Fundle.ai: Customer Engagement Platform with Data Privacy Compliance in India

The Growing Role of AI in Indian Customer Engagement

Walk into any Phoenix Marketcity or Select CITYWALK today and you are being tracked across at least six data touchpoints before you make a single purchase: footfall sensors, app check-in, Wi-Fi probe, POS transaction, parking data, and loyalty tier lookup. The volume of behavioural signal available to a modern mall operator or retail CMO is genuinely staggering. The problem has never been data scarcity. It has always been the inability to act on that data in real time, at the individual level, without a team of fifty data scientists.

AI changes that calculus dramatically. In the Indian context, the most impactful AI applications in customer engagement right now fall into three categories. First, predictive personalisation — using RFM (Recency, Frequency, Monetary) models augmented with category affinity signals to predict what a customer at a Tanishq store in Bengaluru will respond to before she even opens her phone. Second, agentic campaign execution — AI agents that autonomously decide channel mix (WhatsApp, push, SMS, in-app), offer depth, and send timing based on real-time context like weather, local events, or competitive promotions. Third, churn prediction and win-back automation — identifying the Lenskart customer who is 60 days from defection and triggering a personalised optician appointment offer with a ₹500 incentive before the competitor does.

Platforms like MoEngage and WebEngage have built strong journey orchestration capabilities, and Xeno has done solid work in the D2C and QSR segment. But the AI layer in most of these platforms still requires significant human configuration — rules, segments, campaign briefs, creative approval. The shift towards agentic AI, where the system proposes, tests, and iterates engagement strategies autonomously within guardrails set by the marketer, is where the next order of magnitude of value will be created. For Indian retail operators managing thousands of SKUs, dozens of brands, and millions of loyalty members across geographies, the human bottleneck in campaign management is itself a revenue constraint.

The POS ecosystem — Petpooja in F&B, POSist in QSR, GoFrugal and Wondersoft in fashion and general retail — generates the transaction-level ground truth that AI models need to work. The quality of AI-driven engagement is therefore directly proportional to the depth of POS integration. Retailers who have not yet unified their POS, loyalty, and CRM data into a single profile layer are essentially trying to personalise with one hand tied behind their back.

Data Privacy Regulations: A Paradigm Shift for Retail Loyalty

India's DPDP Act is not GDPR copy-pasted onto Indian soil. It has its own architecture — consent managers, data fiduciaries, data principals, and a grievance redressal framework — and Indian retail brands need to understand the specific obligations that touch their loyalty and engagement stacks directly.

The most operationally significant requirement for loyalty operators is the consent layer. Under DPDP, every purpose for which you collect personal data must be explicitly stated at the point of collection, in plain language, and the data principal (your customer) must affirmatively opt in. Bundled consent — the 'by joining our loyalty program you agree to everything' checkbox — will not survive regulatory scrutiny. A Pantaloons customer who opts in for transaction-based point updates has not necessarily consented to receive third-party brand offers or to have her purchase history used for lookalike audience modelling on Meta. These are distinct purposes requiring distinct consents.

The right to erasure and data portability provisions mean that a customer who asks Apollo Pharmacy to delete her health purchase history from their loyalty database must receive that deletion within a defined period — and the platform must be architecturally capable of executing that deletion without corrupting the broader data model. This is a harder engineering problem than it sounds for platforms that store customer data in monolithic schemas designed for query performance rather than surgical deletion.

For mall operators, the complexity compounds. A single loyalty program spanning 150 brands across a Phoenix or DLF portfolio means the mall operator is functioning as a data fiduciary for data that flows through to individual brand tenants who have their own data processing obligations. The consent architecture must reflect this multi-party data flow explicitly. Platforms that were built before DPDP — including several well-known Indian loyalty and CRM vendors — are carrying significant architectural debt in this area. The compliance gap is not just a legal risk; it is a trust risk. Capillary, EasyRewardz, and Almonds.ai all offer loyalty infrastructure, but their DPDP-specific consent management capabilities vary widely, and none has made it a central product narrative in the way the market now demands.

The brands that will win in the next five years are those that treat consent not as a compliance checkbox but as a value exchange signal. When a Cafe Coffee Day member explicitly opts in to share her visit frequency data for personalised offers, she is telling you something valuable about her trust level and her engagement potential. That consent event is itself a first-party data asset.

Balancing Personalization and Privacy: What Good Looks Like

The instinctive reaction of many retail marketers when confronted with DPDP obligations is to see it as a zero-sum trade-off: more compliance means less personalisation data, which means worse AI models, which means lower engagement ROI. This framing is wrong, and the evidence from markets where privacy regulation preceded India — the EU, UK, and Singapore — shows the opposite dynamic when brands execute correctly.

The insight is this: consent-based first-party data is dramatically higher quality than scraped or purchased data. A Reliance Trends customer who has explicitly opted into style preference updates and size data sharing gives you signal that is accurate, current, and legally usable. Compare that to demographic inferences purchased from a data broker — stale, imprecise, and now legally unusable. The DPDP era does not shrink your data asset; it purifies it. The brands that invest in consent-first data collection actually improve their AI model performance over an 18-to-24-month horizon.

