OpenAI Buys Roi: What Finance Teams Should Do Next
OpenAI has acquired AI-powered personal finance app Roi. Roi's co-founder and CEO Sujith Vishwajith said the team built the app to make investing more personal, and that personalization is the future of software. According to reports, he will be the only member of the four-person team moving to OpenAI. Terms were not disclosed, and Roi will wind down client service on Oct. 15.
This move fits a bigger trend: banks are moving from menu-based UX to systems that infer intent and suggest next-best actions. As AI sits on top of permissioned transaction data, the experience shifts from reactive service to proactive guidance. Research shows 75% of bank customers want more personalization, and embedded conversational AI could win back 72% of them.
Why this matters: Personalization is becoming the default in banking
Finance leaders should expect more M&A, acqui-hiring, and partnerships around data, recommendation engines, and conversational interfaces. The play is clear: combine high-quality, consented data with models that can infer context and offer precise next steps in real time. That raises the bar for data governance, model risk, and explainability.
- Data foundation: Centralize permissioned data (transactions, balances, linked accounts) and establish a feature store for real-time signals.
- Consent and privacy: Make data permissions explicit, auditable, and revocable. Map data lineage across model inputs and outputs.
- Next-best-action engine: Deploy intent detection and ranking that accounts for liquidity, risk, and customer preferences.
- Controls: Put model governance, bias testing, and human-in-the-loop review in place for sensitive decisions.
- Distribution: Bring the same AI core to app, web, call center, and RM tools so guidance is consistent across channels.
Consumer credit: Cooling, not cracking
August consumer credit barely grew at a 0.1% annualized pace. Revolving credit contracted 5.5% annualized, the steepest pullback this year, while nonrevolving rose 2%. Households still have capacity: the 75th percentile of active card accounts used roughly 50.7% of available lines in Q1 2025.
Expectations are shifting. One-year-ahead inflation expectations rose to 3.4%, while expected earnings growth slipped to 2.4%. Job-loss probability ticked up to 14.9%, and the probability of finding a job in three months was 47.4%. The perceived likelihood of missing a minimum payment eased to 12.6%, a sign of deliberate balance management.
- For lenders: Anticipate softer revolving growth and more selective card spend. Emphasize liquidity planning, installment options for essentials, and proactive fraud alerts.
- Risk posture: Tighten affordability checks, monitor early-stage delinquencies, and adjust line management to preserve healthy utilization.
- Offers and timing: Shift from broad campaigns to context-aware prompts (payment plans, targeted APR reductions, pre-approved term loans) when intent signals fire.
- Employee enablement: Equip bankers and RMs with AI-assisted answers and client context so guidance is fast and consistent.
Action plan for the next 90 days
- 30 days: Inventory data sources and permissions. Define top 10 intents (e.g., "can I afford X?", "optimize cash flow," "reduce interest"). Prioritize 3-5 next-best actions you can operationalize now.
- 60 days: Launch a controlled pilot that pairs a conversational interface with policy-constrained recommendations. Add event triggers (paycheck, large debit, nearing due date).
- 90 days: Roll out to one high-value segment. Measure acceptance rate, financial outcomes (utilization, payment rate), and customer satisfaction. Expand only with clear guardrails.
KPIs to keep front and center
- Utilization bands and changes at the customer level
- Payment rate, early-stage DQs, and cure rates
- Offer acceptance and outcomes by intent
- Time-to-answer for employee tools and accuracy vs. policy
- Privacy complaints and consent revocations
If you need source data for planning, see the Federal Reserve's consumer credit release and the New York Fed's Survey of Consumer Expectations:
Upskill your team
Building conversational guidance and next-best-action workflows requires product, risk, data, and engineering working off the same playbook. For a practical overview of tools used in finance, this resource can help:
Bottom line: OpenAI's move underscores the shift to personal, proactive finance. Consumers are cautious yet capable, and they'll reward firms that offer timely, context-aware guidance with clear guardrails.
Your membership also unlocks:
