Natixis CIB in New York reimagined its invoice reconciliation by shifting from manual email-based processing to an automated, AI-driven workflow built on Microsoft Power Platform.

Previously, invoices arrived via email, a dedicated employee filtered them, and then manually keyed each invoice into the purchasing system. This created a heavy workload, delays, limited controls, and a higher risk of human error.

To address this, the team implemented:

• Power Automate to automatically extract invoice data from emails.
• AI Builder with a prebuilt invoice processing model to handle data extraction. Their earlier custom AI model only reached about 60–70% accuracy, but the prebuilt model delivered a stable success rate of 93.3% and an overall ~90% accuracy in production.
• Microsoft Dataverse to store invoice data centrally.
• Power BI to visualize and analyze invoice information.
• Power Apps to monitor AI flow activity and adjust invoices manually when needed.

This combination helped Natixis CIB NY significantly improve accuracy, reduce manual intervention, and gain better control over the full invoice lifecycle.

Automating invoice processing with Power Platform reshaped how the finance and IT teams at Natixis CIB NY work day to day.

Key outcomes include:

• Higher accuracy and fewer errors: The AI model for invoice processing now operates at about 90% accuracy, with the prebuilt AI Builder model showing a 93.3% success rate in tests. This reduces the risk of errors from manual data entry.
• Significant workload reduction: With 3,000 to 5,000 invoices processed each year, automation cut the workload of one full-time employee by more than half. Staff can now focus on more critical, value-added tasks instead of repetitive data entry.
• Faster, more reliable reconciliation: Timelier and more consistent invoice reconciliation supports smoother payment cycles and reduces operational risk.
• Better visibility and control: Using Power BI, senior managers can see how many invoices are being processed, reconcile key invoice attributes with purchasing records, and quickly identify gaps or anomalies. This supports more informed budget and spend decisions.

Overall, the automation helped Natixis CIB NY rethink its invoice process from a manual back-office task into a more controlled, data-driven operation.

To make sure the benefits of Power Platform extended beyond a single project, Natixis CIB NY worked closely with Groupe BPCE’s Tech Expertise and Solutions department to put structure and governance in place.

They established a Power Platform center of excellence (CoE) with a clear mandate to:

• Define governance: Set standards and guardrails for how Power Platform is used across the group, ensuring alignment with IT and security policies.
• Ensure compliance: Make sure all solutions comply with internal IT and security rules.
• Support and training: Provide training, guidance, and a community user group for developers and administrators.
• Encourage reuse and scale: Share best practices and reusable components so other teams can automate operational tasks more easily.

This CoE approach helps Natixis CIB NY and the wider Groupe BPCE organization scale low-code and AI-driven automation in a controlled way, while continuing to explore new use cases beyond invoice processing.