Skip to main content

Isaac | Transaction Monitoring Investigator

Version 2.3.1
About Release
  • Release date: August 14, 2026
  • Compatible platform version: Work.AI 10.3+

The Isaac v2.3.1 release expands transaction coverage with new internal transfer types, enhances narratives, and eliminates unused processing flows.

  • Isaac now supports the internal_transfer_debit and internal_transfer_credit transaction types, allowing for more comprehensive handling of internal transfer activity.

  • To make Isaac's narrative more accurate and easier to review, the following changes have been introduced:

    • Final narratives use the counterparty name deduced from the Organization Classification Skill parameter to ensure consistency.

    • Significant counterparties are highlighted in narratives, making it easier to identify those counterparties that represent a substantial portion of alerted funds.

    • Industry information is shown alongside the counterparty name in narratives, giving reviewers more context about the organizations involved in the alerted activity.

    • The alerted transaction status is updated from the matching all-transactions list after pre-built rules are applied. This improves transaction analysis and ensures alerted transaction details reflect the latest rule-based processing results.

  • Business Processes that created mock data have been removed to reduce unnecessary processing paths.

Version 2.3.0
About Release
  • Release date: July 7, 2026
  • Compatible platform version: Work.AI 10.3+

The Isaac v2.3.0 release enhances investigation efficiency and the quality and clarity of outputs. It introduces new capabilities to better manage counterparty data, refine narratives, and facilitate processing of large data volumes. With this release, we also announce the upgrade to Java 21 to deliver improved security and performance.

New features

  • Isaac can now automatically switch to CSV output when transaction counts exceed Excel file limits. You can set the maximum transaction threshold for this switch in the Output step of the configuration interface. The feature enables reliable handling of high-volume outputs without manual intervention or data loss.

  • You can explicitly include trusted counterparties or exclude selected ones from processing. The feature increase investigation efficiency by minimizing unnecessary alerts and focusing on relevant entities.

  • Isaac now offers counterparty classification and risk assessment, enabling enrichment of counterparty data with relevant risk insights. The new capability improves investigation quality, supports consistent decision-making, and reduces manual analysis effort.

  • The configuration interface now includes settings to fine-tune narrative generation, allowing you to create clear, consistent investigation summaries at both entity and alert levels. In the Output step, you have options to generate entity- or alert-level narratives and to include expected and actual transactions in alert-level narratives. The enhancement lets you tailor report detail, improve transparency, and support efficient reviews.

  • You can configure counterparties to be displayed in a tabular format in Word transaction reports by enabling the Format the counterparty transactions in the Word transaction report as a table? setting in the Output step. This feature improves report readability and structure, making result analysis easier.

Improvements

  • Isaac's Word transaction report has been updated to include additional elements, which makes generated outputs more complete and easier to understand.

  • To improve Isaac's performance and reliability, enhancements were made to the processing of large transaction volumes.

  • We have upgraded Isaac to Java 21 to align with modern technology standards. This update also enhances security and improves concurrent workload handling for more efficient processing.

Version 2.2.0
About Release
  • Release date: May 10, 2026
  • Compatible platform version: Work.AI 10.2.9+

The Isaac v2.2.0 release delivers new capabilities for counterparty identification and targeted investigation, account‑level risk assessment support, SMB file ingestion, and performance improvements for larger alerts.

New features

  • The configuration interface now has a section with counterparty selection filters, so you can control which counterparties to include in transaction investigations. You can set Isaac to review all alerted counterparties, those from transaction lists, those associated with high-risk jurisdictions and round-dollar transactions, or a selection of counterparties per type. This feature helps you tailor the investigation scope to your exact needs and reduce unnecessary alerts, improving the investigation efficiency and decision quality.

  • Isaac now allows counterparty names to be automatically overwritten from transaction descriptions, making it easier to identify the right parties. The overwriting mechanism helps you extract more accurate and relevant names from your transaction data. As a result, you get clearer context, less confusion, and faster investigations.

  • A new model for account-level analysis is now available. Choose the Model ID in the configuration interface, and it will assess risks at the entity level and get you a summary status, comments, and detailed results. By surfacing a holistic view of account behavior and risk drivers, the feature helps you spot risks more accurately and make decisions faster with less manual work.

