“Most platforms automate marketing. Fundle automates outcomes — incremental revenue, retention lift, attributed footfall. The number is the product.”
- •Understand why rule-based loyalty programs fail Indian retail operators at scale
- •See exactly how Orchid Hotels and NewU Beauty deployed Fundle's Agentic AI loyalty stack
- •Measure the business outcomes: retention lift, revenue per member, and campaign ROI
- •Follow a five-step implementation playbook any CRM Head can replicate in under 90 days
- •Evaluate Fundle against legacy platforms like Capillary, EasyRewardz, and WebEngage
Indian retail loyalty is having a reckoning. After a decade of stamp-card mechanics dressed up in mobile apps, operators are waking up to a brutal truth: a points balance is not a relationship. The average Indian mall loyalty member earns points at three to five brands, redeems at fewer than two, and quietly lapses within nine months. For a mall the size of Phoenix Marketcity Mumbai — anchoring over 600 brand touchpoints — that attrition curve represents hundreds of crores in recoverable revenue left on the table every single year.
The core problem is architectural. Most loyalty platforms sold to Indian retailers today — including several otherwise capable tools from Capillary, EasyRewardz, and Xeno — are fundamentally rule engines. A marketing manager writes a condition: 'If customer has not transacted in 60 days, send a ₹100 voucher.' The system obeys. The customer receives the same voucher as 40,000 other lapsing members regardless of whether she is a ₹2,000-per-month apparel buyer at Pantaloons or a ₹18,000-per-visit jewellery customer at Tanishq. The intervention is blunt, the economics are leaky, and the customer experience is indistinguishable from spam.
Agentic AI for retail loyalty changes the fundamental contract between the platform and the operator. Instead of executing rules, AI agents autonomously observe customer behaviour across channels, infer intent, select the optimal intervention type, determine the right moment, personalise the content, dispatch the communication, monitor the response, and update the customer model — all without a human writing a campaign brief. This is not marketing automation with a neural network bolted on. It is a category shift from reactive CRM to proactive, self-correcting engagement intelligence.
This is precisely the architecture that Fundle was built to deliver. Orchid Hotels and NewU Beauty — two operationally distinct but strategically aligned hospitality and beauty retail brands — became early proof points of what agentic loyalty looks like in production at Indian scale. Their story is instructive for any CRM Head or Mall Marketing Director evaluating whether the next generation of AI tools is ready for real deployment, or whether it remains a vendor slide-deck aspiration.
Indian Retail Loyalty: The Benchmark Gap
Overview of Orchid Hotels and NewU Beauty Loyalty Challenges
Orchid Hotels is a mid-premium hospitality brand operating across India with a guest database that straddles business travellers, leisure families, and frequent weekend escapees. Their loyalty challenge was not volume — they had hundreds of thousands of enrolled members — it was depth. Stay frequency among their top-quartile members was declining, and the marketing team at Orchid was running the same four campaign types on rotation: welcome bonus, anniversary offer, win-back voucher, and festival promotion. Average email open rates had dropped below 12%. WhatsApp broadcast opt-outs were climbing. The programme felt transactional because it was transactional.
NewU Beauty, the retail pharmacy-and-beauty chain under the Dabur umbrella, faced a structurally different but equally acute problem. Their customer base skews young, urban, and purchase-driven by impulse and trend cycles — think skincare launches, festive makeup hauls, and healthcare consumable replenishment. The challenge was basket depth and cross-category pull. A customer who bought sunscreen twice was not being connected to tinted moisturiser, SPF lip balm, or the derma consultation service. Each transaction was being treated as a standalone event rather than a signal inside a longer customer journey. With stores across Select CITYWALK, Inorbit Mall, and Phoenix Palladium, NewU needed engagement that felt as curated as the store experience itself.
Both brands shared three systemic failure modes common across Indian retail loyalty: first, their data was siloed — POS, app, WhatsApp, and email were not talking to each other in real time. Second, their campaign logic was flat — the same customer segment received the same message regardless of recent behaviour signals. Third, their teams were overwhelmed — a two-person CRM team at NewU was expected to manage 14 campaign workflows simultaneously across 80+ store locations. Neither brand lacked ambition. They lacked the infrastructure to execute at the speed and granularity that modern Indian consumers now expect from brands they choose to remain loyal to.
Orchid Hotels & NewU Beauty saw improved loyalty retention using Fundle's AI engagement platform — and the path to that outcome began with a hard-headed audit of where their existing engagement stack was breaking down, not a feature comparison exercise.
