Communication Federal Credit Union goes live with Scienaptic AI credit decisioning

Source: BusinessWire — 2026-07-20

Summary

Communication Federal Credit Union, a $2.3 billion-asset lender based in Oklahoma City, has gone live on Scienaptic AI's credit-decisioning platform to automate underwriting and expand fair-access lending for its members. The deployment runs on Scienaptic's iCUE (Intelligent Credit Underwriting Experience), which folds large language models and agentic AI into the decisioning engine to deliver faster, more consistent loan decisions. Scienaptic projects the rollout will unlock $134.5 million in incremental vehicle-loan originations while cutting losses across the credit union's consumer-loan portfolios by up to 20%.

Key Takeaways

• Communication Federal Credit Union ($2.3B in assets, Oklahoma City) is now live on Scienaptic AI's platform, automating credit-decisioning workflows that previously required manual underwriting review.
• Projected impact: $134.5 million in incremental vehicle-loan originations and up to a 20% reduction in consumer-loan losses.
• The platform runs on iCUE (Intelligent Credit Underwriting Experience), Scienaptic's newest product combining LLMs and agentic AI with predictive credit models.
• Billy McDaniel, SVP/CLO at Communication FCU, said the system "is automating our decisioning workflows, allowing us to deliver instant, consistent loan decisions while maintaining the prudent risk management our members trust."
• The stated goal beyond speed is fair-access lending — using richer, AI-driven risk assessment to approve more creditworthy members who might be declined under older scorecard-only rules.

Reel Script

This item has real projected metrics and a diagrammable credit-decisioning workflow — likely reel_eligible: true.

Hook (~18s, 40 words) A member walks into a credit union for a car loan. Old system: days of manual review, maybe a flat "no" from a rigid scorecard. New system: an instant decision — and a projected $134.5 million more in loans getting approved that used to get rejected.

Core Concept (~75s, 165 words) Here's what's actually happening under the hood. Traditional credit decisioning leans on a handful of hard-coded rules — credit score cutoffs, debt ratios, maybe five or ten variables — and if you don't clear the bar, you're declined, full stop, no matter the fuller picture. Scienaptic's platform, called iCUE, replaces that gate with a much richer model that pulls in way more signal — transaction patterns, alternative data, behavior over time — and layers large language models and agentic AI on top so the system can actually reason through edge cases instead of just checking boxes. Think of it like the difference between a bouncer with a strict dress code versus one who actually looks at the whole person before deciding. The "agentic" part matters too — it's not just scoring, it's automating the workflow end to end, so a decision that took a loan officer days now comes back in seconds, consistently, every time.

Hands-On (~85s, 190 words) So what did Communication Federal Credit Union actually do here? This is a $2.3 billion-asset credit union out of Oklahoma City, and they just flipped the switch on Scienaptic's platform across their consumer lending operation. The numbers Scienaptic is putting behind this launch are specific, not vague marketing fluff: they're projecting $134.5 million in additional vehicle-loan originations that wouldn't have happened under the old process — that's loans to members who'd have been turned away or never even applied because approval felt like a coin flip. On the flip side, they're projecting up to a 20% cut in losses across the consumer-loan portfolio, meaning the AI isn't just approving more people loosely — it's supposed to be better at spotting the risky applications too, not worse. The credit union's SVP and chief lending officer framed it as keeping "prudent risk management" intact while making decisions instant. That's the pitch in a nutshell: grow the loan book and cut losses at the same time, which under a static scorecard system is usually a trade-off, not a twofer.

Takeaway (~25s, 58 words) Verdict: this is a real production deployment with real projected numbers attached, not a pilot press release. Whether Scienaptic hits $134.5 million and a 20% loss reduction is something to check back on in a year — but the mechanism, richer signal plus agentic automation replacing rigid scorecards, is the actual trend worth tracking in lending right now.

Discussion

(No questions yet — ask follow-ups via a Claude Code chat session on this repo; answers get appended here.)