Fundle.ai: Real-Time Coupon Offers Loyalty Platform: The Mall CMO's Playbook

Challenges of Coupon Implementation in Mall Ecosystems

A mall is not a single retailer. It is a federation of 150 to 400 independent brand stores, food & beverage operators, entertainment anchors, and service providers — each running their own POS, their own promotions calendar, and frequently their own loyalty scheme. Reliance Trends, Cafe Coffee Day, and a regional jewellery chain occupying the same mall floor might share zero data infrastructure. This federation problem is the single biggest reason real-time coupon offers on a loyalty platform fail in the Indian context.

The technical fragmentation is compounded by commercial complexity. Mall management agreements (MMAs) in India typically define revenue-sharing structures, marketing fund contributions, and brand-specific restrictions on promotional depth. A blanket '20% off all brands this weekend' coupon campaign likely violates three MMAs simultaneously. Every coupon offer, therefore, needs to be governed by brand-level rules: which SKU categories are eligible, what the minimum spend threshold is, whether the discount is funded by the brand or the mall marketing fund, and whether the coupon stacks with the brand's own promotions. Without an automated rules engine, this governance happens via WhatsApp threads and spreadsheets — and breaks down the moment campaign velocity increases.

Consumer behaviour in Indian malls adds a third layer of complexity. The average Indian mall visitor touches 3.2 stores per visit and spends 87 minutes on premise, according to RetailQ's 2023 India Mall Shopper Study. But within that visit, the decision to enter a store is made within 8 seconds of passing the entrance. A coupon delivered 20 minutes after the shopper has already left a brand's vicinity is operationally useless. Real-time means sub-60-second delivery from behavioural trigger to push notification — a standard that batch-CRM tools like basic EasyRewardz or standalone WebEngage deployments cannot meet without significant custom engineering.

Finally, there is the coalition loyalty problem. Mall-level loyalty programs — where points earned at Lifestyle can be redeemed at a food court — require a central ledger that reconciles multi-brand transactions in real time. Most Indian mall operators running homegrown or first-generation loyalty stacks (many of which were built on Capillary's early infrastructure or basic Antavo deployments) process transaction reconciliation in batch cycles of 4 to 24 hours. That lag makes real-time coupon triggers — which depend on knowing a customer's current tier, available balance, and recent purchase history — structurally impossible.

Technical Infrastructure Needed for Real-Time Coupons

Building a real-time dynamic coupon engine on top of a fragmented Indian mall ecosystem requires four distinct infrastructure layers, and getting any one of them wrong collapses the entire stack.

The first layer is unified event ingestion. Every POS terminal in the mall — whether it runs on POSist, GoFrugal, or Wondersoft — must emit a transaction event to a central event bus within seconds of payment completion. Similarly, non-transactional events — app opens, location check-ins via BLE beacon or Wi-Fi triangulation, QR scans at brand outlets — must feed the same stream. This is not a trivial integration; Indian malls average 200+ POS terminals across tenants, and each terminal may run a different API version. The event bus must normalise these heterogeneous payloads into a canonical customer event schema before downstream systems can act on them.

The second layer is the real-time customer profile store. A shopper walking into a Lenskart outlet needs to be resolved — from an anonymous device signal or a loyalty card tap — into a known customer profile within 200 milliseconds. That profile must surface her RFM segment, her active coupon wallet, her category purchase history, and any suppression flags (e.g., she redeemed a Lenskart coupon 12 days ago and is therefore outside the eligible re-trigger window). This requires an in-memory profile store — typically Redis or a purpose-built CDP layer — not a traditional SQL database queried on demand.

The third layer is the decisioning engine. Given a customer event and a resolved profile, which coupon should be issued — or should no coupon be issued at all? This is where AI earns its keep. A rules-only decisioning engine can handle 50 to 100 campaign configurations before it becomes unmanageable. A machine-learning decisioning engine can simultaneously evaluate thousands of offer permutations, predict the probability of redemption for each, factor in the brand's margin contribution, and select the optimal offer — all within the latency budget of a push notification delivery (sub-2 seconds end-to-end). The model must also handle explore-exploit trade-offs: occasionally surfacing non-obvious offers to build training signal rather than always defaulting to the highest-confidence prediction.

The fourth layer is the delivery and redemption infrastructure. A coupon issued in real time is worthless if the redemption experience at POS takes 3 minutes of staff intervention. The coupon code must be scannable — a QR or a short numeric code — validated against the same event bus that issued it (preventing double-redemption), and automatically applied to the transaction in the POS without requiring a manager override. This last requirement — POS-level coupon auto-apply — is where most Indian mall coupon programs break down. It requires either deep POS integration or a middleware layer that intercepts the payment flow. Both are achievable; neither is plug-and-play.

