Smart Merchant Recommendations

Overview

Smart Merchant Recommendations are the core of the Credit Card Co-Pilot experience. The system eliminates the guesswork of choosing between multiple credit cards by analyzing every transaction against your specific wallet, current reward rules, and spending milestones.

Instead of memorizing complex reward structures, you receive a single, data-driven recommendation for the "Best Card" to use at any given merchant.

How the Recommendation Engine Works

The engine performs a real-time calculation whenever you search for a merchant or when a new transaction is detected. It processes four primary data points to determine the ranking:

1. Merchant Category: Identifies if the spend is for dining, fuel, travel, or shopping (e.g., "Zomato" is categorized as dining).
2. Transaction Channel: Distinguishes between online, offline (POS), and UPI payments, as rewards often vary significantly by channel.
3. Transaction Amount: Evaluates the spend against reward caps (monthly limits) and minimum spend requirements.
4. Wallet Context: Checks your specific cards to see which one offers the highest effective reward rate for that specific combination.

Key Recommendation Factors

The system doesn't just look for the highest percentage; it looks for the highest net value.

• Accelerated Rewards: Automatically identifies 5x or 10x reward categories.
• Cap Awareness: If you have already reached the monthly cashback limit on one card, the engine automatically pivots to your next best available card.
• Point Valuation: Converts "Reward Points" or "Miles" into their actual INR cash value so you can compare a 5% cashback card directly against a "2 Points per ₹100" card.

Milestone-Aware Recommendations

What sets the Smart Recommendation engine apart is its ability to "see" your progress toward spending milestones. If you are ₹2,000 away from a quarterly milestone that grants a ₹1,000 voucher, the engine will prioritize that card—even if its base reward rate is lower than others.

The system calculates a Milestone Proximity Result for every recommendation:

• Progress Tracking: Visualizes how close you are (e.g., 85% toward a goal).
• Bonus Value: Factors the value of the milestone bonus into the "effective rate."
• Strategic Nudges: Tells you exactly why a card is recommended: "Use this card to stay on track for your ₹50,000 quarterly spend bonus."

Confidence and Transparency

Every recommendation comes with an explainability layer, ensuring you understand the "why" behind the "what."

Confidence Bands

The system assigns a confidence score to its recommendations:

• High: The merchant and category are explicitly matched in our verified database.
• Medium: The recommendation is based on category-wide rules (e.g., "All Dining").
• Low: Based on base reward rates because the merchant category is unknown or ambiguous.

Explainability Details

You will see a short summary of the logic used, such as:

• "5% cashback on online shopping. Best rate in your wallet."
• "Accelerated 10x points detected for this merchant."
• "Switching to this card because your primary card has reached its monthly cashback cap."

Data Integrity and Parser Review

To ensure recommendations remain accurate, the system includes a Parser Review workflow. When transactions are ingested via SMS or Email, the system may occasionally find a merchant name it doesn't recognize (e.g., a local boutique).

Users can "train" their Co-Pilot by:

• Confirming Merchant Labels: Correcting or naming ambiguous merchants.
• Adjusting Channels: Specifying if a transaction was online or in-store.
• Categorization: Helping the engine learn that a specific merchant belongs to a certain reward category.

This feedback loop ensures that the next time you shop at that merchant, the recommendation is 100% accurate.

Integration Example

For developers or advanced users, the recommendation output follows a structured format that captures the full reward breakdown:

// Example of a Card Recommendation Result
{
  "cardName": "HDFC Regalia Gold",
  "rank": 1,
  "reward": {
    "effectiveRate": 0.066, // 6.6% value
    "rewardType": "points",
    "cashValue": 132.00,    // Actual INR value for the transaction
    "explanation": "10x points on SmartBuy. Best value for this amount.",
    "milestoneProgressPct": 75
  },
  "confidence": {
    "band": "high",
    "reasons": ["Direct merchant match", "Current milestone active"]
  }
}

Intended Use Cases

• Pre-Purchase Search: Search for "Amazon" or "Starbucks" before paying to see which card to pull out of your wallet.
• Large Purchases: Check which card has the highest remaining "Cap" before making a significant purchase (like electronics) to avoid losing out on rewards.
• Milestone Chasing: Use the "Strategies" section on the home screen to see which cards you should prioritize this week to hit your monthly or quarterly goals.