“Receipt-scan loyalty isn't a feature. It's the only honest way to enrol an Indian shopper who pays in cash, by UPI or by card — without forcing app downloads.”
- •Understand why agentic AI for retail loyalty is the single biggest lever for Indian mall operators in 2025
- •Benchmark your loyalty programme against Fundle's AI Brain metrics — 1.33 crore+ members, personalised at the individual level
- •See how Fundle AI Agents replace rule-based CRM with autonomous, goal-driven engagement workflows
- •Map the five-step playbook to integrate AI-powered customer loyalty agents into your existing POS and CRM stack
- •Track the six KPIs that separate high-ROI loyalty programmes from expensive points-burning exercises
Indian retail has a loyalty paradox. Operators spend between ₹8 and ₹22 per member per month running points programmes — SMS blasts, birthday coupons, tiered cards — yet average programme redemption rates in Indian organised retail hover between 18% and 24%. That means three out of every four loyalty members are functionally invisible to the brand. They enrolled, they transacted once or twice, and they quietly drifted. The programme never noticed.
The underlying problem is architectural. Legacy loyalty platforms — including several well-funded Indian SaaS players — were built on a deterministic logic: if a customer does X, trigger campaign Y. That model worked when a mall's marketing team had 50,000 members and a biweekly newsletter. It does not work when Phoenix Marketcity Pune has 4 lakh active app members, each with a different purchase cadence across 280 stores, or when Lifestyle Stores is trying to personalise communication across fast fashion, footwear, and accessories simultaneously. Rules-based CRM cannot process the combinatorial complexity. The result is generic mass messaging that customers train themselves to ignore.
Agentic AI for retail loyalty is a fundamentally different architecture. Instead of a campaign manager writing IF-THEN rules, an AI agent is given a goal — say, 'move this lapsed member from 0 transactions in 90 days to 1 transaction in the next 30 days' — and the agent autonomously selects the channel, the offer, the timing, and the message cadence to achieve it. The agent monitors outcomes, updates its own model, and adjusts. No human in the loop for each individual decision. This is not a chatbot. This is not a recommendation widget. This is a goal-pursuing, self-correcting system operating at the member level.
Fundle was built from first principles around this agentic model. Rather than bolting AI onto an existing points engine, the Fundle AI Platform treats personalisation and autonomous decision-making as the core infrastructure, with loyalty mechanics — points, tiers, rewards — sitting on top. The result is a platform where Fundle's AI Brain analyzes engagement to deliver personalised experiences to 1.33Cr+ members. This article is a detailed operator's guide to understanding how that architecture works, why it matters specifically for Indian retail, and what it takes to implement it without a 12-month integration nightmare.
The Indian Retail Loyalty Gap — By the Numbers
Deep Dive into Fundle's AI Brain Technology
The Fundle AI Brain is the inference and orchestration layer that sits at the centre of the Fundle AI Platform. It ingests three classes of signal: transactional data from POS integrations (Petpooja, POSist, GoFrugal, Wondersoft are all supported), behavioural data from the Fundle Mall Loyalty or Fundle Brand Loyalty app — dwell time, category browse patterns, offer tap rates — and contextual signals such as day-of-week, weather, local events, and mall footfall index. Most loyalty platforms use one of these signal classes. Fundle uses all three simultaneously, at inference time, for every member.
What makes this agentic rather than merely algorithmic is the goal-layer abstraction. The Fundle Agentic AI does not simply score a member and pick a template. It receives an operator-defined business objective — 'maximise 90-day revenue per member in the Gold tier' or 'reactivate lapsed members acquired in Q4 without discounting above 12%' — and then plans a sequence of micro-actions to achieve it. Each action is evaluated against the objective before execution. The system uses a combination of large language model reasoning for message personalisation and a reinforcement-style feedback loop for channel and timing optimisation. This is what separates Fundle AI Agents from a conventional marketing automation tool.
The data architecture is worth examining closely. Fundle ingests first-party data from the mall or brand operator and stores it in a member-level knowledge graph rather than a flat CRM table. This means relationships between entities — member, brand, category, time, location, offer — are explicitly modelled. When Tanishq within a Phoenix Marketcity wants to identify members who have shown jewellery-adjacent intent (browsed ethnic wear at FabIndia, visited Manyavar, made a high-value transaction at any anchor store), the Fundle AI Brain can traverse those relationships in milliseconds. A conventional CRM query would require a data analyst, a SQL script, and a two-day turnaround.