Leveraging Fundle's Integrated Retail Media and Loyalty Platform

Fundle powers 3,759+ ad spaces across 123+ malls, driving real-time coupon success. That sentence is worth unpacking operationally, because it describes something no point-solution vendor in the Indian loyalty space — not Capillary, not Almonds.ai, not Customer Capital — currently offers at this scale: a single platform that connects the coupon issuance engine to physical retail media placements inside the mall.

Here is what that means in practice. A shopper enters Phoenix Marketcity Chennai and her Fundle-powered loyalty app registers a beacon event near the food court anchor. The Fundle AI Platform scores her profile — she is a Tier 2 member, last visited 18 days ago, has a strong affinity for casual dining based on her transaction history, and has never redeemed a beverage-category coupon. The Fundle Agentic AI selects a contextual offer: ₹150 off at a Cafe Coffee Day outlet within the food court, funded from the CCD brand marketing budget. Simultaneously, the Fundle AI Workflow triggers a digital display update on the Fundle-managed screen near the food court entrance — showing the shopper's name (if she is opted in for personalised display) or a generic high-relevance creative for her demographic. The coupon lands on her phone within 45 seconds of the beacon trigger.

This physical-digital integration is the architectural moat. Competitors operating purely in the CRM or loyalty software layer cannot orchestrate the retail media placement because they have no contractual or technical relationship with the mall's physical ad inventory. Fundle Mall Loyalty's embedded retail media network means that a coupon offer can be reinforced at the point of physical decision — the storefront, the digital pylon, the food court screen — not just on a mobile screen that may be in a shopper's pocket. The attribution is closed-loop: the same platform that served the coupon tracks whether the shopper walked into the store, transacted, and redeemed.

For brand marketing heads at tenants like Manyavar or FabIndia, the implication is significant. Rather than negotiating separate media buys with the mall's OOH vendor and separate loyalty activations with the mall's CRM team, Fundle Brand Loyalty provides a unified campaign interface where both are configured, budgeted, and reported together. A Manyavar festive campaign can set a ₹5 lakh budget, define the target audience (male, 28–45, Tier 1 spender, not visited Manyavar in 60+ days), and let Fundle AI Agents automatically allocate spend between coupon offers and in-mall display placements based on real-time performance signals.

Feedback Loops and Continuous Campaign Improvement

The most dangerous assumption in dynamic coupon management is that a well-configured system can be left to run autonomously without human oversight. The AI makes the micro-decisions; the marketing team must own the macro-strategy — and that requires structured feedback loops operating at multiple time horizons.

At the daily level, the Fundle AI Platform surfaces a live campaign health dashboard: offers issued in the last 24 hours, real-time redemption rate by brand and category, notification open rates segmented by shopper tier, and any anomaly flags (e.g., a coupon triggering at 3x the expected rate because of a misconfigured suppression rule). A mall loyalty manager should spend 15 minutes each morning on this dashboard. The goal is not to override the AI's individual decisions but to catch configuration errors before they scale into budget bleed.

At the weekly level, the review should focus on offer mix performance. Which categories are generating coupon-influenced incremental spend (i.e., spend that would not have occurred without the coupon trigger)? Which offers are attracting cherry-pickers — shoppers who redeem the discount but show no lift in return visit probability? The distinction matters enormously for budget allocation. A Tanishq jewellery coupon that drives a ₹18,000 incremental transaction is a very different proposition from a coffee chain coupon that attracts the same transactional shopper 12 times a month at a net loss after discount.

At the monthly level, the focus shifts to audience health. Is the loyalty database growing? Is the proportion of members with push notifications enabled increasing? What is the app uninstall rate among members who receive coupon campaigns, and does it correlate with campaign frequency or offer relevance? These metrics tell you whether your coupon program is building a long-term engagement asset or slowly degrading your communication channel through over-communication. Xeno and MoEngage offer some of this reporting in isolation, but the closed-loop attribution — connecting a coupon notification to a physical in-store transaction — requires a platform with both the digital and physical data layer, which is precisely the Fundle AI Platform's design premise.

The quarterly strategic review should benchmark your mall's coupon program performance against the Fundle network aggregate — redemption rates, incremental spend per coupon, repeat visit lift — to understand where you are over- or under-performing relative to comparable properties. This network benchmarking is only possible because Fundle operates across 123+ malls and can anonymise and aggregate performance data at scale.

Compliance and Consumer Privacy in Mall Environments

India's Digital Personal Data Protection Act 2023 is not a distant regulatory concern for mall CMOs — it is an immediate operational constraint that reshapes how real-time coupon offers can be collected, stored, and acted upon. The DPDP Act requires explicit, informed, purpose-specific consent before a data fiduciary can process personal data. For a mall loyalty program, this means that location data used to trigger a contextual coupon — derived from BLE beacon pings, Wi-Fi association, or GPS — requires a separate consent notice that clearly explains: 'We will use your in-mall location to send you personalised offers in real time.' Bundling this into a generic terms-and-conditions checkbox is unlikely to survive regulatory scrutiny.

