Blog · Published · Updated
AI Agent for Compliance Monitoring: Analyst-Reviewed Triage
Short answer
In compliance monitoring, an AI agent prepares an alert for the analyst's decision: it gathers the client's documents, the transaction history and earlier findings, points out gaps and writes a note with sources. A person assesses the risk, decides and files any report. The agent costs from €3,000 net at Syntalith and, with 800 alerts a month at 15 minutes each, pays back in 0.8 months.
How we workDetails ↓
- 01Call30 minutes with an engineer, free
- 02PrototypeA working system on your process and data, free
- 03ImplementationTarget and fixed price in the contract; 50% at signing
- 04Acceptance50% once the agreed target is met
- net per agent for one process, build paid once
- from €3,000
- from signing the contract to acceptance
- 6–16 weeks
- free prototype on your cases
- €0
We belong to the OpenAI and Anthropic partner programmes and learn through their official courses. What this means ↓
Author: Syntalith · 3 min read
Statistics
Finance, compliance and AI: statistics 2026
0.8 mo.
payback of a €3,000 agent with 800 alerts a month: €6,980 of work today, €3,214 with checking and the model. Scenario.
90%
According to the European Commission (2026), 90% of companies that use AI say people take the decisions, and 2% of decisions are fully automated.
Source: European Commission, Flash Eurobarometer on AI and algorithmic management (2026) ↗
52.5%
According to Eurostat (2025), 52.5% of EU enterprises that considered AI and did not use it named a lack of clarity about the legal consequences.
Source: Eurostat, Use of artificial intelligence in enterprises (2025) ↗
40%
According to Gartner (2025), over 40% of agentic AI projects will be cancelled by the end of 2027 because of costs, unclear value or weak risk controls.
14–56%
According to Anthropic (2025), controlled trials found smaller time savings, from 14% to 56%.
Source: Anthropic, Estimating AI productivity gains (2025) ↗
Figures checked: 22 October 2026
A complete set of documents is not a basis for a decision
A client's transaction crosses a threshold and the system raises an alert. The analyst opens several systems: the transaction history, the client's documents, earlier alerts and notes. In Poland, an obliged institution applies a risk-based approach and customer due diligence measures under the AML Act, and the Polish financial intelligence unit GIIF describes the practical side in communication no. 22.
The agent gathers these materials on the alert and writes a note: what triggered the alert, what is known about the client, which documents are current and which are missing. A complete set of documents does not mean there is a basis for a decision: the contract and the statement may be in place while the source of funds is still unexplained. The agent marks this as a gap, and the analyst decides on the next steps.
- 01Alertan event from the transaction monitoring system
- 02Materialshistory, documents, earlier alerts and notes
- 03Notefindings with sources and a list of gaps
- 04Decisionthe analyst assesses the risk, closes or escalates
- 05Recordthe decision, the reasoning and the version of the note in the history
LangGraph describes interrupting a process for a person's decision and resuming after it. The agent has no tool to close an alert or file a report, so it cannot do either even under a wrong instruction.
What it costs and when it pays back
- Today
- 800 alerts × 15 min = 200 h (€6,980)
- With the system, a person checks the output in 45% of the time
- €3,141
- Model, 3 calls per case, Claude Sonnet 5.5
- €73
- Cost of one case
- €8.73 today, €4.02 with the system
- Saving
- €3,766 a month
Payback of €3,000
0.8 months
The volume, the time and the checking share are the scenario's assumptions. €34.90 an hour is the EU average hourly labour cost for 2025 according to Eurostat. Model: 12,000 input and 1,000 output tokens per call at Anthropic's rates, converted at the ECB rate of 1.1205 USD per euro on 22 October 2026. Checking takes 45% of the time because the decision on every alert stays with the analyst.
The count covers the work around a case: searching, reading, writing and passing things on. What remains is the time to check the output, 45% of today's work in the scenario. Before the quote we measure both times on 20 to 100 cases the team has already handled, in a €0 test that takes 5 working days.
AML, the AI Act, GDPR and DORA are separate questions
Classification under the AI Act depends on the use: Annex III, point 5(b) lists creditworthiness assessment and credit scoring of natural persons, with an exception for systems used to detect financial fraud. According to UODO, Poland's data protection authority, the data protection impact assessment belongs to the controller. Poland's financial supervisor KNF describes DORA as rules on digital operational resilience, so the institution assesses the model provider, continuity and incident handling. Before launch, analysts know how to work when the model or the integration does not respond.
A few dozen alerts a month
an analyst's checklist
no build
Observation mode before decisions
the agent writes notes with no effect on the process
pilot
Hundreds of alerts, materials in several systems
an agent that prepares alerts for decision
from €3,000
Where to start
The pilot starts in observation mode: the agent writes notes for closed alerts, and analysts compare them with their own findings before the note enters daily work. We build this as part of AI agents. For the free process review, a description of one case from recent weeks and the tools it ran through is enough. You do not need to send confidential documents, and you get a written plan within 2 working days.
Worked example · Scenario
How we count the payback: AI in finance
We run the same count for your team in the free process review.
€2,094
a month for preparing 20 reports
Payback: 1.7 mo.
3 hours to collect the data and prepare a report and checking in 15% of the old time are our assumptions. €34.90 an hour is the EU average hourly labour cost for 2025 according to Eurostat, for companies with 10 or more employees. Model cost: checking the commentary in our reporting system (about €0.0002) plus compiling the data (Claude Sonnet 5.5, about 30,000 + 2,000 tokens), at the ECB rate of 1.1205 USD per euro on 22 October 2026.
See the full calculation →
