is not just a point of sale but a data-generating node that improves the next decision.
A second layer of intelligence operates within the store itself. TangoEye, its computer vision platform, tracks real-time metrics—foot traffic, dwell time, and conversion rates. If a store underperforms its prediction, the system does more than just record the miss. It diagnoses the cause.
This creates a closed loop. Prediction, execution, correction, and then improved prediction.
Lenskart is Expanding a Feedback System
It is a neat framing, but it also hides a shift. When location becomes a calculated output rather than a strategic bet, the role of brand starts to evolve. It still matters, but it operates on top of a system already designed to reduce failure.
“On algorithm versus brand the honest answer is that we do not fully separate them, and we are not sure it is useful to do so. The algorithm places us in the right neighbourhood. The brand ensures people walk in.”
Underneath both sits another layer that often goes unnoticed: hiring. “The entire frontline hiring process—from screening to onboarding—is now AI-enabled,” Bansal says. “At the pace we open stores, that matters. Scaling hiring without scaling the recruitment team is only possible with automation.”
That combination of precise placement, brand pull, and quick hiring is what allows Lenskart to expand without triggering the classic retail problem of cannibalisation.
“The core principle is that GeoIQ models cannibalisation explicitly before a decision is made. It does not just predict a new store’s revenue in isolation; it is predicting the net revenue impact across all nearby locations simultaneously.”
This is where most expansion strategies fail. A new store may appear profitable in isolation but quietly redistributes demand. Lenskart is trying to eliminate that blind spot before the store even opens.
The inputs extend beyond static demographics. Mobility data, real movement patterns, trade area analysis, and dwell-time flows help map how people actually navigate spaces. The idea is not to find a good location. It is to identify untapped movement.
The Biggest Thing is Remote Optometry
The more fundamental unlock comes from remote optometry.
“The traditional model required a qualified optometrist to be physically present in every store, which made profitable operations in smaller catchments very difficult. Remote optometry decouples the specialist from the location,” Bansal explains.
This is not just a technology layer. It removes a structural constraint. A single optometrist can now serve multiple geographies. Suddenly, markets that could not support a full-time specialist can now become viable.
That reflects in the numbers. Remote eye exams have grown 330%. “The growth in remote eye exams was not the result of a single product change, it was the consequence of solving the right underlying constraint.”
The system combines high definition video with AI-guided diagnostic equipment. The optometrist operates remotely while AI monitors patient positioning and reading stability, flagging inconsistencies before measurements are accepted.
Every exam is recorded. Every recording feeds the model. At scale, this achieves two things. It expands access and builds a powerful dataset.
Lenskart conducted 5.5 million eye tests in a single quarter, with 50% coming from first-time users. That is more than a top-line metric. It is the raw material for an expanding intelligence layer.
“The eye test dataset is valuable for vision care, and forms the foundation for understanding the customer across their entire journey,” Bansal added.
Each test captures prescription data, facial geometry, and the start of a behavioural profile. This data then integrates with browsing patterns, purchase history, and lifecycle signals.
What emerges is a unified customer profile that powers highly personalised recommendations.
“When someone uses our virtual try-on, the system already has their face geometry, prescription, and style history. The frame recommendation they receive is tailored to them, not generic to the category.”
The data architecture spans a few clear layers. Facial biometrics from virtual try-ons and in-store scans map face shape, pupillary distance, and bridge width. These inputs determine both fit and aesthetics.
Prescription data tracks current vision, and how it changes over time, which feeds into lens recommendations. Behavioural signals from the app and website capture intent: what frames a user lingers on, what gets wishlisted, and where they drop off in the buying journey.
Then there is lifecycle data, which involves purchase frequency, membership tier, and replacement cycles. It shows where each customer sits in their eyewear journey.
This is where the offline and online loops converge. The store feeds into the same system as the app.
Lenskart is also in early stages of developing AI-based self-testing, where the examination itself is guided by algorithms rather than a remote specialist. The foundation for this lies in years of accumulated eye test data. The goal, eventually, is to make high-quality eye test available to anyone, anywhere, independent of specialist availability.
First-time users play a disproportionate role here. Each new customer introduces fresh data points that recalibrate the baseline. As that cohort grows, the model improves for everyone.
The company also revealed its plans to introduce ‘B by Lenskart’, its entry into “wearable intelligence”. The smart glasses will feature Gemini AI and Sony Camera, as the company disclosed in November last year.
Outside Metros
The final piece of the puzzle sits outside metropolitan markets. There is a common assumption that algorithmic optimisation for profitability excludes low-income or rural markets. Bansal disagrees.
“In the eyewear category the highest unmet need and the largest untapped demand often exist in the same places,” he suggests. The constraint in these markets has not been demand—it has been access. Remote optometry changes that equation. It allows Lenskart to enter smaller towns without compromising unit economics.
First-time buyers in these regions often show higher conversion rates because the problem is being addressed for the first time.
Beyond its core business, the company runs the Drishti: Har Gaon, Har Ghar programme. As part of this, local women, referred to as Women Eye Marshalls, are trained to conduct preliminary screenings with remote support from certified optometrists. This extends access into areas that remain out the reach of its commercial operations.
What ties all of this together is not AI as a feature, but control as a system. Control over where to expand, how demand is measured, control over feedback loops and cost structures.
Retail has always been about managing uncertainty. Lenskart is working to make it predictable.
Your reaction