  • Isaac now supports file ingestion from Server Message Block (SMB) shares, enabling seamless data intake from Windows file shares and Linux Samba shares. Using the new capability, you can integrate existing file‑based workflows directly into Isaac without additional data transfer steps. This makes the workflow setup easier and allows you to leverage your current file systems for reliable data ingestion.

Improvements

  • Isaac’s heap space has been increased to handle larger and more complex alerts. The change improves the solution's stability and scalability, so you can process more data without performance degradation.
Version 2.1.2
About Release
  • Release date: April 12, 2026
  • Compatible platform version: Work.AI 10.2.9+

The Isaac v2.1.2 release introduces configurable narrative grouping and counterparty search limits, enhances automated deduplication, and improves investigation efficiency and data quality by enabling entity‑centric output handling. Extended rules, more efficient large-volume transaction handling, enhanced narratives, and better error handling ensure more reliable processing and reduced operational overhead.

New features

  • We have introduced an entity-centric approach to handling output data, consolidating multiple alerts related to the same entity into a single sheet rather than producing separate outputs. The enhancement reduces data fragmentation and simplifies analysis and reporting, resulting in improved investigation efficiency and reduced operational overhead when managing high alert volumes.

  • Isaac now allows you to limit the number of counterparties returned during searches using a new dedicated flag in the configuration interface on the Investigation step. The feature decreases processing time and minimizes unnecessary data handling, allowing your teams to optimize productivity without sacrificing accuracy.

  • We have automated deduplication between account holders and counterparties, ensuring that the same entity is not represented multiple times across investigation results. The feature reduces redundant data to let your analysts streamline investigations.

  • Isaac now supports automatic extraction of address information from counterparties, enabling seamless downstream integration with the Evan AI Agent for Adverse Media Monitoring. As a result, your investigations can benefit from richer data inputs and more effective automation across workflows.

  • You can now set transaction narratives to be grouped based on the entry method, transaction code description, or transaction type parameters. You can choose a grouping criteria in the Narrative grouping field on the Output step. The new feature allows you to tailor how transaction narratives are consolidated to match investigation workflows, improve output readability, and reduce review time by helping analysts quickly identify relevant patterns.

Improvements

  • Isaac's decision rules have been extended with the following options to increase the detection coverage for common deposit and transfer risk patterns:

    • No-code rules: LOW-RISK CHECK DEPOSIT and HIGH-RISK CHECK DEPOSIT

    • Pre-built rules: INTERNAL_TRANSFER_PATTERN, SOURCE_OF_FUNDS_ANALYSIS_PATTERN, JUST_UNDER_THRESHOLD_DEPOSITS, and ROUND_DOLLAR_DEPOSITS

  • We have refined counterparty name cleaning and deduplication during processing to improve data quality and reduce duplicate entity records. As a result, Isaac delivers more accurate and consistent counterparty identification, enabling clearer investigations.

  • To improve clarity and consistency in investigation outputs, adjustments were introduced to Isaac's narratives. The clearer, more standardized narratives help analysts communicate results more effectively and support regulatory requirements.

  • Error handling has been enhanced across investigations to ensure issues are identified and reported more accurately within minimum time.

  • We have improved the handling of large transaction lists during ingestion and analysis, which enables faster processing and greater scalability and allows teams to handle high‑volume workloads efficiently without compromising data quality.

Version 2.1.0
About Release
  • Release date: January 19, 2026
  • Compatible platform version: Work.AI 10.2.9+

Isaac 2.1.0 introduces enhancements that simplify decision logic configuration, expand rule coverage, improve investigation depth and transparency, and add flexibility to Excel reporting.

  • To minimize the time and manual effort required to configure decision logic, Isaac provides a collection of pre-built rules that cover more complex scenarios and pattern-based detection. Once you enable the rule engine, you can select the required rules directly in the configuration. Isaac then evaluates transactions against the defined conditions and returns a transaction status (for example, low risk) along with a decision comment.

  • The no-code rules capability has been enhanced to support a wider range of use cases for transaction monitoring and fraud detection. Designed as a starting point, these rules can be customized in the Decision Rules steps of the respective flows to fit your specific business needs while reducing reliance on engineering resources.

  • A new Excel output format, PivotTablePerTab, is now available, allowing you to define your own pivot tables per worksheet and upload a custom template. This document format enables more flexible and tailored transaction analysis and enriches reporting capabilities.

  • Isaac now identifies top counterparties and sends their details to the final investigation step, including their activity across the supplied transactions, which provides additional context for investigators and supports deeper analysis.

