How it Works

How it Works

3-step-process-icons-8

Upload user data

Pave maps multiple data sources to our standardized data model.

Our AI infrastructure ingests:

  • Bank aggregators
  • BaaS platforms
  • Payroll APIs
  • Debt management APIs
  • Credit bureau reports
  • Loan tapes 
3-step-process-icons-9

Get Cashflow Predictions

Pave processes billions of transactions in near real-time to surface signals including:

  • Affordability
  • Changes to income
  • Changes to a financial situation
  • and more…
Our models detect recurring behaviors and predict future events with high precision and recall.
3-step-process-icons-10

Power credit risk models

Supercharge your decisioning with Pave’s Scores and 450+ attributes offering a real-time view of a user’s affordability.

Lenders use these predictions for the following use cases:

  • Underwriting decisions
  • Lead scoring
  • Collections
  • and more…
mission and vission

The Problem

In an increasingly complex credit environment, lenders need to supplement credit bureau data with additional borrower context to understand risk in real-time. To put it in perspective,  it can take up to 90 days after a payment’s due date for a lender to report a missed payment.

By relying on lagging affordability indicators like FICO, lenders can quickly face losses by offering credit to people who can’t afford it – which also puts the borrower under more financial strain.

Lenders want to figure out:

  1. Can this person afford to make monthly payments of $X?
  2. Has anything changed recently about their ability to make their payments?
  3. When is the most optimal payment date to maximize successful collections?

Our Solution

We help to:

  • Detect a user’s historical income and expenses  
  • Predict future cashflow events including expected income and spending 
  • Predict the likelihood a user will be able to afford their payments based on the above

Drive critical credit decisions by unlocking a complete and real-time view of your consumers’ financial profile that typical credit reports fail to provide. 

Inaccurate income detection leads to massive losses

Raw transactions from data aggregators will often miscategorize Cash Advance deposits as “Payroll”. In the sample dataset of 2,843 users, an aggregator miscategorized over half a million dollars worth of cash advance deposits as “Payroll” or income.

Across the board, we’ve seen aggregators incorrectly record people’s income as double or triple what it actually is. Not analyzing this data correctly could mean providing someone with a cash advance amount that they can’t afford to pay back.

The Power of Granularity

Pave’s classifiers return granular and accurate tags to distinguish between different types of income, debt payments, and spending. In this sample dataset of 2,843 users, Pave accurately identified and tagged these same transactions as “Cash Advance” rather than payroll.

Granular tagging improves understanding of user behavior

Get unified insights into user behavior and affordability

Pave analyzes the transaction description, date, amount, and other factors to accurately distinguish between Brigit cash advance deposits, repayments, and subscription payments.  This way, lenders can better understand the nuances of a consumer’s risk.

Power credit decisions with cashflow intelligence

3 product offerings

1
Scores

Reduce defaults and NSFs

Use our pre-built Scores trained on billions of labeled datapoints:

  • Cash Advance Score
  • Likelihood to Repay Score
  • NSF/Overdraft Prediction Scores
  • Your own proprietary Score

2
Attributes Store

Drive lift in risk models

Use a growing number of signals in your models with our Attributes Store including:

  • Income history and stability
  • Debt payment behavior
  • NSF and balance trends
  • Changes in affordability

3
Cashflow API

Predict user cashflow

Transform raw bank transactions with our Cashflow API including:

  • Income detection and predicting next payroll
  • Bill and subscription detection and predicting next payment
  • Obligatory, essential, and non-essential categorization

Transform your data sources into attributes and scores with Pave.dev

Discover how to accelerate your model development from months to days

Benefits of using Pave

Faster time to market

Generate attributes and scores in near-real time via our APIs and turnkey access to Pave’s secure and standardized Snowflake data warehouse. Accelerate your development from quarters to days.

Unlock resources

Reduce the burden on your data science, data engineering, and infrastructure teams to process billions of data points in-house. Focus on proprietary data sources, models, and unique areas of value.   

Improve model performance

Leverage a growing number of Attributes and Scores including loan repayment behavior, affordability metrics, income stability, spend shocks, likelihood to repay, and more.  

Pave integrates with 20+ data source integrations and growing!

Check out our use cases for cash advance risk, collections risk, overdraft risk, lead scoring, and more.