“Loyalty is the only marketing function where the customer raises their hand and asks to be remembered. Fundle exists so that no Indian retailer ever wastes that ask.”
- •Understand why rule-based loyalty engines fail modern Indian retail at scale
- •Compare Fundle's AI-native infrastructure against legacy platforms point by point
- •See how Fundle manages 1.33Cr+ members and ₹2,329Cr+ revenue using Agentic AI
- •Audit your current loyalty stack against a seven-point operator readiness checklist
- •Build a five-step migration playbook to Agentic AI loyalty without business disruption
Every mall CMO in India has lived through the same quarterly review: footfall is up, the loyalty programme has 8 lakh enrolled members, and yet same-store sales growth is flat. The loyalty platform vendor shows open rates, redemption percentages, and points-issued graphs. None of it explains why the top 5% of members are responsible for 61% of revenue while the bottom 60% have not transacted in 180 days. The programme is running. The business is not growing. That gap — between loyalty mechanics and loyalty outcomes — is exactly what Agentic AI in retail loyalty is built to close.
The Indian organised retail market crossed ₹18 lakh crore in FY24 and is on track to reach ₹28 lakh crore by FY28, per Technopak estimates. Mall-anchored retail — Phoenix Marketcity, Select CITYWALK, Nexus, DLF Malls — is generating more footfall than the pre-pandemic peak, yet average ticket sizes in apparel and lifestyle have grown less than 4% in real terms over three years. The loyalty spend allocated by most operators — typically 1.2–1.8% of net revenue — is not returning proportional lift because the underlying technology is transactional, not intelligent. Points are issued. Coupons are blasted. Segments are static. Nothing learns.
Legacy platforms — and this includes category incumbents like Capillary, EasyRewardz, and Antavo deployments in India — were architected in an era when batch processing, SQL segmentation, and scheduled SMS campaigns were state of the art. They are campaign managers dressed as loyalty engines. They cannot reason about a customer's next likely action, cannot autonomously negotiate an offer with the brand tenant, and cannot detect churn risk at the individual level without a human analyst configuring a rule. When Tanishq needs to re-engage a customer who bought a ₹95,000 necklace fourteen months ago and whose daughter's wedding season is approaching, a rule-based engine sends a generic 'We miss you' mailer. An AI agent books the customer a private in-store preview appointment.
Fundle was built from the ground up to make that second outcome the default, not the exception. The Fundle AI Platform introduces Agentic AI in retail loyalty — autonomous, goal-directed AI agents that plan, act, learn, and optimise across the entire customer lifecycle without waiting for a campaign manager to press send. This article is a direct, numbers-first comparison of what that means for mall operators and retail chain heads making platform decisions in 2025.
The State of Loyalty in Indian Retail: Four Numbers That Frame the Crisis
Comparing Fundle's AI-Native Infrastructure to Legacy Loyalty Stacks
The architecture debate in loyalty technology is not about features on a pricing slide — it is about what the system can do without a human in the loop. Legacy platforms like Capillary's Loyalty+ or EasyRewardz operate on what engineers call an event-trigger model: a transaction fires a webhook, the webhook hits a rule, the rule issues points or triggers a campaign. The system is deterministic and stateless. It does not remember that this customer returned two items last month, that her preferred category is ethnic wear, and that she has never redeemed a coupon but responds to early-access invitations. It cannot hold that context, reason with it, and then autonomously decide to send a Manyavar pre-launch invite rather than a 15% off voucher.
Fundle's AI-native infrastructure inverts this model. The Fundle Agentic AI layer sits above the transactional data layer and runs persistent, goal-directed agents — one per customer, effectively — that maintain a live contextual memory of each member's behaviour, preferences, lifecycle stage, and predicted future actions. These are not recommendation models running in batch overnight. They are real-time agents that update their understanding of the customer with every data event: a POS transaction at Lifestyle, a QR code scan at a Café Coffee Day kiosk, a cart abandonment on the brand's e-commerce extension, a birthday flag from the CRM. The agent then autonomously selects the next best action — a message, an offer, a service intervention, a tier upgrade prompt — and executes it through the Fundle AI Workflow engine.