  • New fields have been added to the data schema to support richer investigations and reporting, including AMM and related reports.

  • To ensure data quality and prevent incomplete or invalid transactions from being processed, Isaac validates incoming transaction data and returns an error if essential information is missing.

  • Transaction narratives now include additional data points, improving decision transparency and making it easier to understand the rationale behind Isaac's actions.

Version 2.0.0
About Release
  • Release date: November 26, 2025
  • Compatible platform version: Work.AI 10.2.9+

Isaac v2.0 introduces new features and enhancements to help you reduce operational costs, accelerate decision-making, and strengthen risk defenses. The new features include ingesting data via API, three specialized Business Processes instead of a single all-in-one flow, a no-code rule engine to help guide Isaac's automated decisions, and an optional fraud investigation path. In addition, significant improvements have been made to optimize Isaac’s data schema and enhance reporting.

New features

  • To streamline alert review and enhance your flexibility in data handling, we have split the single Isaac Transaction Monitoring Investigator Business Process (BP) into several separate flows, each serving its own purpose:

    • The Transaction Monitoring Core BP is recommended for batch alert processing. The BP can be used on its own or in combination with the Transaction Monitoring File Ingestion and Transaction Monitoring Alert Investigation BPs. It completes basic Isaac's operations, such as fetching and consolidating deferral and alert data, sending notifications, and generating reports.

    • The Transaction Monitoring Alert Investigation BP is recommended for alert-by-alert real-time processing. The BP fetches and evaluates alerts, enriches them with manual review data (if any) and additional transaction details, and passes them through the transaction monitoring or fraud detection sub-flows. It can be used on its own or in combination with the Transaction Monitoring Core BP.

    • The Transaction Monitoring File Ingestion BP is mainly intended for historical data evaluation. It lets you manually ingest lists of deferrals and alerts uploaded to S3 buckets and triggers the Transaction Monitoring Core BP.

  • You now have the option to ingest alert and deferral data directly from your internal transaction monitoring system using an API. This capability supports data schemas and contracts for seamless integration with alert systems, reducing manual effort and operational overhead required to get Isaac up and running.

    For historical data evaluation, manual input upload remains available as a separate process through the dedicated File Ingestion BP. During configuration, you can choose your preferred method for submitting data: API for real-time integration or File ingestion for manual uploads, giving you flexibility to align with your operational needs.

  • To help you automate transaction evaluation against specific criteria (for instance, fraud risk), we have introduced a rule engine. The engine is basically a sequence of no-code Decision rules steps you can configure to suit your business purposes. Isaac examines each transaction against preconfigured conditions and assigns an evaluation flag (for instance, labels a transaction as low-risk).

  • Isaac now offers an optional fraud investigation path as a sub-flow within the Transaction Monitoring Alert Investigation BP. This new sub-flow enables you to assess transactions for potential fraud risks. By integrating fraud evaluation into alert investigations, you can further strengthen compliance, reduce financial exposure, and enhance overall risk management.

  • In addition to the final summary and preliminary deferral report, Isaac now offers two more reporting options to ensure consistent information delivery with minimal manual effort. Both report types are supplied with default templates you can customize to your needs.

    • The transaction report allows you to export transaction data in a structured, easy-to-use format to an Excel or Word document. With the report, you can quickly review, filter, and share transaction details for audits, compliance checks, and operational insights without relying on manual data compilation.

    • The fraud summary report consolidates all critical fraud-related information coming from the fraud investigation path into a Word document. The report provides a comprehensive case overview: case summary, scope of review, involved parties, and detailed narratives on fraud transactions.

Improvements

  • To improve data completeness and Isaac's decision-making capabilities, a new process structure was added to enrich transactions with additional details (for instance, data from images or external systems). As a result, you can get a more comprehensive overview of each transaction without investing extra manual effort.

  • To provide richer data for downstream investigations and application of rules specifically related to Anti-Money Laundering and fraud detection, Isaac's data schema was updated to cover extended transaction details.

    Now, transaction analysis takes into account customer and industry profiles, allowing you to improve the fraud detection accuracy and provide deeper insights into operational data. The enhancement enables you to compare each transaction against the customer’s historical behavior and benchmark it against patterns typical for similar customers in the same industry. By identifying deviations from expected norms, your teams can detect anomalies faster and make more informed risk decisions.