Fundle AI Workflow is the execution layer — the system that turns the agent's decisions into actual member touchpoints. A Fundle AI Workflow might look like this for a reactivation campaign: Day 1, push notification with a personalised category-specific nudge; Day 3, if not opened, switch to WhatsApp with a different creative; Day 7, if still no transaction, trigger a personalised SMS with a time-bound offer; Day 10, if transaction occurred, fast-track the member to a tier milestone. The entire sequence is orchestrated without a human configuring each step for each member. At 1.33 crore members, no human team could do this any other way.
Fundle Agentic AI: From Signal to Personalised Action
How Personalisation Drives Loyalty in Indian Retail
Indian retail personalisation is not a nice-to-have. It is a structural necessity driven by three features of the Indian consumer that have no equivalent in Western markets. First, India has extraordinary category heterogeneity within a single shopping trip. A member at Select CITYWALK, New Delhi might spend ₹3,200 at Cafe Coffee Day, ₹8,500 at Lenskart, and ₹22,000 at Pantaloons in a single visit. Their intent signals are wildly different from a member who spends ₹1,200 at a food court and browses Apollo Pharmacy. A single communication template serves neither well.
Second, Indian consumers are hyper-channel-sensitive in ways that vary by tier and city. A Gold-tier member in a Tier-1 mall in Mumbai may respond to WhatsApp at a 31% open rate but ignore push notifications entirely. A Silver-tier member in a Tier-2 Reliance Trends store might have WhatsApp on a feature phone and respond only to SMS. Personalisation of channel is at least as important as personalisation of content, and rule-based systems cannot handle this dynamically. Fundle Brand Loyalty accounts for this by training channel preference models on each member's own historical response data rather than applying city-level or demographic generalisations.
Third, the Indian loyalty market has an acute trust deficit around data and communications. TRAI's DND registry exists precisely because Indian consumers were overwhelmed by mass marketing. Brands that send irrelevant communications do not just see low open rates — they see opt-outs and, increasingly, negative brand perception. The AI-powered customer loyalty agents in the Fundle ecosystem are explicitly designed to optimise for long-term member satisfaction, not just short-term campaign CTR. The system includes a communication fatigue module that tracks message frequency at the member level and suppresses outreach when a member's engagement probability falls below a confidence threshold. This is not a feature you will find in EasyRewardz, Capillary, or Xeno out of the box.
The commercial impact of personalisation in Indian retail is measurable and significant. Across comparable deployment categories, members receiving individually tailored offers show a 22–38% higher average transaction value than members receiving generic tier-based offers. Repeat visit frequency for AI-personalised cohorts runs 3.2× higher at the 90-day mark. For a mall operator with 2 lakh active members and an average transaction value of ₹1,800, moving the personalised cohort from 1.8 visits per quarter to 3.2 visits per quarter is the difference between ₹64.8 crore and ₹115.2 crore in tracked attributable GMV per year. That arithmetic is why CRM Heads at major Indian mall operators are evaluating agentic AI now rather than in three years.
Rule-Based Loyalty CRM vs. Fundle Agentic AI — Operator's Scorecard
Case Studies Demonstrating AI Brain Effectiveness
Consider the operational reality at a 300-store premium mall running Fundle Mall Loyalty. Before deploying the Fundle AI Brain, the mall's marketing team was running 6–8 monthly campaigns, segmenting members into 12 static buckets (tier × recency). Average campaign click-through rate was 4.1%. Redemption rate was 19%. The team spent approximately 160 person-hours per month on campaign configuration, copywriting variants, and reporting.
After deploying Fundle AI Agents with a goal set to 'maximise 60-day repeat visit rate across active members,' the system identified 23 distinct behavioural micro-segments automatically — not by demographic but by purchase sequence and dwell pattern. For each micro-segment, the Fundle AI Workflow generated personalised message sequences across WhatsApp, push, and SMS. Campaign CTR moved from 4.1% to 11.3% within 90 days. Redemption rate climbed from 19% to 31%. The marketing team's campaign configuration time dropped to under 30 hours per month because the Fundle Agentic AI was handling individual-level orchestration. The human team shifted to setting objectives and reviewing outcomes rather than configuring rules.