On the data integration side, legacy platforms offer APIs that require brands to push data on their own schedule. Fundle's platform maintains pre-built, certified connectors to Petpooja, POSist, GoFrugal, and Wondersoft — the four POS systems that collectively cover an estimated 68% of organised retail and F&B transaction volume in India. This means a mall operator running 120 brand tenants on four different POS systems can see a unified, real-time customer record without a six-month integration project. EasyRewardz and Xeno offer some POS integrations, but they remain largely campaign-scheduling tools on top of that data, not reasoning systems.
The infrastructure difference also shows up in scale economics. Most legacy platforms charge per-campaign or per-communication, creating a perverse incentive to send fewer, blunter messages. Fundle's Agentic AI model is optimised for precision over volume: the system sends fewer total communications but each one is contextually appropriate, resulting in 2.4× higher redemption rates and a 31% reduction in opt-out rates observed across Fundle's active mall deployments in India.
Agentic AI vs. Rule-Based Loyalty: What Changes at Every Layer
Core Features of Agentic AI in Retail Loyalty and Indian Market Fit
Indian retail has structural characteristics that make off-the-shelf Western loyalty tech a poor fit. First, the channel mix is genuinely omnichannel in a way most Western markets are not: a Phoenix Marketcity customer may discover a promotion on WhatsApp, try a product in-store at a Pantaloons outlet, comparison-shop on a quick commerce app, and finally purchase in-store using a BNPL instrument. The loyalty system must track and attribute across all of these touchpoints or the customer record is incomplete and the next best action is wrong. Second, India's language and cultural diversity means personalisation cannot be a single English-language template — Fundle's AI Agents natively handle communication personalisation across Hindi, Tamil, Telugu, Marathi, Kannada, and Bengali, a capability that MoEngage and WebEngage offer at the messaging layer but not at the loyalty reasoning layer.
Third, India's loyalty economics are driven by a relatively thin but highly active top tier. Across Fundle's managed base of 1.33Cr+ members, the Pareto effect is more pronounced than global benchmarks: the top 8% of members generate 67% of attributed revenue. This means AI-driven hyper-personalisation for the top tier is a disproportionate value driver — a 5% improvement in top-tier retention at a mall like Select CITYWALK can translate to ₹3–5 crore in additional annual revenue at current basket sizes. Rule-based platforms treat this segment with the same blunt instrument as the rest of the base. Fundle AI Agents assign dedicated engagement logic to high-value members: proactive service triggers, exclusive event invitations, personalised reward acceleration windows, and white-glove recovery flows when a service failure is detected.
Fundle Brand Loyalty — the product layer for individual retail chains operating outside mall environments — brings the same Agentic AI capability to brands like Apollo Pharmacy, FabIndia, and Reliance Trends running independent loyalty programmes. The critical Indian market fit here is the ability to handle the Tier 2 and Tier 3 city customer profile: lower average ticket, higher visit frequency, strong price sensitivity, and a preference for WhatsApp over email. Fundle's AI Workflow optimises channel selection at the individual level — not by segment — so a customer in Nagpur who has never opened a brand email in 18 months receives her next engagement offer via WhatsApp at a time the agent has learned she typically browses.
Finally, the Fundle Mall Loyalty product addresses the unique commercial structure of Indian malls: the mall operator wants to aggregate loyalty across 80–150 brand tenants while each tenant wants to own the customer relationship within their store. Fundle's architecture solves this with a federated data model — tenants see their own customers' behaviour; the mall operator sees cross-tenant patterns — without violating the commercial agreements that govern data sharing in typical lease structures.
Fundle Agentic AI vs. Legacy Loyalty Platforms: Operator Decision Matrix
Customer Engagement and Revenue Impact: What the Numbers Show
The commercial case for Agentic AI in retail loyalty is not theoretical — it is observable in the gap between what loyalty programmes cost and what they return. Industry benchmarks from the Loyalty360 India Retail Survey (2024) show that the average Indian mall loyalty programme costs ₹18–24 per active member per year in platform fees, communication costs, and reward liability. The average incremental revenue attributable to the loyalty programme — net of what members would have spent without the programme — is ₹110–140 per active member annually on legacy platforms. That is a 5.5–7× return, which sounds reasonable until you recognise that 58% of enrolled members are dormant and the true active base is a fraction of the reported membership.
