The One-Line Truth
Casap replaces the fragmented spreadsheet-and-processor patchwork that banks and credit unions use to resolve payment disputes by running the entire lifecycle -- intake, investigation, provisional credit, chargeback filing, and member communication -- through AI agents in a single system, while scoring every claim for first-party fraud before the institution absorbs the loss.
The Role: VP of Card Operations, Head of Disputes, Chief Strategy Officer at a credit union or regional bank Founded: 2023 (incorporated December 2022) | HQ: New York City (product and engineering in San Francisco) | Funding: $33.5M total ($25M Series A led by Emergence Capital, August 2025) Founders: Shanthi Shanmugam (CEO, ex-Robinhood Group PM -- launched 24/7 support and crypto trading for 20M members) and Saisi Peter (Co-Founder, ex-Chime -- built tools supporting 1,200+ agents, saved $10M through automation)
The Disruption Connection
In December, The Heed Report showed that back-office financial operations remain one of the most heavily regulated and least automated functions in business technology. Casap is the response built specifically for card issuers.
Dispute resolution sits at the intersection of three forces pressing on every regional financial institution simultaneously: federal compliance timelines that do not flex, consumer expectations that have moved to real-time, and a fraud vector -- first-party abuse -- that exploits the very process designed to protect honest cardholders. The tools institutions use to manage this intersection were built before any of those pressures existed at their current intensity.
The Problem It Kills
Payment disputes at community banks and credit unions follow a workflow that has barely changed in two decades.
When a cardholder calls to dispute a transaction, a staff member manually enters claim details into the institution's system. Posting a provisional credit takes approximately 20 minutes of labor per case. Once filed, up to 70% of smaller institutions outsource the claim to a third-party processor, paying flat fees of $20 to $40 per case. The claim then enters what John Mays, Card Services Manager at MidSouth Community FCU, described as "a black hole" -- no real-time updates, no visibility into status, and no control over resolution.
Under Regulation E (covering debit and ATM transactions) and Regulation Z (covering credit card disputes), issuers face strict statutory timelines. A missed deadline is not a service failure -- it is a regulatory violation. The 90-day resolution window that processors typically consume is not a technical limitation; it is the maximum the regulation allows.
Meanwhile, first-party fraud -- where cardholders dispute legitimate purchases to secure unauthorized refunds -- now accounts for 30% to 50% of total fraud losses at financial institutions, costing the industry an estimated $100 billion annually. Because flat processor fees make investigating low-value claims economically unviable, many institutions set high write-off thresholds, absorbing smaller losses automatically. That vulnerability is precisely what serial disputers exploit.
The operational consequence: the same staff handling legitimate member disputes is also absorbing fraudulent ones, with no systematic way to distinguish between the two before the money leaves the institution.
Who This Is For / Who Should Skip It
Build with this if: You are a bank, credit union, or fintech issuer processing at least 200 dispute cases per month and want to bring disputes in-house, eliminate third-party processor fees, and reduce fraud losses. Casap is designed for institutions with assets ranging from $200 million to $20 billion, though it also serves larger public fintechs and top-25 banks. The sweet spot is any institution where a small team handles a high volume of card disputes and cannot hire additional headcount.
Skip this if: You are on the merchant/acquirer side of disputes (Casap is built exclusively for card issuers, not merchants defending against chargebacks). Skip if your dispute volume is under 200 cases per month -- the ROI math requires sufficient case volume to offset platform costs. Skip if you need a general-purpose fraud monitoring platform -- Casap automates post-transaction dispute operations, not pre-transaction fraud detection. Pair it with an enterprise fraud monitoring system (Featurespace, NICE Actimize, Verafin) rather than replacing one.
How It Actually Works
Minute 1. Casap connects to your core banking system and digital banking platform through an API integration layer. The system maps your existing card network relationships (Visa, Mastercard) and loads your regulatory profile -- which Regulation E and Z timelines apply, what your current write-off thresholds are, and how your provisional credit policies work. The intake interface replaces the manual data entry your team currently performs.