The Retail Loyalty Engagement Funnel: Before vs. After Agentic AI
Adoption of Fundle's Agentic AI Loyalty Solutions
The decision to move to an agentic AI for retail loyalty architecture was not taken lightly by either brand. Both Orchid Hotels and NewU Beauty had existing vendor relationships — one with a mid-tier CRM automation tool, another with a points-management platform that had been customised over three years. Migration carries switching costs: data cleansing, staff retraining, integration redevelopment, and the political capital required to retire a system someone internally championed.
What tipped the evaluation in favour of the Fundle AI Platform was a three-part proof-of-concept run over six weeks on a subset of each brand's active member base. The Fundle AI Agents were connected to live POS data, the brand's existing WhatsApp Business API, and historical transaction records going back 24 months. Within the first two weeks, the agents had autonomously identified seven distinct behavioural micro-segments at NewU that the marketing team had not explicitly defined — including a high-frequency single-category buyer cluster (body care only, 8+ transactions per year, zero cross-category) and a lapsing premium buyer group showing early churn signals four to six weeks before conventional RFM models would have flagged them.
For Orchid Hotels, the Fundle Agentic AI surfaced a guest cohort that had stayed twice in 18 months but only during long weekends — a segment with demonstrably higher lifetime value potential if engaged with contextual staycation offers in the three weeks preceding each major Indian holiday calendar window. No rule had been written to find these guests. The agent found them, constructed the hypothesis, tested a micro-campaign, measured the uplift, and escalated the finding to the CRM manager as a recommended workflow — all within the proof-of-concept window.
The Fundle Brand Loyalty module was deployed for NewU across its retail footprint, while Orchid Hotels implemented a hybrid of the Fundle Loyalty Platform and the hospitality-configured engagement workflows within Fundle AI Workflow. Crucially, both brands retained full control over spend caps, channel permissions, and brand tone-of-voice guardrails — the agents operated within defined boundaries, not as black-box automation.
Agentic AI Loyalty Platform vs. Rule-Based CRM: Head-to-Head
Implementation Process and User Experience
The implementation of agentic AI for retail loyalty across two operationally different brands illuminates the practical realities that vendor decks routinely obscure. Neither deployment was a flick-of-a-switch affair, but both ran materially faster than comparable CRM migrations in Indian retail — and the reasons are instructive.
At NewU Beauty, the data foundation work consumed the first three weeks. Transaction records from GoFrugal POS were ingested, cleaned, and merged with app registration data and WhatsApp opt-in lists. Roughly 23% of member records required deduplication — a common but underestimated problem in Indian retail where the same customer may have enrolled via a store cashier using a slightly different name spelling, via the app using a Google login, and via a third-party mall loyalty programme. The Fundle AI Platform's entity resolution layer handled the bulk of this automatically, flagging a small residual set for manual review. By week four, the Fundle AI Agents were live on a pilot cohort of 50,000 members across ten stores.
At Orchid Hotels, the integration challenge was different: guest data lived across a property management system, a legacy email CRM, and a manually maintained Excel-based VIP list that the reservations team had been curating for years. The Fundle AI Workflow ingestion pipeline connected to the PMS via a standard API and absorbed the email CRM export in batch. The VIP Excel — culturally sensitive because the reservations head had built it — was onboarded as a high-signal enrichment layer rather than being overwritten, which avoided internal friction and preserved institutional knowledge.
From a user experience standpoint, both CRM teams reported that the Fundle dashboard's 'Agent Activity Feed' was the single feature that most accelerated internal adoption. Rather than showing a black-box log of automated actions, the feed presents each agent decision in plain language: 'Sent a ₹500 spa upgrade offer to 340 guests who checked out on Sunday and have a second stay within 45 days — predicted redemption rate 18%, based on similar cohort response last quarter.' This explainability layer matters enormously in Indian corporate culture, where marketing heads are accountable to category and brand leads who will ask 'why did we send that?' The agent had to be able to answer.
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 Loyalty in Indian Retail
Audit and Unify Your First-Party Data
Before any AI agent can act intelligently, the data substrate must be coherent. Map every customer touchpoint — POS (POSist, GoFrugal, Wondersoft), app, WhatsApp, email, in-store staff interactions — and run an entity resolution exercise to deduplicate member records. Target: a single customer view with 85%+ match rate before go-live.
Define Agent Guardrails, Not Campaign Rules
Shift the briefing model. Instead of writing 'If X then Y' rules, define the boundaries within which AI agents operate: maximum discount depth (e.g., never exceed 15% off), approved channels per customer, brand voice guidelines, and spend caps per segment. The Fundle AI Agents operate autonomously within these parameters.
