Client: Large global software M&A group.

Challenge: Partial automation in the finance function still required users to complete complex data structures manually. This created significant effort at scale and increased the risk of human error - especially in high-volume activities like purchase order creation across global teams. The client wanted major productivity gains without investing in a costly replacement finance tool.

What we did: 

Coucal’s consultants:

• designed and implemented an AI-assisted workflow automation approach that integrated LLM APIs with simple, low-cost tooling. For example, for high-volume purchase order creation, the solution parsed key data from invoices, matched it to vendor and finance-system records, and generated structured inputs automatically.

• built a human-in-the-loop verification and approval step to ensure appropriate oversight while eliminating repetitive manual entry.

• trained operations personnel to extend the approach and build additional automations based on proven use cases, enabling continuous improvement beyond the initial deployment.

Outcomes:

• Reduced manual effort by 99% for a high-volume purchase order creation use case

• Improved accuracy and reduced risk of human error through structured extraction and verification

• Delivered significant productivity gains and cost savings without purchasing a new finance platform

• Enabled sustained automation expansion by upskilling internal operations teams

Capability built:

• Reusable AI-assisted automation pattern: LLM integration + lightweight tooling + human-in-the-loop controls

• Workflow playbooks and templates for invoice-to-PO and related processes

• Internal enablement for operations teams to identify, build, and scale additional automations