First Hour. A cardholder disputes a transaction. The AI agent handles intake by pulling the cardholder's identity, transaction metadata, and card details automatically. Before opening a formal dispute, the system runs transaction proximity analysis: if the cardholder has a history of successful, undisputed purchases with the same merchant, the platform prompts them to contact the merchant directly, deflecting unnecessary chargebacks. If the dispute proceeds, the system calculates a predictive win score and a first-party fraud score simultaneously. Provisional credits post in seconds instead of 20 minutes. The entire claim -- from intake through chargeback filing and member communication -- lives in one system.
First Week. Your team sees a real-time dashboard showing every active dispute, its stage in the regulatory timeline, the predicted outcome, and the fraud risk score. Staff who previously spent their days on manual data entry and status checking now focus on the cases the AI flags for human review -- high-risk fraud indicators, edge cases where the evidence is ambiguous, and the small percentage of disputes that require judgment. The system files chargebacks directly with card networks (no third-party processor queue), tracks merchant responses, and generates compliant notifications automatically. MidSouth Community FCU eliminated their processor fees entirely in the first week.
The Features That Matter
AI-Powered Dispute Agents. The system manages the full dispute lifecycle end-to-end: intake, triage, investigation, decisioning, chargeback filing, and member communication. Unlike workflow management tools that organize the manual steps, these agents execute them. The gotcha: complex edge cases still require human review, and the system is designed to surface those rather than over-automate them.
First-Party Fraud Scoring. A proprietary risk score -- described by the company as analogous to a FICO score for dispute fraud -- that evaluates every claim before the institution absorbs the loss. The engine ingests cardholder IP addresses, device signatures, linked phone numbers, dispute filing frequency, and cross-merchant behavioral patterns. Serial disputers get flagged; honest cardholders get instant resolution. The limitation: the score improves with data volume, so newer institutions see less precision initially.
Direct Network Integration. Casap files chargebacks directly with Visa and Mastercard, bypassing the third-party processor queue entirely. This eliminates the $20 to $40 per-case processor fees, gives the institution real-time visibility into chargeback status and merchant responses, and provides access to Visa transaction metadata (IP addresses, phone numbers) that the third-party model never exposed.
Regulatory Compliance Engine. Regulation E, Regulation Z, Nacha guidelines, and card network rules are embedded in the routing logic. The system calculates statutory deadlines, schedules provisional credit postings, and tracks investigation timelines automatically. This is not compliance reporting -- it is compliance execution. The gap: compliance rules vary by state and institution type, so the initial configuration requires careful mapping.
Predictive Win Modeling. The system evaluates each dispute against historical network outcomes to calculate a win probability. High-probability cases proceed automatically; low-probability cases are flagged so staff can decide whether to invest the administrative effort. This prevents institutions from wasting resources on cases they are statistically likely to lose.
Automated Member Communication. Throughout the dispute lifecycle, the system sends automated status updates, evidence summaries, and resolution notifications to cardholders. This eliminates the follow-up calls that consume staff time and frustrate members waiting in the dark.
Real Cost
Casap does not publish standard pricing tiers on its website. The platform operates on an institutional contract model, with pricing scaled to dispute volume and institution size.
The ROI math from published case studies provides the clearest cost picture:
Chartway FCU (a ~$3.2B-asset credit union) achieved an estimated $875,000 in net savings during its first year with Casap. Per-claim processing costs dropped by approximately 90%. The credit union eliminated third-party processor fees entirely and reduced write-offs by 72%. MidSouth Community FCU saw positive ROI within weeks of deployment.
The cost baseline Casap replaces: $20 to $40 per-case processor fees, 20 minutes of manual staff labor per provisional credit, and the fraud losses absorbed through high write-off thresholds. For an institution processing 500 disputes per month at $30 per case in processor fees alone, that is $180,000 per year in fees before counting staff time or fraud losses.