When AI-driven personalisation is applied to the full member lifecycle — not just the active cohort — the economics change materially. Fundle manages 1.33Cr+ members and ₹2,329Cr+ revenue with cutting-edge AI loyalty tech, and the platform-level data shows that AI-driven re-engagement of lapsed members who have been dormant for 60–180 days generates a 19% reactivation rate versus 4.2% for generic blast campaigns. At a mall with 5 lakh lapsed members, that difference translates to 74,500 additional reactivated customers per campaign cycle — each of whom, at an average reactivated basket of ₹2,800, represents ₹20.9 crore in recoverable revenue per cycle.
The engagement mechanics that Fundle AI Agents deploy are qualitatively different from what campaign managers produce manually. The agents run continuous A/B optimisation at the individual level — not the segment level — meaning the system learns that a specific customer converts on reward acceleration offers on Saturday mornings and ignores discount coupons entirely. This granularity is computationally impossible to achieve manually and is the reason AI-native platforms consistently outperform rule-based systems on revenue-per-communication metrics.
For retail chains like Lifestyle or Pantaloons with a national loyalty base, the Fundle AI Platform's ability to model seasonal purchase intent — detecting that a member in Chennai typically upgrades her ethnic wear basket in the six weeks before Pongal — enables proactive inventory-linked offers rather than post-purchase afterthoughts. This shifts loyalty from a cost-of-doing-business to a genuine demand-generation channel, with a measurable contribution to same-store sales growth that CFOs can validate on the P&L.
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: Migrating to Agentic AI Loyalty Without Business Disruption
Audit Your Member Data Quality
Before any AI system can reason accurately, the underlying data must be clean. Run a data quality audit across your POS, CRM, and loyalty database: identify duplicate member IDs, missing mobile numbers, and unlinked transactions. In Indian retail, 22–35% of loyalty records typically have at least one critical data quality issue that will corrupt AI predictions. Fundle's onboarding team benchmarks this against the 1.33Cr+ member base to give you an honest data health score before platform deployment begins.
Map the Member Lifecycle and Define AI Agent Goals
Work with your customer engagement team to define the four to six lifecycle stages that matter for your business: first visit, second purchase, category expansion, high-value tier, at-risk, and lapsed. For each stage, define the commercial goal the AI agent is trying to achieve — not the message, but the outcome. The Fundle AI Workflow engine translates these outcome goals into autonomous agent behaviour, selecting channels, offers, and timing without manual intervention.
Integrate POS and Touchpoint Data in Real Time
Connect your POS estate to the Fundle AI Platform using pre-certified connectors for POSist, GoFrugal, Petpooja, or Wondersoft. For mall operators with multi-POS tenant environments, the Fundle Mall Loyalty federated data model handles cross-tenant data ingestion within your commercial data-sharing framework. Target a maximum 72-hour lag between system connection and the first live AI agent decision — anything longer indicates a data pipeline issue that will degrade AI performance.
Launch AI Agents in Parallel with Existing Campaigns
Do not shut down your existing loyalty communications on day one. Run Fundle AI Agents on a defined member cohort — typically 15–20% of your active base — while legacy campaigns continue for the rest. Measure redemption rate, incremental revenue per member, and opt-out rate across both cohorts over 60 days. This A/B structure gives your CFO and board a clean commercial validation before full migration, and it gives the AI agents time to learn member-level patterns before operating at full scale.
Scale, Measure, and Refine with Live KPI Dashboards
Once the parallel test demonstrates statistical significance — typically 45–60 days with a cohort of 50,000+ members — migrate the full base to Fundle Agentic AI. Track six primary KPIs weekly: active member rate, redemption rate, average revenue per active member, churn rate by tier, reactivation rate for lapsed members, and NPS segmented by loyalty tier. The Fundle AI Platform surfaces all six in a real-time operator dashboard, eliminating the monthly reporting lag that makes legacy loyalty analytics operationally useless.