Run a Six-Week Proof of Concept on a Live Cohort
Select a representative slice of your member base — ideally 40,000 to 80,000 members across your top stores or properties. Let the Fundle Agentic AI run autonomously against a holdout control group. Measure incremental transaction rate, redemption rate, and average order value lift. Use the PoC data to build your internal business case.
Migrate Campaign Workflows Progressively, Not All at Once
Identify the three to five highest-volume, lowest-complexity campaign types your team runs manually — win-back, birthday, tier-upgrade nudge — and migrate these to Fundle AI Workflow first. This frees CRM headcount immediately for higher-judgment work while demonstrating ROI to internal stakeholders within the first 60 days.
Instrument KPIs and Create a Weekly Agent Performance Review
Establish a standing weekly review where the CRM team examines agent recommendations, overrides where needed, and feeds qualitative brand intelligence back into the system. Track: active member rate, revenue per engaged member (₹), redemption rate, campaign contribution to total store revenue, and NPS delta between programme members and non-members.
Business Outcomes and Customer Feedback
Outcomes in loyalty technology are notoriously easy to cherry-pick and difficult to attribute cleanly. The discipline applied by both Orchid Hotels and NewU Beauty in their Fundle deployments — specifically the use of randomised holdout groups during the proof-of-concept and the first 90 days of full rollout — makes the results worth examining seriously rather than treating as marketing testimony.
At NewU Beauty, the most significant metric shift was in cross-category purchase rate among previously single-category buyers. Within 90 days of the Fundle Brand Loyalty agents going live, the body-care-only cohort identified during the PoC showed a 26 percentage-point increase in the proportion making at least one purchase in a second category. This is not a trivial number: cross-category buyers at NewU carry an average basket 2.1× larger than single-category buyers and visit 40% more frequently. The unit economics of acquiring a new customer versus converting a single-category buyer to a cross-category buyer at NewU favour the latter by a factor of roughly 6:1 in cost-per-incremental-revenue terms.
At Orchid Hotels, the long-weekend-stay cohort identified by the Fundle AI Agents converted at 22% on contextual staycation offers — more than three times the brand's historical win-back campaign conversion rate of 6-7%. Revenue per communication sent improved by approximately 4.1× compared to the broadcast campaigns the same team had been running in the prior quarter. Guest feedback collected post-stay in this cohort showed higher satisfaction scores specifically on 'the hotel understood what I was looking for' — a proxy measure for perceived personalisation that has direct implications for review scores on platforms like MakeMyTrip and Google.
Beyond the quantitative outcomes, both brands reported a qualitative shift in how the CRM function was perceived internally. At NewU, the marketing director presented agent-generated segment insights to the category buying team — intelligence that directly influenced a reorder decision on a skincare subcategory that was about to be delisted. The loyalty programme, historically seen as a cost centre, was now producing actionable category intelligence. That repositioning of loyalty as a strategic data asset rather than a discount-distribution mechanism is, arguably, the most durable value created by the deployment.
- Single customer view exists or data unification project is scoped and funded before vendor selection
- POS system (POSist, GoFrugal, Wondersoft, or equivalent) has an accessible API or standard export for transaction data
- WhatsApp Business API opt-in list is clean, consented, and separated from broadcast-only contacts
- Internal brand guardrails — maximum discount depth, channel preferences, tone-of-voice guidelines — are documented and signed off by brand lead
- A randomised holdout group methodology is agreed with the analytics team before go-live to enable clean incrementality measurement
- CRM team has been briefed on the shift from 'campaign manager' to 'agent supervisor' role and weekly review cadence is scheduled
- Executive sponsor (CMO or VP Marketing) has signed off on a 90-day evaluation window with agreed KPIs before ROI judgement is made
“India's loyalty problem has never been a data shortage. It has always been a speed-of-action deficit. By the time a rule fires, the customer has already made her decision. Agents act before the moment passes.”
How Fundle solves this
The Fundle AI Platform was designed from the ground up to serve the specific structural conditions of Indian retail loyalty: fragmented POS ecosystems, high member-to-staff ratios, WhatsApp as the dominant engagement channel, and the acute need for explainability in hierarchical marketing organisations. It is not a Western enterprise loyalty tool adapted for India — it is an India-first architecture that happens to be internationally deployable.