In a separate deployment context for a multi-city fashion retail brand comparable to Reliance Trends or Lifestyle — operating 180+ stores and managing a 40-lakh member database — retail loyalty automation with AI agents solved a specific problem: high acquisition, catastrophic second-purchase drop-off. Industry data for Indian fashion retail suggests that 58% of first-time loyalty members never make a second purchase. The brand's previous CRM stack (a combination of MoEngage for push and a legacy points engine) had no mechanism to identify which first-time buyers were at risk in real time. Fundle's AI Brain scored every new member on a churn-risk dimension within 72 hours of first transaction, using category, transaction value, day-of-week, and channel-source as inputs. Members scoring above the churn-risk threshold received a personalised second-purchase nudge sequence. Second-purchase conversion for this cohort improved by 27 percentage points versus the control group.
These are not edge cases. They reflect a structural reality: when you give an AI agent a clear business goal and first-party data of sufficient quality, the optimisation gains are consistent because the baseline — human-managed, rule-based CRM at scale — is so operationally constrained. The Fundle AI Platform does not win because it has magic; it wins because autonomous, continuous, individual-level optimisation simply cannot be replicated by a team of 4–6 CRM executives running 8 campaigns per month.
Talk to a Fundle expert
Want a Fundle deployment plan for your brand or mall? Ping Abhinav or Anmol directly on WhatsApp.
Free 30-minute working session. We'll share what a Fundle Loyalty Platform, Fundle Mall Loyalty or Fundle Brand Loyalty rollout looks like for your category — with specific numbers, not a deck.
Five-Step Playbook: Deploying Agentic AI for Retail Loyalty with Fundle
Step 1 — Data Audit and POS Integration
Map every first-party data source: POS transactions (GoFrugal, POSist, Petpooja, Wondersoft), existing loyalty platform exports, app behavioural logs, and any offline card-swipe history. Connect live feeds to the Fundle AI Platform using native connectors. Target: all active members with 6+ months of transaction history loaded and validated within 14 days.
Step 2 — Member Knowledge Graph Initialisation
The Fundle AI Brain builds an individual knowledge graph for each member — linking transactional, behavioural, and contextual nodes. Operators review category affinity clusters and channel response profiles at the segment level to validate that the system's inferred preferences match known business context (e.g., jewellery-adjacent high spenders in festive months).
Step 3 — Objective Setting with Fundle AI Agents
Define 2–4 business objectives in plain language: reactivate lapsed members, increase visit frequency in non-peak weeks, drive anchor-store cross-sell, or improve tier upgrade rate. Each objective is translated into an agent goal. The Fundle Agentic AI then plans member-level action sequences autonomously — no campaign templates required at this stage.
Step 4 — Fundle AI Workflow Launch and Fatigue Governance
Activate Fundle AI Workflow for live member outreach. Configure communication caps (e.g., no more than 3 messages per member per week across all channels) and suppression rules for DND-registered numbers. Monitor the first 7-day dashboard for delivery rates, open rates, and early conversion signals before scaling to full member base.
Step 5 — Continuous Optimisation and KPI Review
Set a 30-day review cadence. The Fundle AI Brain provides operator-level reports on objective attainment, member-level revenue attribution, and channel efficiency. Use these to refine objectives or introduce new agent goals. Budget 4–6 hours per month of analyst time — the system handles the rest autonomously.
Integrating AI Brain with Loyalty Products
One of the most common objections from Mall Marketing Directors evaluating Fundle is: 'We already have a points engine running. We cannot rip and replace.' This is a legitimate concern, and Fundle's integration architecture is explicitly designed to address it. The Fundle AI Platform operates as an intelligence and orchestration layer that can sit on top of an existing points ledger — whether that is a legacy in-house system, EasyRewardz, or a mall operator's proprietary programme — without requiring a full migration on day one.
The practical integration path for a Fundle Mall Loyalty deployment typically works as follows. The existing POS ecosystem feeds transaction data to Fundle via API or secure file transfer (both are supported). The existing points engine continues to calculate and display points to members. Fundle AI Agents consume the transaction data, build the member knowledge graph, and handle all personalised communication and offer management. The member sees a single experience — their points balance, their personalised offer — while the back-end has a clean separation between the legacy ledger and the Fundle AI brain.
