Synchrony taps AI to refine auto credit decisioning as metrics stabilize

Synchrony is increasing its use of AI-driven models in auto underwriting now that its auto credit metrics have stopped deteriorating and are showing signs of stabilization.

What the headline means

• Synchrony’s auto loan/lease portfolio went through a period of rising delinquencies and charge‑offs, but those metrics (e.g., roll rates, delinquencies, loss rates) are now leveling out rather than worsening.

• With that stabilization, management is more comfortable letting AI and machine‑learning models play a larger role in credit decisioning for auto, versus leaning as hard on conservative overlays and traditional score‑cut strategies.

How Synchrony is using AI

• Synchrony has built a next‑gen decisioning platform called PRISM, which uses thousands of data signals (credit bureau, cash‑flow, rent payments, partner data, fraud indicators) to make a credit decision in roughly six seconds.

• The same PRISM framework that’s been used heavily in cards is being applied more aggressively to auto, with AI/ML models helping to segment risk, identify underserved but creditworthy borrowers, and adjust cut‑offs dynamically as performance improves.

Why “metrics stabilize” matters

• When credit performance is under pressure and metrics are worsening, lenders typically tighten: higher score floors, less reliance on experimental models, and more manual or policy overlays.

• Once performance stabilizes, a lender like Synchrony has more room to:

  • Relax some overlays and let model‑recommended approvals flow through

  • Expand near‑prime/underserved segments where AI suggests attractive risk‑adjusted returns

  • Test new model features (e.g., more cash‑flow or alternative data) without materially compromising portfolio quality.

Strategic implications for auto finance

• For dealers and OEM partners, more AI‑driven auto decisioning should mean faster responses, more automated approvals at the margin, and potentially higher capture rates at the point of sale.

• For regulators and risk teams, it raises all the usual questions: model risk management, explainability, fair lending and adverse action logic, and governance around alternative data and dynamic decision trees.

• For competitors, the move signals that Synchrony believes its data and infrastructure advantages (PRISM, private cloud, real‑time ML platform) are mature enough to extend beyond cards into auto at scale.