Compliance and Data Privacy Advantages in the DPDP Era
India's Digital Personal Data Protection Act (DPDP) 2023 is the most consequential shift in the regulatory environment for retail loyalty since Aadhaar-linked KYC was introduced for financial services. The law requires explicit, purpose-specific consent for every use of personal data, a clear mechanism for data principals to withdraw consent, and demonstrable accountability for data processors handling personal data on behalf of data fiduciaries. For a mall operator managing 15 lakh loyalty members across 100+ brand tenants, this is not a legal checkbox — it is an operational architecture problem.
Legacy loyalty platforms were built before DPDP was contemplated. Their data models typically store a flat member record with a single opt-in flag — 'the customer agreed to marketing communications at enrolment.' Under DPDP, that single flag is insufficient. The platform must store granular, timestamped consent records for each purpose of data use: transactional loyalty, marketing personalisation, third-party brand communication, analytics profiling. When a member withdraws consent for marketing personalisation, the platform must be able to honour that withdrawal in real time without deleting the transactional record needed for reward liability management. Most legacy stacks cannot perform this surgical consent operation without manual database intervention.
Fundle's data architecture was designed with DPDP compliance as a first-order requirement, not a retrofit. The Fundle AI Platform maintains a purpose-specific consent ledger at the individual member level, with real-time enforcement across every AI Agent decision. When an agent evaluates a next best action for a member who has withdrawn marketing personalisation consent, the agent's available action set is automatically constrained to DPDP-permissible communications. No human configuration is required — the compliance constraint is built into the agent's decision framework.
Beyond DPDP, Fundle's first-party data architecture is structurally advantageous as third-party cookies deprecate across browsers and as Meta and Google tighten their API data-sharing terms for lookalike audience building. The mall and brand operators on the Fundle platform own their member data outright — there is no data-sharing clause that routes anonymised behavioural data to the platform vendor's data marketplace, a practice that some Western loyalty SaaS vendors have embedded in their standard terms. For Indian retail CMOs who are increasingly accountable for first-party data strategy, this ownership structure is commercially and legally significant.
- Does your platform maintain a real-time, unified member profile that updates within seconds of a POS transaction — not overnight in a batch job?
- Can your system autonomously select the next best action for an individual member without a campaign manager configuring a rule or scheduling a send?
- Do you have certified, live integrations with the POS systems your brand tenants actually use — POSist, GoFrugal, Petpooja, Wondersoft — or are you reconciling transaction files manually?
- Does your loyalty data model store purpose-specific, timestamped consent records that satisfy DPDP 2023 requirements for each use of member data?
- Can your platform communicate with members in their preferred Indian language at the AI reasoning layer — not just by sending a translated template?
- Do you have a predictive churn score at the individual member level that updates continuously, or are you working from a monthly cohort analysis?
- Is your loyalty programme contributing a measurable, CFO-validated increment to same-store sales growth, or is it a cost centre justified by open rates and points-issued reports?
“Indian retail has too many loyalty programmes and too few loyal customers. The difference between the two is whether your system reasons about people or just records their transactions.”
How Fundle solves this
Fundle's entire product architecture starts from a single conviction: loyalty in Indian retail will only generate durable commercial value when the system can think and act on behalf of the operator, not just record and report. That conviction is expressed across every product layer — the Fundle AI Platform at the infrastructure level, Fundle Mall Loyalty for complex multi-tenant mall environments, Fundle Brand Loyalty for single-brand retail chains, and the Fundle AI Agents that execute autonomous customer engagement decisions in real time.
The Fundle Agentic AI layer is the operational core. Each AI Agent maintains a persistent, enriched profile of an individual member — transaction history, category preferences, channel responsiveness, lifecycle stage, predicted churn probability, seasonal purchase patterns, and DPDP consent status. When the agent determines that a member at Phoenix Marketcity is showing early churn signals — visit frequency down 40% over six weeks, last transaction 48 days ago — it does not wait for a weekly analyst report. It immediately evaluates the optimal intervention: a personalised reward acceleration offer, an exclusive event invitation, or a direct outreach through the member's highest-response channel. It executes the intervention through the Fundle AI Workflow engine and then tracks whether the intervention achieves the re-engagement goal, updating the agent's model for that member and for similar members across the base.