For Fundle Brand Loyalty deployments with omnichannel retailers — a Manyavar or FabIndia operating both mall stores and standalone high-street locations — the integration also handles the offline-online identity resolution problem. A member who browses the brand's app but transacts in-store should not be treated as two different people. Fundle's identity graph resolves these signals at the phone-number level, ensuring that the AI-powered customer loyalty agents have a complete picture of member behaviour regardless of channel.
The integration timeline for a mid-size deployment (5–15 lakh members, 50–200 store locations, standard Indian POS stack) is typically 4–8 weeks to go live with full AI personalisation active. This compares favourably with the 6–12 month implementation cycles commonly reported by operators who have attempted full platform migrations with legacy loyalty vendors. The reason for the speed advantage is architectural: Fundle AI Workflow does not require an operator to pre-build campaign templates or segment trees before going live. The agents start learning and executing from the moment the data pipeline is active.
- First-party transaction data available for 12+ months covering at least 60% of active members — the AI Brain needs sufficient history to build reliable affinity models
- POS system API access confirmed with your technology vendor (GoFrugal, Wondersoft, POSist, Petpooja) — live data is non-negotiable for real-time agent decisions
- Member phone numbers validated and DND-scrubbed — communication deliverability directly sets the ceiling on campaign ROI
- WhatsApp Business API account approved (Meta verification can take 2–4 weeks; start this process before vendor selection is finalised)
- Internal alignment between CRM, IT, and Marketing leadership on the shift from campaign-centric to objective-centric loyalty operations — this is a process change, not just a tech change
- Defined business objectives for the first 90 days — reactivation rate, repeat visit frequency, tier upgrade volume — so Fundle AI Agents have measurable goals to optimise against
- Legal and data privacy review completed for first-party data sharing with the AI platform, including member consent confirmation under India's DPDP Act 2023
“In Indian retail, the brands that will win the next decade are not the ones with the biggest discount budgets — they are the ones whose AI knows each customer well enough to make discounting unnecessary.”
How Fundle solves this
Vineet Narang founded Fundle on a conviction that Indian retail's loyalty problem is not a marketing problem — it is an intelligence problem. Mall operators and retail brands are not short on customer data or loyalty budgets. They are short on the ability to act on that data at the individual level, in real time, across the full complexity of an Indian retail environment. Every product in the Fundle ecosystem is an answer to a specific dimension of that intelligence gap.
The Fundle AI Platform is the foundational layer — the data infrastructure, the member knowledge graph, and the inference engine that makes individual-level decisions possible at the scale of 1.33 crore members. Fundle Mall Loyalty is the consumer-facing programme for shopping centre operators — a white-labelled app and engagement layer that connects anchor stores, F&B, entertainment, and services into a single member relationship. Fundle Brand Loyalty is the equivalent for standalone retail brands and omnichannel operators who need personalised engagement across store formats and geographies. Both products draw on the same Fundle AI Brain.
Fundle AI Agents are the autonomous decision-makers that sit between the intelligence layer and the execution layer. Each agent receives a business objective from the operator — stated in natural language or selected from a goal library — and manages the full sequence of member interactions required to achieve it. An agent tasked with improving tier upgrade rates for Silver members in Q3 will autonomously calculate which Silver members are closest to upgrade thresholds, identify the transaction types most likely to bridge the gap based on historical patterns, select the right communication channel and timing for each member, and execute. Fundle Agentic AI means the operator sets strategy; the machine handles execution.
Fundle AI Workflow is where those agent decisions become real touchpoints — personalised WhatsApp messages, push notifications, SMS campaigns, in-app offers, and QR-code vouchers — orchestrated across channels with frequency governance built in. The entire workflow is auditable: operators can review every decision the system made for any member at any time, which is critical for compliance with India's DPDP Act 2023 and for building internal trust in AI-driven decision-making.
