PII in the Age of AI: Balancing Personalization with Customer Guardrails

As artificial intelligence models become central to Philippine financial services, retail ecosystems, and capital markets, consumer data is becoming far more predictive. An individual’s credit history, daily grocery shopping basket, or personal investment portfolio can reveal intimate lifestyle shifts, financial pressure, or major family milestones long before a customer explicitly states them.
While hyper-personalization allows organizations to deliver highly relevant recommendations, it also raises critical data privacy questions. In a region where digital adoption is surging, brands must navigate the fine line between helpful personalization and intrusive surveillance.

Defining Sensitive Data in Smart Ecosystems

During a panel session at the summit, industry leaders explored this ethical balance. The panel brought together Poonam Anandani, Head of Technology, Process, and Data at Home Credit Philippines; Sally Garcia, AVP of Data Management at Robinsons Retail Holdings; and Andoy Beltran, VP of Digital Solutions and Investor Engagement at First Metro Securities. The conversation was moderated by Ritesh Gandhi, Senior Director of Customer Success for SEA and ANZ at MoEngage.

Between consumer lending, multi-brand retail, and capital markets, these three organizations handle some of the most sensitive personally identifiable information (PII) a customer generates. The panel discussed how modern AI models have fundamentally altered what companies consider “sensitive” data across three distinct sectors:

Alternative Credit Scoring in Consumer Lending: In non-traditional lending, Home Credit frequently services first-time borrowers who lack formal bank credit histories. Poonam Anandani discussed how alternative data sources allow institutions to evaluate credit risk responsibly without crossing privacy boundaries or reading too much into a customer’s personal life. Drawing on her experience in global investment banking technology, she noted that data governance playbooks must be adapted thoughtfully to fit local consumer needs without compromising protection.

Retail Basket Analytics and Health Cues: In retail ecosystems, a grocery basket reveals subtle household changes, such as health condition adjustments or expanding family sizes. Sally Garcia shared that while unifying customer data across loyalty platforms like Go Rewards allows brands to tailor relevant offers, the primary challenge lies in moving profile data safely between store formats. Brands must draw a firm line between using purchasing signals for helpful recommendations and making customer feel monitored.

Financial Suitability and Portfolio Guidance: In wealth management, First Metro Securities handles income brackets, risk tolerances, and complete portfolio histories. Andoy Beltran addressed the ongoing tension between regulatory suitability rules and personalized investment guidance. While suitability regulations exist to protect investors, AI models must be designed so that smart recommendations never push users toward products outside their explicit financial intent.

Setting Uncrossable Lines for AI Execution

A central theme of the discussion was that predictive capabilities must never outpace customer consent. When algorithms begin drawing inferences from combined data streams, organizations can easily overstep if clear boundaries are not set in advance.
The panelists agreed that hyper-personalization should never compromise data transparency. When customers share data, they expect that information to be used to improve their direct service experience, not to be sold or processed in opaque ways behind closed doors.

To maintain trust as AI capabilities advance, the panel emphasized several structural non-negotiables:

Explicit Purpose Consent: Customers must retain clear visibility over how their transactional records are shared, especially when data moves between distinct partner brands or sister companies.

Preventing Vulnerability Exploitation: Predictive algorithms must be barred from using financial stress or sudden lifestyle changes to target users with high-risk offers.

Human Oversight over Automated Outcomes: High-stakes decisions involving credit approvals or investment suitability must always maintain clear auditing trails and human checks.

The panel concluded with a shared directive: no matter how much an AI model might improve engagement or conversion metrics, customer privacy boundaries are non-negotiable. In an increasingly automated digital economy, respecting customer limits remains the single most reliable driver of long-term brand loyalty.

——————-
MoEngage #GROWTH Summit Philippines 2026 convened growth marketers, product owners, and executive leaders from leading regional consumer brands to share actionable strategies on customer engagement, AI automation, and data governance.