The Filene Research Institute's FiLab program independently validated these economics across a pilot with multiple credit unions, with published results confirming the operational savings claims.
What Customers Say
Rob Keatts, Chief Strategy Officer at Chartway FCU, in the company's Series A announcement: the platform was the most meaningful operational upgrade to their claims process in decades, replacing a patchwork of spreadsheets, emails, and manual tracking with real-time visibility. The credit union can now resolve issues faster without relying on third parties or worrying about missed deadlines.
John Mays, Card Services Manager at MidSouth Community FCU, in a published case study: the biggest frustration before Casap was the complete lack of visibility once a dispute left the team's hands. After running an ROI analysis before switching, Mays found the platform exceeded even his most optimistic projections.
The FiLab independent evaluation across multiple credit unions found that staff reported a 122% improvement in ease-of-use compared to previous systems, error frequency dropped by 63%, overall satisfaction averaged 4.8 out of 5, and 93% of employees said the platform made their job easier.
The complaint patterns are harder to surface -- Casap's customer base is concentrated in credit unions and regional banks, which produce less public review content than enterprise SaaS. No G2, Capterra, or TrustRadius reviews were found at the time of research. The independent FiLab evaluation is the closest thing to an unfiltered signal, and it is strongly positive.
The Competitive Read
Legacy core processors and third-party outsourcers remain the primary alternative for most institutions. They handle dispute settlement but offer no real-time tracking, no fraud scoring, and no compliance automation. Casap's advantage is not incremental -- it is architectural. The disadvantage: switching from a processor requires the institution to take operational ownership, which demands internal commitment.
Acquirer-side chargeback platforms (Chargeback Gurus, Riskified) serve the opposite side of the table. These are built for merchants defending against chargebacks. Casap is built for card issuers managing the disputes. They are complements, not competitors.
Enterprise fraud monitoring platforms (Featurespace, NICE Actimize, Verafin) detect suspicious transactions and generate alerts. They do not automate the dispute resolution workflow itself -- posting credits, filing chargebacks, tracking regulatory deadlines, or communicating with members. Casap acts as the execution layer that picks up where monitoring platforms stop. Pair them, do not choose between them.
Where competitors are actually better: Enterprise fraud platforms offer broader coverage across pre-transaction and cross-channel fraud detection. Casap is narrowly focused on post-transaction dispute operations and first-party fraud. If your primary problem is account takeover or payment fraud before the transaction clears, you need a fraud monitoring platform first and Casap second.
The Honest Verdict
Excellent for: Credit unions and regional banks with growing dispute volumes and shrinking staff. The platform is purpose-built for the institution that cannot hire another three people to handle the caseload increase, needs to bring disputes in-house for cost and control reasons, and wants to stop absorbing first-party fraud losses because investigating small claims is too expensive under the current model. The independent FiLab validation and the Chartway results ($875K first-year savings, 95% chargeback win rate, 72% write-off reduction) make this one of the most empirically validated tools in the entire 30-day series.
Breaks at: Institutions with very low dispute volume (under 200 cases per month) may not generate enough ROI to justify the switch. The platform's AI scoring improves with data volume, so early-stage deployments see less precision than mature ones. Integration with legacy core banking systems requires real technical effort -- this is not a plug-and-play widget. And the company is still young (founded 2023, approximately 44 employees) -- institutions that require multi-year vendor stability guarantees may hesitate.
