Composite scenarios. Real workflow patterns.
Trust, delivered.

The challenge

Three separate tools — a categorisation API, a reconciliation SaaS, and a manual AML process — generating three fragmented audit trails. When the FCA opened a supervisory visit, Paysona's compliance team spent 11 days assembling documentation from three sources. The FCA found gaps.

The solution

Trustinera AI replaced all three tools with a unified platform. TrustFlow handles live ingest from Barclays and Starling feeds. Categorise processes 4.5M transactions per month at 97% accuracy. Sentrise handles AML screening and case management. OpsIQ monitors pipeline health 24/7.

Timeline

6 weeks from contract to full production deployment, including data migration and team training.

The challenge

Meridian's compliance team of four was spending 18 hours per week on manual AML monitoring — pulling transaction reports, running spreadsheet-based watchlist checks, and writing SAR narratives from scratch. With AUM growing 40% year-on-year, the team knew the manual approach wouldn't survive the next year.

The solution

Sentrise replaced the manual AML workflow with real-time transaction monitoring, automated watchlist screening (OFAC, HM Treasury, PEP), and a case management system with auto-drafted SAR narratives. TrustFlow connected to Meridian's three custodian data feeds via API.

Timeline

4 weeks from signed contract to live AML monitoring in production.

The challenge

Clearway's affordability assessments relied on manual bank statement review — categorising income and expenditure by hand for each applicant. With 800 applications per week, the process took three underwriters three days each week. Categorisation inconsistency was flagged in an internal audit.

The solution

Trustinera AI's Categorise module processes uploaded bank statements and returns structured income/expenditure categories with confidence scores. High-confidence items flow straight to the affordability model. Low-confidence items queue for an underwriter review — reducing manual review from 100% to under 8% of applications.

Timeline

3 weeks integration via Python SDK, including underwriter training and threshold calibration.

The challenge

Nortex wanted to add AI transaction categorisation to their bookkeeping product but didn't have the ML engineering capacity to build it. Two previous vendors had poor accuracy on UK merchant data and required months of professional services to integrate.

The solution

Trustinera AI's Categorise API was integrated via the TypeScript SDK in 11 days by two engineers. The API's OpenAPI 3.1 spec and sandbox environment made integration testing straightforward. Nortex users now see AI-categorised transactions in their bookkeeping timeline. User corrections are fed back via the corrections API, improving accuracy for the whole Nortex customer base.

Timeline

11 days from API key to production feature launch.

The challenge

Ashfield's finance team of eight spent five working days on every period close — manually downloading bank statements from four banks, reconciling against Sage 200 entries, and emailing exception lists back and forth. The process was error-prone and delayed board-level financial reporting by a week every month.

The solution

TrustFlow connected to Ashfield's four bank feeds (NatWest, HSBC, Barclays, Starling). Categorise processes and enriches all transactions. Reconcile matches against Sage 200 automatically, surfacing only genuine exceptions. Period-close reports are generated on demand and delivered directly to the Finance Director dashboard.

Timeline

8 weeks including bank feed onboarding, Sage 200 integration configuration, and team changeover.

The challenge

Corda's finance ops team was processing 2.8M transactions monthly across seven payment rail types, with a manual categorisation process that took three people 40 hours a week. Their false positive rate in fraud detection was 18% — generating more noise than signal for the risk team.

The solution

Trustinera AI processes all 2.8M monthly transactions via TrustFlow's webhook ingestion. Categorise handles all seven payment types with a custom taxonomy configured for Corda's chart of accounts. Sentrise reduced false positives via configurable risk scoring thresholds tuned against Corda's historic transaction patterns.

Timeline

5 weeks from first API call to full production volume.