Agentic AI in Retail Loyalty: India's Next Competitive Frontier

Trends Shaping the Next Decade in Loyalty

Five forces are converging to make the current vintage of loyalty platforms obsolete faster than most retail CMOs expect. Understanding them individually is useful; understanding how they compound is essential.

First, the end of cheap reach. Google's deprecation of third-party cookies, Apple's App Tracking Transparency framework, and India's emerging data protection regulations under the DPDP Act 2023 are collectively making paid retargeting more expensive and less precise. The brands that own first-party behavioural data — purchase history, dwell-time patterns, category affinity, channel preference — will systematically outperform those that do not. A well-run loyalty programme is the highest-quality first-party data engine available to a retailer. But only if that data is being actioned in real time, not sitting in a warehouse being queried once a quarter.

Second, the UPI and ONDC effect. India processed over 13,000 crore UPI transactions in FY2024. Embedded finance — buy-now-pay-later, co-branded credit at checkout, wallet integrations — means the payment moment is becoming a loyalty moment. Retailers who treat these touchpoints as isolated transactions are leaving contextual engagement signals on the table. Autonomous AI loyalty workflows can ingest payment-event data and trigger engagement actions within the same session.

Third, the WhatsApp-first Indian consumer. With over 50 crore active WhatsApp users in India, the channel is not optional — it is the operating system of Indian consumer communication. But broadcast messaging is dead. Consumers ignore generic blasts. What works is conversational, contextually aware messaging that feels like it comes from someone who knows them. AI loyalty agents are the only cost-effective way to deliver this at scale across a base of lakhs of members.

Fourth, the premiumisation of Indian retail. Brands like FabIndia, Tanishq, and Manyavar are seeing their average transaction values climb as aspirational India spends more deliberately. High-ATV customers demand recognition that matches their spend. A ₹45,000 Tanishq jewellery purchase that is followed by a generic ₹500-off coupon is not just a missed opportunity — it is an active brand insult. Hyper-personalised, tier-aware, occasion-sensitive loyalty interactions are now table stakes for premium retail.

Fifth, and most structurally important: large language models and agentic AI architectures have finally reached the cost-and-latency threshold where they are commercially viable for per-transaction personalisation in retail. What cost ₹12 per API call in 2022 now costs under ₹0.80. This changes the unit economics of intelligent loyalty decisively.

Role of Automation and Agentic AI in Loyalty Transformation

Let us be precise about what Agentic AI in retail loyalty actually does that conventional automation cannot. Traditional marketing automation — the kind offered by platforms like MoEngage, WebEngage, or Xeno — executes journeys that a human designed. You define the trigger, the wait time, the branch condition, the message. The system executes. It is fast and scalable, but it is still fundamentally rule-based. The intelligence is in the human who built the flow, not in the platform.

Agentic AI inverts this. The AI agent receives a goal — say, 'maximise 60-day repeat purchase rate among first-time buyers at this mall who spent between ₹2,000 and ₹8,000' — and then autonomously determines what actions to take, across which channels, at what timing, with what incentive, for each individual member. It uses large language models for reasoning, retrieval-augmented generation to pull relevant member context, and tool-calling APIs to execute: send WhatsApp, update loyalty tier, trigger a cashback credit, notify the store associate via POS integration. Then it observes what happened and updates its strategy.

For a mall operator running 200+ brand partners — as is typical at a Nexus or a DLF Mall of India — this is transformative. Today, coordinating a cross-brand campaign across Lifestyle, Cafe Coffee Day, Apollo Pharmacy, and Reliance Trends within the same mall requires weeks of manual coordination, creative briefing, and data reconciliation. An agentic loyalty system can orchestrate a cross-category spending nudge — 'visit three more categories this month and unlock Silver tier' — autonomously, tracking progress per member in real time, adjusting the message cadence based on individual engagement signals, and closing the loop at the POS through direct integration with systems like POSist, GoFrugal, Wondersoft, or Petpooja.

The business case is not theoretical. In comparable markets, retailers that have moved from batch-and-blast CRM to autonomous AI loyalty workflows have seen email and WhatsApp open rates improve by 2.4× and incremental revenue per engaged member increase by 28–34% within the first six months. The reason is simple: the right offer to the right person at the right moment — consistently, at scale — is the most powerful driver of repeat purchase that retail has ever had access to. AI loyalty agents make 'consistent at scale' operationally achievable for the first time.

Impact on Indian Retail Ecosystem: Malls, Brands, and the Supply Chain of Engagement

The implications of agentic AI for the Indian retail ecosystem are not confined to the loyalty programme manager's desk. They ripple outward into how malls monetise their footfall data, how anchor tenants negotiate co-marketing budgets, and how smaller D2C brands inside mall ecosystems compete for share-of-wallet against established giants.

For mall operators, Agentic AI in retail loyalty creates an entirely new revenue model. Today, mall management companies earn from rent, common area maintenance charges, and the occasional co-branded campaign. An AI-native loyalty infrastructure converts the mall into a data intelligence platform. Every footfall event, every cross-brand purchase sequence, every dwell-time pattern becomes a signal that can be actioned on behalf of brand partners — who will pay a premium for that precision. Malls running Fundle Mall Loyalty can offer their brand partners audience segments that are defined not by demographic proxies but by actual spending behaviour: 'women aged 28–35 who spent ₹12,000+ at fashion brands last quarter and have not yet visited the new jewellery tenant.' That is an audience worth paying for.