At the core is the Fundle AI Agents layer — a set of autonomous engagement agents that each specialise in a distinct customer lifecycle problem: acquisition conversion, first-purchase activation, cross-category development, churn prediction and intervention, tier-upgrade acceleration, and high-value member retention. These agents run continuously, not on a campaign schedule, which means the intervention window is always the optimal one rather than the one that fits the marketing calendar. For a brand like Manyavar — where the purchase cycle is inherently event-driven and concentrated around Diwali, wedding season, and Eid — an agent that tracks life-event signals and acts within 72 hours of a trigger is worth more than twelve months of batch campaigns.
Fundle Mall Loyalty extends the same agentic architecture to the mall operator context, where the unit of engagement is the shopper's wallet share across the entire property, not a single brand. A shopper who visits Phoenix Marketcity primarily for FabIndia and Cafe Coffee Day but has never transacted at the food court anchor can be identified, understood, and engaged by a cross-brand agent that operates within the mall's commercial priorities. Fundle AI Workflow manages the orchestration layer — ensuring that a shopper does not receive competing offers from three brands in the same hour, and that the mall-level experience feels coherent rather than cacophonous.
Fundle Agentic AI also addresses the integration reality of Indian retail, with pre-built connectors for Petpooja, GoFrugal, POSist, and Wondersoft — the four POS systems that collectively power the majority of organised Indian retail and F&B transactions. For brands like Apollo Pharmacy or Reliance Trends operating large, multi-location networks on diverse POS infrastructure, this means a genuine single-stack deployment rather than a parallel data-wrangling project. Vineet Narang's founding vision for Fundle was that intelligence should remove operational burden from retail teams, not add to it — and the integration architecture reflects that principle directly. Fundle Brand Loyalty and Fundle Mall Loyalty are not separate products requiring separate implementations; they are views into the same agentic intelligence layer, configured for the brand or the property as needed.
Frequently asked
What exactly does 'agentic AI for retail loyalty' mean in practice — how is it different from marketing automation?+
Marketing automation executes predefined rules: if condition A, send message B. Agentic AI for retail loyalty means the system autonomously observes customer behaviour, forms a hypothesis about the best intervention, selects the channel and content, acts, measures the outcome, and updates its model — without a human writing each campaign. Fundle AI Agents do this continuously across your entire member base, not on a scheduled campaign cycle.
How long does it typically take to go live with the Fundle AI Platform for a mid-size retail brand in India?+
For a brand with a reasonably clean POS data export and an existing WhatsApp Business API setup, the Fundle AI Platform can reach a live-agent state on a pilot cohort in four to six weeks. Full rollout across all stores and channels typically completes in 10 to 14 weeks. Brands with more complex data environments — multiple POS systems, legacy CRM exports — should budget for an additional two to four weeks of data unification work upfront.
How does Fundle's approach compare to platforms like Capillary, EasyRewardz, or MoEngage for Indian retail loyalty?+
Capillary and EasyRewardz are strong rule-based loyalty engines with deep Indian retail integrations — well suited for straightforward points-and-tiers programmes. MoEngage and WebEngage are excellent campaign automation tools but are not loyalty-native. Fundle AI Platform occupies a different category: it is agentic-first, meaning the intelligence layer autonomously drives engagement rather than executing marketer-defined rules. For brands that have outgrown rule-based loyalty and need autonomous, self-optimising engagement, Fundle is purpose-built for that transition.
How did Orchid Hotels handle the privacy and data consent requirements for using AI agents on guest data?+
All Fundle AI Agent actions operate strictly within the consent permissions collected at enrolment and updated at each touchpoint. For Orchid Hotels, WhatsApp engagement was restricted to opted-in members only, and the agent's action log was available to the data privacy officer at any point. Fundle AI Workflow includes a consent-state check as a mandatory pre-condition before any outbound communication is dispatched — no communication bypasses this gate.
What KPIs should a CRM Head track to measure the ROI of an agentic AI loyalty deployment in the first 90 days?+
Track five core metrics against a holdout control: active member rate (transacted at least once in the period), revenue per engaged member in INR, redemption rate as a percentage of earned points, cross-category purchase rate for single-category buyers, and campaign contribution to total store revenue. Secondary metrics worth monitoring include WhatsApp opt-out rate (should decline) and NPS delta between programme members and non-members.
Can Fundle AI Agents work across both a mall loyalty programme and individual brand loyalty programmes simultaneously?+
Yes. Fundle Mall Loyalty and Fundle Brand Loyalty are designed to coexist on the same platform, with the Fundle AI Workflow layer managing the orchestration to prevent offer conflicts and ensure a coherent shopper experience. A customer who is a member of both the Phoenix Marketcity mall programme and a specific brand programme within the mall will receive interventions that are coordinated across both, not competing with each other.
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.