[ Industry case study — AI & Automation ]
Automated reconciliation saves 14,000+ staff hours a year
STX Next built ReconSync, an Azure-native platform that automatically reconciles 320,000+ bank transactions a year for UK wealth manager Mattioli Woods.
Original project by STX Next ↗Industry reference — not a NexSum Labs engagement.
- Client
- Mattioli Woods
- Industry
- Wealth management & financial services
- Services
- Process automation, Data engineering, Cloud solutions, Agile transformation
- Year
- 2026
- Location
- United Kingdom

The client
Mattioli Woods is a leading UK wealth management and employee benefits firm founded in 1991, serving more than 20,000 clients with over £16 billion in assets under management, administration and advice. After its 2024 acquisition by private equity firm Pollen Street Capital, the firm launched an ambitious operational transformation programme.
The challenge
Daily bank reconciliation, which underpins the government-mandated Annual Valuation Statement, was done by hand by a 200-person CRM team, costing over 100 hours a day and 14,000+ hours a year. For four of five core banks, someone had to log in each morning and download transaction files manually. The reconciliation logic lived in people's heads with no documented architecture, and internal teams lacked a shared Agile delivery framework.
The solution
01Automated bank data collection
Scheduled Azure Durable Functions pull transaction files from bank APIs and SFTP endpoints and store them securely in Azure Blob Storage, with no morning log-ins required.
02End-to-end ReconSync pipeline
Service Bus-driven functions fetch expected transactions from the internal MWeb platform, send complete batches to the ReconArt reconciliation engine, and sync the matched results back automatically.
03Secure, observable infrastructure
Cosmos DB, Key Vault with Managed Identity, private endpoints, GitHub Actions with modular Bicep, and App Insights give the platform auditable storage, secrets handling and monitoring.
04Embedded Agile coaching
An embedded Scrum Master introduced ceremonies, backlog refinement and estimation across three internal product teams, lifting planned sprint delivery to 50-90%.



The results
A process that consumed a full working day across a 200-person team now runs as an automated pipeline in under two hours, monitored by a team of four. More than 320,000 transactions a year are reconciled across five banking partners without manual touchpoints, saving 14,000+ staff hours annually. Coverage reached 60-70% of transaction volume at soft launch, with a clear path to full automation.
Tools & stack
Original project by STX Next ↗ — published by the delivering agency, included here as an industry reference.
[ More industry examples ]
More from ai & automation.
AI & AutomationARC Europe · 2026
LLM chatbot routes stranded drivers in 2-4 questions
2-4questions to reach the right service
AI & AutomationDisasters Emergency Committee · 2026
An AI proofreader to speed up appeal content at the Disasters Emergency Committee
100%pilot participants who found the tool valuable
AI & AutomationMerck · 2024
AI assistant cuts compound research from months to hours
6 hoursto find molecules, vs. 6 months manually
[ Next step ]
Want this kind of result? Let's scope it.
contact@nexsumlabs.tech — we reply within one business dayNext case study — Zego