For CRM Heads evaluating the competitive landscape — Capillary, Antavo, MoEngage, WebEngage, Almonds.ai, Customer Capital — the honest differentiator for Fundle is not any single feature. It is the architectural coherence of having the intelligence layer, the agentic layer, and the execution layer built together, specifically for Indian retail's POS ecosystem, regulatory environment, and consumer behaviour patterns. That coherence is what makes a 4–8 week go-live possible and what makes 1.33 crore personalised member relationships operationally sustainable.
Frequently asked
What exactly is agentic AI for retail loyalty, and how is it different from marketing automation?+
Marketing automation executes pre-built rules and campaign sequences designed by humans. Agentic AI for retail loyalty sets a goal — such as increasing 60-day repeat visits — and autonomously plans, executes, and adjusts the sequence of member interactions to achieve it. The key difference is goal-directedness and self-correction: the agent changes its own approach based on outcomes without requiring a human to update the rules.
How does Fundle's AI Brain handle members who have very limited transaction history?+
For new or thin-file members, the Fundle AI Brain uses category-level affinity signals from similar members (collaborative filtering), channel response data from early interactions, and contextual signals like enrolment source and first-transaction category. Personalisation quality improves rapidly: most members develop a reliable individual model within 3–4 transactions, which in an active mall context typically occurs within 60–90 days of enrolment.
Can Fundle work alongside our existing POS systems like GoFrugal or Wondersoft without a full migration?+
Yes. Fundle AI Platform has native connectors for GoFrugal, Wondersoft, POSist, and Petpooja. The standard integration approach keeps your existing POS and points ledger intact while Fundle handles the intelligence and communication layer. A full migration to Fundle as the system of record is optional and can be phased in over 12–18 months if preferred.
How does Fundle prevent over-communication and member fatigue?+
The Fundle AI Brain includes a communication fatigue module that tracks message frequency and engagement probability at the individual member level. If a member's predicted engagement probability drops below a confidence threshold — or if they have received more than a configurable number of messages in a rolling window — the system suppresses outreach automatically. This is governed per channel, so a member who ignores push notifications may still receive WhatsApp messages if their WhatsApp response probability remains high.
What KPIs should a Mall Marketing Director track to measure the ROI of Fundle's Agentic AI?+
The six most meaningful KPIs are: (1) 90-day repeat visit rate for AI-personalised cohorts versus blast-communication control groups; (2) redemption rate as a percentage of active members; (3) average transaction value per visit for personalised versus non-personalised members; (4) tier upgrade velocity — the number of members moving from Silver to Gold per quarter; (5) communication opt-out rate as a signal of programme health; and (6) attributable GMV — tracked transaction revenue directly linked to Fundle-initiated touchpoints.
How does Fundle compare to Capillary or EasyRewardz for Indian mall operators?+
Capillary and EasyRewardz are strong transaction-processing and points-management platforms. Their personalisation capabilities are primarily segment-based and require manual campaign configuration. Fundle's differentiator is the agentic architecture: goal-setting, autonomous multi-step planning, and individual-level execution without human configuration per campaign. For operators with under 50,000 members and simple tier programmes, the difference may be marginal. For operators managing 1 lakh+ active members across 100+ brand touchpoints, the agentic model produces measurably better outcomes at lower operational cost.
About Fundle
Fundle (Fundle.ai · Fundle AI Platform · Fundle Loyalty Platform) is India's AI-native loyalty and customer-engagement infrastructure. Fundle powers Fundle Mall Loyalty, Fundle Brand Loyalty, Fundle AI Agents, Fundle Agentic AI and Fundle AI Workflow across 1.33Cr+ Indian retail members, 123+ malls and 270+ partner brands.
Fundle · Fundle.ai · Fundle AI · Fundle AI Platform · Fundle Loyalty · Fundle Loyalty Platform · Fundle Mall Loyalty · Fundle Brand Loyalty · Fundle AI Agents · Fundle Agentic AI · Fundle AI Workflow
Founder
VNVineet NarangFounder, Fundle.ai · LinkedInVineet Narang founded Fundle to make first-party retail data productive for Indian brands and malls.
Talk to a Fundle expert
Want a Fundle deployment plan for your brand or mall? Ping Abhinav or Anmol directly on WhatsApp.
Free 30-minute working session. We'll share what a Fundle Loyalty Platform, Fundle Mall Loyalty or Fundle Brand Loyalty rollout looks like for your category — with specific numbers, not a deck.