A custom AML transaction monitoring system costs $100,000 to $500,000. Here is what it does, the modules it needs, and the build-vs-buy decision.
A custom AML transaction monitoring system usually costs $100,000 to $250,000 for a rules and case-management module tied to one market, and $250,000 to $500,000 or more for a standalone platform with machine learning, sanctions screening, and regulator-ready reporting. You are building three things at once: a rules engine that flags suspicious activity, a case workflow for investigators, and an audit trail a regulator can actually read. Here is what that involves and where the money goes.
A transaction monitoring system watches every payment, deposit, and transfer that moves through your product and decides which ones need a human look. It does not stop money laundering by itself; it produces alerts that a compliance team investigates, documents, and sometimes reports to authorities as a suspicious activity report (SAR) or suspicious transaction report (STR).
The core loop has four steps. First, the system pulls transactions from your ledger, core banking system, or payment rails. Second, it runs each one against rules and behavioral models. Third, alerts that clear a risk score are routed to investigators in a case queue. Fourth, investigators either close the alert with a reason or escalate it, and every action is logged.
The rules are the hard part. A scenario like "cash deposits just below the $10,000 reporting threshold across multiple branches in one day" catches structuring. "Rapid movement of funds in and out within 24 hours" catches layering. "A dormant account suddenly receiving a large transfer" is another standard one. Most systems ship with a library of 50 to 200 predefined scenarios that you tune to your own risk appetite.
| Module | What it does | Why it costs money |
|---|---|---|
| Data ingestion | Pulls transactions from ledgers, payment rails, or core banking APIs | Normalizing messy data is most of the work |
| Rules engine | Runs scenarios (structuring, layering, rapid in/out) | Rule tuning and thresholds need compliance input |
| Alert scoring | Ranks alerts so analysts see the riskiest first | Model tuning and backtesting |
| Case management | Queue, investigation notes, decision trail | Workflow and permissions for auditors |
| Screening | Sanctions, PEP, and adverse-media checks | Depends on a data provider |
| Reporting | SAR/STR filing, regulator extracts, audit logs | Format varies by jurisdiction |
The single biggest cost driver is false positives. Industry data puts false-positive rates for rule-based monitoring above 90% in many banks, which means every alert a model eliminates saves analyst time directly. That is why a large share of the engineering budget goes into tuning thresholds and backtesting scenarios before go-live.
These are planning ranges, not quotes. The number moves with transaction volume, the number of markets, and whether you integrate a third-party screening provider or build the detection logic yourself.
| Scope | Typical range | Timeline |
|---|---|---|
| Rules + alerting + case management for one market | $100,000 to $250,000 | 3 to 5 months |
| Standalone platform with ML, screening, and reporting | $250,000 to $500,000+ | 5 to 9 months |
| ML behavioral layer on top of an existing rules engine | $50,000 to $150,000 | 2 to 4 months |
For context, US financial institutions spent an estimated $59 billion on BSA/AML compliance in 2023, and nearly all report that costs keep rising. The people side is expensive too: a BSA officer averages about $126,000 and a compliance analyst about $97,000 a year in the US. A system that cuts even a third of manual alert reviews pays for itself quickly.
The realistic choice is rarely "build everything from nothing." Most teams either license a monitoring platform such as Unit21, Hawk AI, ComplyAdvantage, AMLYZE, or Sardine and configure it, or build a custom engine on top of their own ledger.
License a platform when your volumes are standard, your markets are covered, and you want to be live in weeks. You pay a setup fee plus a per-transaction or per-alert fee, and you hand over control of the detection logic. This works until you hit a market or a product the vendor does not model well.
Build custom when the rules are your product's edge, when you operate in markets the vendors under-serve (common in Africa, Latin America, and parts of Asia), or when you already run your own ledger and need the monitoring to live inside it. A custom engine keeps your detection logic private and lets you backtest it against your own transaction history. The trade-off is time: a credible build is three to six months, and you own model validation afterwards.
We have built financial systems across Brazil, South Africa, Singapore, and the US, and we treat transaction monitoring as part of the core platform rather than a bolt-on. In practice that means the monitoring rules read from the same event stream as the ledger, so nothing can be switched off silently. We start with a short discovery phase that fixes the scenario list, the risk thresholds, and the reporting formats for your regulator before any code is written.
Most builds we do use Java or Go for the rules engine, Redis for in-memory thresholds and counters, and a relational store for the case and audit trail. Sanctions and PEP screening integrate through a provider API so the data stays current. The result is a system where every alert, decision, and report is reproducible for an audit.
If you are planning a FinTech platform and want transaction monitoring designed in from day one rather than retrofitted, talk to us about your project. We can help you scope the rules, the reporting, and a realistic budget.
It is the software that watches transactions for patterns that suggest money laundering, flags them, and routes them to investigators for review and possible SAR/STR filing. It combines a rules engine, alert prioritization, case management, and reporting.
As a planning range, a rules and case-management module for one market runs $100,000 to $250,000, and a standalone platform with machine learning and screening runs $250,000 to $500,000 or more. Licensed platforms charge a setup fee plus per-transaction or per-alert fees instead.
Structuring (deposits just under a reporting threshold), smurfing (many small deposits across accounts), rapid in-and-out movement, dormant accounts receiving large transfers, and transfers to high-risk jurisdictions. Systems usually ship with 50 to 200 scenarios that you tune.
The FATF Recommendations set the global baseline, and each jurisdiction implements them its own way: FinCEN BSA/AML rules in the US, the AMLA and AMLD framework in the EU, plus the Travel Rule for originator and beneficiary information. Check your own regulator's specific thresholds.
Buy a licensed platform if your volumes are standard and you want to be live fast. Build custom if the detection logic is part of your product, you operate in under-served markets, or you run your own ledger and need monitoring inside it.
A rules-plus-case-management module for a single market typically takes three to five months. A full standalone platform with machine learning, screening, and reporting takes five to nine months, including backtesting and model validation.
GlobeSoft builds custom software, AI applications, SaaS, FinTech and enterprise systems for clients worldwide. Message us on WhatsApp for the fastest answer, or email us — we typically reply within 24 hours.
Keep reading
What a software development retainer is, how monthly retainer pricing works, typical 2026 costs, and when a retainer beats project-based billing.
7 min readHow Pix payment integration works for businesses selling in Brazil: gateway vs direct routes, fees of 0.22% to 0.99%, Pix Cobranca and Automatico, and a build checklist.
7 min readWhat mobile money integration costs, the M-Pesa, MTN MoMo, Flutterwave and Paystack routes, timelines, and the licensing and reconciliation that matter.
10 min readWhat white-label software development really involves, how a commissioned build works, what it costs, and how to choose between rebranding and building.
10 min readA practical playbook for working with a China-based software team: communication rhythm, contracts, IP, quality checks, and the mistakes that sink projects.
5 min readOffshore development has a bad reputation that half of it does not deserve. What it costs, what goes wrong, and how to run it so it actually works.
