Imprint’s unicorn moment highlights the rise of AI-ready co-branded cards. Learn what it signals for fraud, routing, and modern payments infrastructure.
Imprint’s Unicorn Moment: AI-Ready Co-Branded Cards
Co-branded cards are back—and not as a nostalgia play.
A co-branded card provider reaching unicorn status is a signal that the “card as plastic” era is over. The modern prize isn’t the card itself; it’s the data exhaust, routing decisions, fraud signals, and lifecycle automation that sit behind every swipe and tap. Imprint hitting unicorn territory (as reported in industry coverage) fits a broader pattern: payments winners are building infrastructure that behaves like software, and software is increasingly optimized by AI.
This matters because co-branded programs used to be slow, bank-heavy projects with long timelines and blunt rewards. Now brands expect measurable ROI, fast iteration, and a tight feedback loop between customer behavior and payment performance. In the “AI in Payments & Fintech Infrastructure” series, I keep coming back to the same point: payments is becoming an optimization problem. Co-branded cards are a perfect case study.
Why co-branded cards are booming again (and what’s actually changed)
Co-branded cards are growing because brands want a direct line to customer spend and loyalty—and they want it without waiting a year to ship.
The classic co-branded model worked, but it was rigid: a brand, a bank, a processor, and a rewards engine stitched together with heavy operations. The new model looks more like a product platform. Providers like Imprint have been winning by focusing on:
- Speed to launch: compressing timelines from “many months” to something closer to software releases.
- Better economics: smarter onboarding, improved authorization rates, and more targeted rewards reduce wasted cost.
- Brand control: tighter integration into apps and commerce flows means the card behaves like part of the brand experience, not a generic financial product.
The seasonal reality: Q4 is when the cracks show
It’s December 2025. For payments teams, this is the high-stakes period: holiday peaks, returns, gift cards, elevated fraud, and customer service load.
Co-branded portfolios feel that pressure intensely. If approvals dip by even a couple points during a peak week, you’ll see it in revenue. If fraud controls are too strict, you’ll see it in angry customers. If they’re too loose, you’ll see it in losses. This is exactly where AI-optimized payments infrastructure becomes less of a “nice-to-have” and more of a survival trait.
The real product is the payments decisioning layer
A unicorn valuation for a co-branded card provider isn’t just about issuing cards. It’s about owning a decisioning layer that improves over time.
In a co-branded program, every transaction creates a decision moment:
- Should we approve this authorization?
- How confident are we it’s legitimate?
- Which controls should trigger (step-up, decline, notify)?
- What reward or message should follow?
- What should we learn for the next transaction?
Traditional stacks treat these as separate systems. Modern providers try to treat them as one coordinated loop.
Authorization rate is an underrated growth lever
Brands obsess over rewards and marketing, but authorization rate often delivers faster ROI.
A simple example: if a portfolio processes $500M in annual spend and improves approval rate by 1%, that’s potentially $5M more in captured volume (before you even talk about interchange, margin, or downstream retention effects). Even modest improvements matter.
AI can help here, but only if it’s paired with the right infrastructure: clean data models, real-time feature pipelines, and feedback that ties outcomes (chargebacks, disputes, returns) back to decisions.
Fraud isn’t one problem—it's several
Fraud is not just “stolen cards.” Co-branded programs face:
- Account takeover (especially around password resets and promo periods)
- Synthetic identity during application/onboarding
- First-party fraud (friendly fraud and abuse)
- Merchant dispute patterns that look like fraud but aren’t
AI-based fraud detection can reduce losses, but the goal isn’t “minimum fraud.” The goal is maximum good spend with controlled loss.
A payments risk system that declines legitimate customers at scale is just a revenue leak dressed up as security.
Where AI actually fits in a co-branded card stack
AI is most valuable in co-branded cards when it’s applied to decisions that repeat millions of times and compound.
Here’s the practical map—what I’d prioritize if I were building (or buying) this capability.
1) Smarter onboarding and underwriting
The best co-branded card programs don’t just approve more people; they approve the right people.
AI can support underwriting with:
- Alternative data signals (used responsibly and legally)
- Behavioral features from the brand’s own ecosystem (purchase frequency, tenure, return rates)
- Real-time identity verification outcomes feeding risk scoring
This tends to reduce two expensive failure modes: over-approving (losses) and under-approving (missed growth).
2) Real-time fraud + adaptive controls
Static rules don’t survive peak season. Fraud patterns change fast.
AI-based systems can:
- Adjust risk thresholds based on live attack patterns
- Recommend control actions (step-up vs decline) based on expected customer impact
- Detect merchant- and category-specific anomalies
The key is governance. You need auditable decisions, clear escalation paths, and well-defined “human override” controls.
