Risk engines, fraud models & payment rails
Estimate your credit →Building risk models, fraud systems, payment infrastructure, and compliant core platforms requires experimentation to meet accuracy, latency, integrity, and security requirements.
The credit is calculated on qualified research expenses (QREs) — primarily wages for engineers and technical staff who spend time on qualifying activities, plus supplies, cloud/compute, and 65% of qualified contract research.
The table below lists the business components that most often qualify, a common example, and the technical uncertainty that makes each one defensible.
| Business component | Example | Technical uncertainty |
|---|---|---|
| Risk & underwriting engines | Credit or insurance risk model | Which model predicts risk accurately and fairly? |
| Fraud detection | Real-time fraud-scoring model | How to catch fraud while minimizing false positives? |
| Payment rails & infrastructure | Payment processing, ledger system | How to process reliably at scale and low latency? |
| Trading & pricing systems | Pricing engine or execution algorithm | Which approach prices and executes correctly? |
| Compliance & RegTech automation | KYC/AML automation | How to automate compliance accurately? |
| Core platform & ledger | Double-entry ledger, reconciliation engine | How to ensure consistency and integrity at scale? |
| Identity & data security | Identity verification, encryption | Which architecture is both secure and performant? |
| Embedded finance & APIs | Banking-as-a-service APIs | How to integrate reliably across partners? |
Potentially qualifying: Engineers compare model architectures and feature strategies to resolve whether fraud can be detected at required latency without excessive false positives, documenting backtests and controlled trials.
Generally not qualifying: Connecting a standard payment-provider API, changing report fields, or routinely adjusting a settled fraud threshold without technical uncertainty is implementation or maintenance—not experimentation.
Find your own job titles below. Each row shows the wage category a role usually falls under and the kinds of work that typically do and do not count. Titles alone never make wages eligible; the activity must meet the four-part test.
7 of 7 roles shown
| Role | What they do | Wage category | Typically qualifies | Typically does not |
|---|---|---|---|---|
| Builds the systems that move and reconcile transactions. | Direct research | Testing transaction consistency approaches, Uncertain reconciliation architecture +1 more | Standard gateway setup, Routine reconciliation +1 more | |
Typically qualifies
Typically does not
|
||||
| Builds and tests models that detect suspicious activity. | Direct research | Model architecture experiments, Resolving inference constraints +1 more | Routine model operation, Manual case reviews +1 more | |
Typically qualifies
Typically does not
|
||||
| Develops secure ways to authenticate users and protect transactions. | Direct research | Experimental authentication architecture, Encryption implementation alternatives +1 more | Compliance questionnaires, Routine patching +1 more | |
Typically qualifies
Typically does not
|
||||
| Makes high-volume financial data arrive accurately and on time. | Direct research | Real-time pipeline experiments, Uncertain consistency and recovery methods +1 more | Scheduled reporting, Dashboard upkeep +1 more | |
Typically qualifies
Typically does not
|
||||
| Implements and tests technically uncertain modeling systems. | Direct research | Experimental computational methods, Numerical stability testing +1 more | Economic forecasting alone, Investment strategy research +1 more | |
Typically qualifies
Typically does not
|
||||
| Builds test systems for new transaction and risk technology. | Direct support | Experimental load-test harnesses, Data preparation for qualifying trials +1 more | Manual regression of shipped features, Routine release signoff +1 more | |
Typically qualifies
Typically does not
|
||||
| Directly guides engineers through technical design alternatives. | Direct supervision | First-line review of qualifying designs, Direction of technical experiments +1 more | Fundraising, Compliance administration +1 more | |
Typically qualifies
Typically does not
|
||||
Wage categories follow IRC Section 41. Only the portion of time spent on qualified research counts, and the percentage must be documented.
Illustrative example using sample figures. Your actual credit depends on your facts; see Form 6765 and consult a tax professional.
The IRS requires qualifying research to satisfy four tests. Here's how they typically map for this industry:
The activity aims at improving a product, process, software, or technique used in the business — not just a business outcome.
The work relies on principles of engineering, computer science, biology, chemistry, or another hard science to resolve uncertainty.
There was genuine uncertainty at the outset about whether — or how — the component could be built or improved.
The team evaluated alternatives: A/B tests, prototypes, benchmarks, simulations, or iterative design-build-test cycles.
Take these into a 15-minute conversation with your team.
The last question matters most. Customer-funded work and work performed outside the US are generally excluded. Funded work is excluded to the extent your business does not bear the financial risk or does not retain substantial rights in the results.
No. A compliance goal can motivate development, but the work still needs qualifying technical uncertainty and experimentation. Policy writing and compliance reporting are different from developing an uncertain technical solution.
Potentially, when the development meets the research requirements and involves testing technical alternatives. Business scoring decisions, economic analysis and routine use of an existing model do not qualify by themselves.
An integration may involve qualifying work when the technical approach is uncertain and alternatives are tested. Connecting a standard vendor API using established instructions is not automatically research.
Software primarily used for internal general and administrative functions may face additional internal-use software requirements. Review the intended use and separate customer-facing components where appropriate. Review the IRS rules
Eligible computer-rental costs for qualified research and a portion of eligible U.S. contract research may count. General hosting, routine operations and vendor subscriptions are not automatically eligible.
Keep architecture decisions, model experiments, test results, pull requests and supported time allocations. Document the technical uncertainty rather than only the commercial or regulatory objective.
Most companies underestimate what they can claim. Our study covers every qualifying role, every component, and every year still open for amendment.
Get your free estimate →