2027 Event Agenda

For utilities, contractor invoice review is more than an administrative task. It is a frontline financial control.

At PPL, a single team can process more than 400 contractor invoices each month, with every invoice requiring a manual 3-way match across handwritten field reports, submitted invoices, and contracted rates. Because each review takes 10 to 20 minutes, the process consumes substantial staff time annually while also introducing risk related to billing errors, inconsistent validation, and potential overpayments.

To address this, PPL implemented Alfred, an AI-enabled invoice validation agent designed to automate the end-to-end review process. Alfred ingests incoming invoices, interprets handwritten field activity reports using document understanding capabilities, compares quantities and unit rates against active contract terms, and generates a structured payment recommendation. It also identifies discrepancies and provides transparent indicators to support approval or rejection decisions, creating a more consistent and auditable review process.

Since deployment, Alfred has reduced average review time from 10 to 20 minutes to just 2 to 3 minutes per invoice, an approximately 85% improvement. Beyond efficiency gains, the solution has introduced a stronger layer of financial control by surfacing subtle mismatches between field records, rate cards, and invoiced amounts that were difficult to detect consistently through manual review alone. The result is a previously fragmented process that is now more scalable, more controlled, and better equipped to support high-volume contractor payment operations.

Since deployment, Alfred has reduced average review time from 10–20 minutes to just 2–3 minutes per invoice, an approximately 85% improvement. Beyond efficiency gains, the solution has introduced a stronger layer of financial control by surfacing subtle mismatches between field records, rate cards, and invoiced amounts that were difficult to detect consistently through manual review alone. The result is a previously fragmented process that is now more scalable, more controlled, and better equipped to support high-volume contractor payment operations.

This session will explore Alfred’s architecture, the challenge of interpreting unstructured handwritten documentation in a utility environment, and the lessons learned from deploying AI into a high-stakes financial approval workflow. Attendees will gain practical insights into how utilities can apply AI to improve financial controls, reduce manual effort, and modernize document-intensive operational processes.