3) AI-optimized routing (yes, even for cards)
People associate routing with ACH or payments orchestration. But cards have routing-like decisions too: retries, network options in some contexts, tokenization paths, and how you handle soft declines.
AI can improve:
- Soft decline recovery (when to retry, when to prompt the customer)
- Network token performance (token vs PAN behavior, lifecycle updates)
- Issuer/merchant messaging strategies that increase approvals
If you’ve ever watched a checkout conversion dip due to issuer behavior, you already know this is where infrastructure earns its keep.
4) Reward economics and personalization that don’t blow up the P&L
Rewards can become a slow-motion margin disaster if they aren’t tightly managed.
AI can help brands avoid “blanket generosity” by:
- Targeting incremental behavior (not rewarding what customers would’ve done anyway)
- Detecting reward abuse and arbitrage patterns
- Forecasting reward liability with more accuracy
Personalization should be constrained by unit economics. If the model can’t explain the margin impact, it shouldn’t ship.
What Imprint’s unicorn status signals about fintech infrastructure
A unicorn moment in co-branded cards signals that investors (and the market) value repeatable, scalable payments primitives.
Three signals stand out.
Signal #1: Brands want fintech outcomes, not fintech projects
The winners are packaging complexity into an outcome: launch faster, approve more good spend, reduce fraud loss, and keep customers happy.
Co-branded card providers are increasingly judged on measurable operational metrics:
- Time to launch
- Approval and decline rates (and reasons)
- Fraud loss rate and chargeback rate
- Activation rate and spend per active account
- Customer service contact rate per 1,000 accounts
When those metrics improve, valuation follows.
Signal #2: Payments data is now a competitive asset
The co-branded provider sits at the intersection of brand behavior and payments behavior. That’s a unique vantage point.
Done right, this data becomes an engine for:
- Better risk decisions
- Smarter lifecycle messaging
- Stronger retention and reactivation
- Faster iteration of benefits
This is why AI shows up naturally: the system has enough repetition and feedback to learn.
Signal #3: Compliance and trust are part of the product
At unicorn scale, “move fast” has a ceiling.
Modern co-branded platforms need strong controls around:
- Model governance and audit trails
- Fair lending considerations (where applicable)
- Data minimization and privacy-by-design
- Incident response and vendor risk management
Infrastructure providers that treat trust as a feature—rather than a legal checkbox—tend to last.
Building an AI-ready co-branded card program: a practical checklist
If you’re a brand, fintech PM, or payments leader evaluating a co-branded card platform, here’s what I’d ask for. Not a glossy deck—proof.
Platform questions that predict success
- Decision latency: What’s the end-to-end latency for authorization decisioning and fraud scoring?
- Feedback loops: How quickly do chargebacks, disputes, and returns feed back into models and rules?
- Explainability: Can you explain why a decision happened in plain language to compliance and support teams?
- Testing framework: Can you A/B test rewards, controls, and messaging without breaking compliance workflows?
- Data portability: If you switch providers, do you retain usable historical data and event logs?
Metrics to track in the first 90 days
- Application-to-approval rate (by segment)
- Activation rate within 7 and 30 days
- Approval rate at point-of-sale (overall and by top merchants)
- Fraud loss rate (with a breakdown by type)
- Dispute rate and average resolution time
- Incremental lift: spend and frequency vs comparable non-card customers
If a provider can’t commit to a measurement plan, they’re not selling a platform—they’re selling hope.
People also ask: what’s next for co-branded cards and AI?
Will AI replace rules-based fraud systems? Not fully. Rules are still essential for clear policy controls and regulatory clarity. The best stacks combine rules for guardrails and AI for pattern detection and adaptation.
Do co-branded cards still matter if wallets and BNPL keep growing? Yes, because co-branded cards plug into wallets and can coexist with pay-over-time options. The durable advantage is the account relationship and data continuity, not the form factor.
What’s the biggest risk in AI-driven payments infrastructure? Over-automation without governance. If your model decisions aren’t auditable and your teams can’t override quickly, you’ll create compliance and customer experience problems at scale.
Where this goes next in payments infrastructure
Imprint’s unicorn status is a milestone, but the more interesting story is what it represents: co-branded cards are becoming software-defined, data-driven payment products.
If you’re building in this space, don’t treat AI as a bolt-on feature. Treat it as a set of measurable capabilities: better approvals, controlled fraud, smarter routing, and rewards that pay for themselves. That’s how fintech infrastructure turns into durable advantage.
If you’re evaluating a co-branded card program for 2026 planning, ask yourself one forward-looking question: Which parts of your payments stack learn from outcomes—and which parts repeat the same mistakes every peak season?