For a cross-border payments provider, growing transaction activity meant more information to examine, more cases to explain and more reports to prepare. Novarcs helped bring AI-assisted analysis and reporting into a consistent workflow on AWS, giving teams clearer evidence while keeping people responsible for consequential decisions.
A payment business with a growing review workload
The provider serves businesses that collect and send money internationally, including cross-border e-commerce and international trade customers. Alongside payments, its teams support fraud detection, anti-money-laundering screening and risk reporting. AI already played a role in examining transactions and preparing analysis, so the challenge was to make that support easier to operate as the business expanded.
A useful review depended on more than a generated risk label. Analysts needed transaction history, relevant screening results and an explanation they could examine. Reporting teams then needed to turn that material into a coherent record. Novarcs treated transaction analysis, human review and report preparation as connected parts of the same business workflow.
When inconsistent AI creates extra work
The existing environment made AI governance and operating costs difficult to manage together. Teams maintained different arrangements for access, monitoring and billing, while AI use grew with transaction activity. It was difficult to see which business tasks were driving spending or where changes would have the greatest effect.
Output consistency created a separate burden. Differences in format and explanation made results harder to compare and less convenient for analysts and reporting teams to reuse. The business needed AI assistance that presented supporting information in a repeatable form, with clear review responsibilities for sensitive payment and customer information.
A shared workflow for evidence and reporting
Novarcs consolidated the AI workflow around Amazon Bedrock and adapted the instructions used for transaction analysis, screening support and report drafting. Shared output checks gave teams a consistent format for reviewing context and supporting evidence. The work also brought report templates into the workflow, connecting case analysis with the material compliance teams needed to prepare.
Human responsibility remained part of the design. AI organised information, explained findings and prepared draft reports; the customer retained responsibility for risk criteria, case review and reporting decisions. Novarcs introduced the changes in stages and compared results against existing workflows before expanding their use. This kept validation close to the work that analysts and reporting teams actually performed.
More consistent review and more manageable operations
Risk analysts gained more consistent explanations and a clearer basis for examining suspicious activity. Shared templates and organised evidence made report preparation more repeatable, while complex cases retained a route to human review. The customer could use AI assistance within its review process without removing the people responsible for assessing findings and approving reports.
Operations teams also gained a clearer view of AI consumption by business task, supporting more informed decisions about ongoing expenditure and service improvements. By connecting output quality, reporting needs and cost visibility, Novarcs helped the customer move from fragmented AI support to a workflow its teams could operate and refine together.