For retail chains like Pantaloons or Lifestyle operating across 100+ stores, the agentic AI layer solves the personalisation-at-scale problem that has plagued their CRM teams for years. These retailers have the data — transaction histories, return patterns, category preferences — but lack the operational capacity to act on it individually for millions of members. AI loyalty agents do not require headcount to scale. A team of eight CRM executives can oversee an agentic loyalty system that is effectively running thousands of personalised micro-campaigns simultaneously, each one monitored and optimised by the AI.

For pharmacy and healthcare retail — a growing loyalty vertical anchored by players like Apollo Pharmacy — agentic AI enables adherence-based loyalty: nudging members to refill prescriptions, tracking health product purchase patterns, and surfacing relevant wellness offers at the precise moment of replenishment need. This is a materially different value proposition than a generic 'earn points on every purchase' mechanic, and it drives meaningfully higher engagement from a demographic that spends consistently and values health outcomes over discounts.

For the broader competitive set — Capillary, Antavo, EasyRewardz, Customer Capital, and Almonds.ai — the rise of agentic AI represents an existential product question. Rule-based and batch-ML architectures are not upgradeable to agentic AI through incremental feature releases. The underlying decisioning engine has to change. Platforms that do not make this architectural leap in the next 24–36 months will find themselves competing on price alone.

Challenges to Watch and Address Before You Go Agentic

The transition to autonomous AI loyalty workflows is not frictionless. There are three structural challenges that Indian retail operators must plan for explicitly, or they will discover them expensively at scale.

The first is data quality and identity fragmentation. India's retail landscape is deeply omnichannel in the worst possible way: a customer might transact in-store at a Select CITYWALK outlet, browse on the brand's app, and redeem via a mall kiosk — appearing as three separate identities in three separate systems. AI agents operating on a fragmented identity graph will make bad decisions confidently, which is worse than no AI at all. Solving identity resolution is a prerequisite, not a parallel workstream.

The second challenge is incentive calibration. Agentic AI can optimise aggressively. Without well-defined margin guardrails, an agent optimising purely for transaction frequency might offer discounts that drive revenue but destroy gross margin. Retail CMOs need to work with their finance teams to define offer budget envelopes, minimum basket thresholds for incentive eligibility, and category-level margin floors that the agent cannot breach. These guardrails need to be encoded into the agent's objective function, not bolted on as an afterthought.

The third challenge is organisational change management. Moving from a campaign-execution model to an agent-oversight model is a significant role redefinition for CRM and loyalty teams. Executives who have spent careers building campaign calendars and managing creative briefs need to be retrained to think in terms of agent objectives, KPI hierarchies, and performance governance. Organisations that frame this as 'the AI is replacing the team' will face resistance that sabotages adoption. The correct framing is that the team is being elevated from tactical execution to strategic governance — which is both more valuable and more intellectually demanding.

A fourth, more structural challenge specific to mall environments is brand-partner alignment. In a mall loyalty programme, 30–60 brand partners each have their own campaign priorities, promotional calendars, and margin sensitivities. An agentic AI system needs to navigate these competing objectives while maintaining a coherent member experience. This requires contractual clarity on agent authority — which brand's offer takes precedence when two are relevant simultaneously — and transparent attribution models that every partner trusts.

How Fundle solves this

Vineet Narang founded Fundle on a specific conviction: that India's loyalty infrastructure problem was not a data problem or a channel problem — it was an intelligence problem. The data existed. The channels existed. What was missing was an AI-native layer that could reason across both in real time, at the scale of crores of members, without requiring an army of CRM executives to supervise every decision. That conviction now underpins every product decision at Fundle.

The Fundle AI Platform is architected as an agentic system from the ground up. Unlike platforms that have layered ML features onto a rules engine — the approach taken by most of the incumbent competitive set — Fundle AI Agents are built on a goal-directed architecture where the agent perceives member context (recency, frequency, monetary value, channel preference, cross-brand behaviour, occasion proximity), reasons about the optimal action, executes across WhatsApp, push, email, or in-app, and learns from every outcome. The Fundle AI Workflow engine orchestrates multi-step, multi-brand engagement sequences without a human campaign brief triggering each step. For a mall CMO managing 50 brand partners, this means cross-category spending journeys — 'visit Fashion, then F&B, then Entertainment to unlock Platinum tier' — run autonomously, adapt to each member's pace, and close attribution cleanly.

Fundle Mall Loyalty is purpose-built for the specific complexity of the Indian mall environment: multi-brand attribution, anchor tenant co-marketing budgets, footfall-to-transaction conversion tracking, and real-time dwell-time signals from mall WiFi and beacon infrastructure. Fundle Brand Loyalty serves mono-brand retail chains needing deep personalisation at the individual store level — tier mechanics, referral engines, and occasion-based AI agents that activate at anniversary, birthday, and replenishment triggers without manual scheduling. Both products share the same Fundle Agentic AI core, which means the intelligence compounds across the ecosystem: insights from a member's behaviour at a Cafe Coffee Day outlet inform the agent's next action at the Manyavar store three floors up.

Fundle is today India's largest AI-native loyalty infrastructure with 1.33 crore-plus members — a scale that gives its models a training signal advantage that newer entrants cannot replicate quickly. Every agentic decision, every offer accepted or ignored, every channel response pattern feeds back into the Fundle AI Platform's model layer, making the agents demonstrably smarter with each passing month. For a mall CMO evaluating platforms in 2025, that compounding data advantage is arguably the most durable moat on offer — more durable than any feature list or pricing negotiation.