Trajectory: Named to the CB Insights 2026 AI 100. 642% year-over-year growth. 45 financial institutions and expanding rapidly through CUNA Strategic Services, Q2 Marketplace, and state credit union league partnerships. The $25M Series A at a reported $105M valuation (per TipRanks) positions the company to expand beyond card disputes into account takeover investigations, deposit reviews, and cross-institution risk intelligence. The hiring signals -- VP of Engineering at $290K to $360K, Disputes and AI Operations Leaders -- indicate a company building deep technical infrastructure, not chasing quick feature additions. Emergence Capital's track record (Salesforce, Veeva, Zoom) suggests patient, infrastructure-grade capital allocation. If the Chartway and MidSouth results replicate at scale, Casap has the potential to become default infrastructure for every mid-tier card issuer in the United States.
Set It Up with AI
Dispute Operations Assessment Prompt: "I run dispute operations at a [credit union / community bank] with [X] in assets and approximately [X] card dispute cases per month. Our current process involves [manual entry / third-party processor / hybrid]. Walk me through the operational steps I would need to evaluate whether bringing disputes in-house with an AI-powered platform would reduce our per-claim costs, improve our chargeback win rate, and help us comply with Regulation E and Z timelines. Include a framework for calculating current cost-per-dispute including staff time, processor fees, and fraud write-offs."
First-Party Fraud Audit Prompt: "Analyze our dispute data for first-party fraud patterns. I will provide a sample of [X] recent dispute cases with transaction amounts, dispute reasons, cardholder history, and outcomes. For each case, flag indicators of potential first-party fraud: repeat disputers, disputes on merchants where the cardholder has a history of successful purchases, disputes filed shortly after delivery confirmation, and patterns of disputing then re-purchasing from the same merchant. Calculate what percentage of our disputes show first-party fraud indicators and estimate the annual dollar impact."
Regulatory Compliance Mapping Prompt: "Create a compliance checklist for card dispute resolution at a federally-chartered credit union. Map the specific requirements of Regulation E (electronic fund transfers, debit card disputes) and Regulation Z (credit card billing disputes) including: provisional credit posting deadlines, investigation completion timelines, written notification requirements, and the conditions under which provisional credits can be revoked. For each requirement, identify where our current manual process creates compliance risk and where automation could eliminate deadline exposure."
Vendor Evaluation Prompt: "I am evaluating AI-powered dispute automation platforms for a credit union with $[X]B in assets. Build a structured evaluation framework with weighted criteria across: regulatory compliance coverage (Reg E, Reg Z, Nacha, card network rules), integration requirements with [our core system], first-party fraud detection capabilities, direct network filing (Visa/Mastercard), implementation timeline and resource requirements, pricing model transparency, and independent validation or case study evidence. Include questions I should ask each vendor during demos that would reveal gaps in their compliance automation."
Sources
Independent / Third-Party
- 2025 FiLab Results: Casap -- Filene Research Institute
- Chartway Credit Union Saved an Estimated $875K in One Year -- McKaye Black, Filene Research Institute
- AI fraud and disputes platform Casap raises $25m -- FinTech Global
- AI Startup Casap Raises $25 Million to Fight First-Party Fraud -- PYMNTS.com
- Casap Named to the 2026 CB Insights AI 100 -- PR Newswire (CB Insights announcement)
- FinovateFall 2025 Best of Show Winners Announced -- David Penn, Finovate
- Women in Technology: Casap's Shanthi Shanmugam -- W.B. King, Finopotamus
Customer-Attributed
- How MidSouth Cut Fraud Loss in Half with Casap -- Casap (John Mays, MidSouth Community FCU)
- Casap Raises $25M Series A -- BusinessWire (Rob Keatts, Chartway FCU)
Company-Owned / Investor Sources
- A Pioneering Approach to Fixing First-Party Fraud -- Carlotta Siniscalco, Emergence Capital
- Investing in Win-Win-Win Companies in Fintech -- Lightspeed Venture Partners
- Solving the $100B First-Party Fraud Problem -- Primary Venture Capital
Day 30 of 30. This is the final entry in the 30 Tools in 30 Days series. Thirty AI-native companies. Six business layers. Every one founded after ChatGPT. The full series is available at 30 Tools in 30 Days.