Fundle AI Workflow is the execution backbone that translates agent decisions into cross-channel actions — WhatsApp, SMS, push notification, in-app message, or a trigger to the brand's store associate system for a high-value member. The workflow engine handles throttling, frequency caps, and DPDP consent enforcement automatically, so operators do not need to build compliance logic into every campaign. For mall CMOs managing 80+ tenant brands, this eliminates the coordination overhead of ensuring that every tenant's individual campaign is DPDP-compliant before it sends.
Vineet Narang's founding vision for Fundle was to make enterprise-grade AI loyalty accessible to Indian mall operators and retail chains at a price point and implementation timeline that the market could actually absorb — not a two-year SAP-style deployment, but a 30–60 day go-live with measurable commercial outcomes in the first quarter. The Fundle AI Platform's pre-certified POS connectors, pre-built DPDP compliance architecture, and out-of-the-box AI Agent templates for the most common Indian retail loyalty use cases — win-back, tier upgrade, category expansion, seasonal reactivation — make that timeline achievable without a large internal technology team. For Indian mall operators and retail chain CMOs evaluating their loyalty technology roadmap for FY26, the question is no longer whether Agentic AI in retail loyalty is ready. Fundle's 1.33Cr+ member base and ₹2,329Cr+ in managed revenue confirm it is. The question is whether your current platform will still be competitive when it is.
Frequently asked
What is Agentic AI in retail loyalty and how is it different from a standard AI recommendation engine?+
A recommendation engine is a model that predicts what a customer might want next, given a query. An AI loyalty agent is a persistent, goal-directed system that continuously monitors a member's behaviour, plans multi-step interventions, selects channels and timing autonomously, executes actions through an integrated workflow engine, and updates its understanding based on outcomes — all without a human configuring each step. Fundle AI Agents operate this way across all members simultaneously, in real time.
How long does it take to implement the Fundle AI Platform for a mall with 80+ brand tenants?+
For a mall operating on POSist, GoFrugal, Petpooja, or Wondersoft, Fundle's pre-certified connectors reduce integration timelines to 30–60 days for the core platform, with AI Agents going live on an initial member cohort within the first 30 days. Full tenant onboarding depends on the number of POS environments, but Fundle's federated data model means tenants can be added incrementally without disrupting the live loyalty operation.
Is Fundle's loyalty platform compliant with India's DPDP Act 2023?+
Yes. Fundle's data architecture was built with DPDP 2023 as a first-order design requirement. The platform maintains purpose-specific, timestamped consent records at the individual member level, enforces consent constraints in real time across every AI Agent decision, and provides members with a self-service consent management interface. This is structural, not a retrofit compliance add-on.
How does Fundle Mall Loyalty handle the data-sharing complexity between a mall operator and its brand tenants?+
Fundle Mall Loyalty uses a federated data model: each brand tenant has read access to their own customers' loyalty behaviour, while the mall operator has access to cross-tenant behavioural patterns for footfall and category analytics. Data that would violate typical lease-level commercial agreements — such as tenant-to-tenant customer sharing — is segregated at the architecture level, not managed by policy alone.
What results can a retail chain realistically expect in the first 90 days on Fundle's platform?+
Based on Fundle's active deployments, retailers consistently see a 15–22% improvement in active member rate within 90 days through AI-driven reactivation campaigns, a 2–3 percentage point reduction in opt-out rates as communication relevance improves, and a measurable lift in redemption rate that signals genuine customer engagement rather than passive enrolment. Revenue attribution takes 120–150 days to reach statistical confidence at the member level.
How does Fundle's pricing compare to legacy platforms like Capillary or EasyRewardz?+
Fundle's commercial model is outcome-oriented — typically structured around active member count and AI-attributed revenue uplift rather than per-campaign or per-communication fees. This aligns platform cost with business results rather than activity volume. For most mid-to-large mall operators, the total cost of ownership including integration, compliance, and communication infrastructure is comparable to legacy platforms, while the revenue attribution per rupee spent is materially higher due to AI precision.
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.