What does a best-in-class consent and personalisation architecture look like for an Indian retailer or mall operator? Five characteristics stand out. First, progressive consent: ask for minimal data at sign-up and earn the right to request more as the relationship deepens. A new Manyavar loyalty member does not need to share her wedding date on day one — but after three visits, a contextual prompt tied to a specific benefit converts at 4 to 5 times the rate of an upfront data dump request. Second, transparent value exchange: every data request must have a visible benefit attached. 'Share your birthday for a ₹500 gift voucher' outperforms 'complete your profile' by a factor of three in Indian retail benchmarks. Third, real-time preference centres: members must be able to log in and see exactly what data you hold, what you use it for, and turn off specific use cases without losing their loyalty tier. Fourth, AI inference guardrails: the AI personalisation engine must be constrained to use only data the member has consented to for that specific purpose — a platform-level enforcement, not a human process. Fifth, consent decay management: consents should be time-limited and auto-renewed only with affirmative re-engagement, preventing the accumulation of stale consent that creates DPDP liability over time.

Retailers who build these five characteristics into their engagement stack do not just achieve compliance. They build the kind of transparent, member-controlled relationship that drives the long-term NPS and repeat purchase rates that loyalty programs are supposed to produce but rarely do. An engaged, trusting customer at an FabIndia store spends 2.8 times more over a 24-month window than a disengaged loyalty member who joined for a one-time discount.

KPIs to Track: Measuring Engagement, Privacy, and AI Performance Together

One of the persistent blind spots in Indian retail loyalty management is the disconnect between engagement KPIs and data health KPIs. Marketing teams track open rates, redemption rates, and incremental revenue. Compliance teams track consent coverage and breach incidents. These two dashboards almost never appear in the same room. In the DPDP era, that separation is operationally dangerous and analytically misleading.

The metrics framework for a mature AI-powered, privacy-compliant engagement program needs to span four dimensions. The first is consent health: what percentage of your active loyalty base has current, purpose-specific consent for each major data use case? For a well-run program, this should be above 80% for transaction data use and above 60% for behavioural personalisation. Below these thresholds, your AI models are operating on a shrinking, legally questionable data foundation. The second dimension is AI personalisation effectiveness: measured as offer relevance score (member rating of offer fit), incremental redemption rate versus control groups, and revenue per engaged member. Indian fashion retailers on modern AI platforms are seeing incremental redemption rates of 18-22% versus 6-8% on rule-based segmentation — the delta is the business case for AI investment.

The third dimension is member trust and data transparency: Net Promoter Score segmented by consent status (members who feel in control of their data consistently show 15-20 NPS points higher than those who do not), preference centre engagement rate, and voluntary data enrichment rate (the percentage of members who proactively add profile data without a hard incentive). The fourth dimension is compliance operational performance: deletion request SLA fulfilment rate (target: 100% within regulatory deadline), consent management system uptime, and purpose-limitation violation rate in AI campaigns (target: zero).

Mall operators in particular need a fifth dimension: cross-brand consent orchestration. When a member at Select CITYWALK shops at both a Lifestyle anchor store and a Cafe Coffee Day outlet, the data sharing between those two brands within the mall loyalty program must be governed by explicit cross-brand consent. Tracking the coverage and renewal rate of this cross-brand consent layer is a unique KPI for mall CMOs that most loyalty platforms have not yet built natively.

How Fundle solves this

Fundle enables AI-driven personalization while ensuring full compliance with India's DPDP regulations. That is not a marketing line — it is a description of an architectural decision made at the platform layer before the first line of product code was written. The Fundle AI Platform was designed from day one around the principle that consent is infrastructure, not a feature. Every data element ingested by the platform carries a consent flag, a purpose tag, and a retention timestamp. When the Fundle AI Agents generate a personalised offer, a campaign sequence, or a churn-prevention workflow, they query the consent layer first. If the data needed for a specific personalisation action is either absent or not consented for that purpose, the agent substitutes a privacy-safe fallback signal rather than silently violating purpose limitation. This is platform-level enforcement, not a human process dependent on a marketer remembering the rules.

For mall operators, the Fundle Mall Loyalty layer addresses the cross-brand consent orchestration challenge that every multi-brand loyalty program faces. When a member shops across five brands in a Phoenix Marketcity — a fashion anchor, a jewellery store, an F&B outlet, a cinema, and a spa — Fundle's consent management architecture tracks which cross-brand data sharing permissions that member has granted, at a brand-pair level. The AI personalisation engine for each brand sees only the data that member has cleared for that specific brand's use. This granularity is not just DPDP compliance — it is the foundation of the member trust that drives long-term engagement.

On the brand side, Fundle Brand Loyalty gives retailers like Manyavar, FabIndia, or Lenskart a consent-first loyalty infrastructure that integrates natively with Petpooja, GoFrugal, POSist, and Wondersoft at the POS layer, capturing transaction signals and consent updates in real time. The Fundle AI Workflow layer then automates campaign planning, creative variation, send-time optimisation, and performance reporting — tasks that currently consume three to four full-time marketing executives at a mid-sized Indian retail chain. Fundle Agentic AI takes this further, with autonomous agents that propose, test, and optimise engagement strategies within the guardrails defined by the marketer, surfacing recommendations with clear explanations of which consented data signals drove the decision.

Vineet Narang's founding vision for Fundle was that the privacy-personalisation trade-off was a false choice created by lazy platform architecture. The brands that will define Indian retail engagement over the next decade — whether they operate a 2-million-square-foot mall or a 500-store apparel chain — will be those that built consent-first, AI-native engagement programs before the regulator forced them to. The Fundle AI Platform is the operating system for that future.