The practical implication: consent architecture must be built into the loyalty app onboarding flow, not bolted on afterward. A shopper enrolling in the Select CITYWALK loyalty program must be presented with granular permission choices — transaction-based personalisation, location-based triggers, third-party brand data sharing — each individually toggleable. This is both a compliance requirement and a trust-building mechanism. Research consistently shows that shoppers who actively grant location permissions are 2.3x more likely to redeem contextual offers, because they understand and expect the value exchange.

Data minimisation is the second compliance discipline. A real-time coupon engine does not need to store a shopper's precise GPS coordinates — it needs to know she is within 50 metres of a specific brand outlet. Storing raw location streams is both a privacy liability and a data cost. The correct architecture processes location events at the edge (on-device or at the beacon gateway), emits only a zone-level event ('shopper entered Zone C, Floor 2'), and discards the raw signal. The Fundle AI Platform is architected on this principle — events are processed and anonymised before they hit the central data store, and raw location data is never retained.

For multi-brand coalition programs, data sharing between the mall operator and individual brand tenants is an additional compliance vector. If a shopper's Lifestyle transaction data is used to trigger a Cafe Coffee Day coupon, that cross-brand data use requires explicit disclosure and, depending on the data's nature, a data processing agreement between both entities. Mall operators running Fundle Loyalty programs receive standard DPA templates and consent language as part of the platform's compliance toolkit — a detail that matters enormously when a brand's legal team asks 'who owns this data and what can you do with it?'

Finally, the right to erasure under the DPDP Act means that when a loyalty member requests deletion of their data, the deletion must cascade across the event store, the customer profile, the coupon wallet, and any brand-level data shares — ideally within 72 hours. This requires that the loyalty platform maintain a data lineage map: for every piece of data, where it came from, where it went, and how to delete it end-to-end. Platforms built on legacy batch architectures — where data exists in multiple derived tables and export files — struggle enormously with this. Real-time streaming architectures, by contrast, make lineage tracking native.

How Fundle solves this

Fundle was designed from the ground up to solve the specific, compounded problem that Indian mall operators face: fragmented tenant data, physical-digital disconnection, real-time latency requirements, and regulatory complexity — all simultaneously. Vineet Narang's founding thesis was that Indian mall retail needed a platform that treated the mall as a single interconnected retail organism, not a collection of independently managed brand fiefdoms.

The Fundle AI Platform delivers this through four integrated product pillars. Fundle Mall Loyalty provides the coalition loyalty infrastructure — a single points ledger, tiering engine, and member profile store that operates across all tenant brands in a mall. Transaction events from POSist, GoFrugal, Wondersoft, and Petpooja integrations are ingested in real time, member profiles are updated within seconds of each transaction, and coupon eligibility is recalculated continuously rather than in batch cycles. This is the data foundation that makes everything else possible.

Fundle Brand Loyalty gives individual brand tenants — Tanishq, Lenskart, FabIndia, Manyavar, Reliance Trends — their own campaign management interface within the shared infrastructure. Brand marketing teams can configure offer parameters, set budget caps, define target audiences using the shared mall-level behavioural data, and monitor redemption performance in real time — without needing to negotiate data access with the mall operator on a campaign-by-campaign basis. The governance model is contractual and automated: brands see only their own customers' data plus the anonymised audience signals they have been granted access to.

Fundle AI Agents handle the real-time decisioning layer — the sub-2-second evaluation of which coupon to issue, to whom, via which channel, and at what discount depth. These agents are trained on the Fundle network's cross-mall behavioural dataset, which means they arrive at a new mall deployment with prior knowledge about Indian shopper behaviour patterns, seasonal redemption curves, and category affinity clusters — rather than starting from zero. The Fundle AI Workflow orchestrates the end-to-end campaign automation: from trigger event ingestion through decisioning, coupon issuance, push delivery, POS validation, and post-redemption model update — all without human intervention in the execution loop.

The retail media integration — 3,759+ physical ad spaces across 123+ malls — is what elevates Fundle Agentic AI beyond a software platform into a complete marketing operating system for Indian malls. No other platform in the competitive set — not Capillary's loyalty cloud, not Antavo's enterprise tier, not EasyRewardz or Customer Capital — offers the combination of digital loyalty infrastructure and physical retail media activation in a single closed-loop system. For a mall CMO looking to implement real-time coupon offers on a loyalty platform at genuine scale, Fundle is the only architecture that closes the loop between offer issuance and physical shopper behaviour — and proves it with attribution data rather than panel surveys.