Plum Cuts AP Processing Time by 90% with Agentic Automation

Plum Cuts AP Processing Time by 90% with Agentic Automation

Client

Plum Goodness

Industry

Consumer Brand / DTC

Key Metrics

  • 90% reduction in AP processing time

  • 20+ invoices processed daily with zero manual entry on clean matches

  • 16 person-hours saved every day (4 team members × 4 hrs)

  • Live in production in 1–2 weeks, no custom middleware

Understanding the Problem

Plum, a Mumbai-based consumer goods company, runs its finance operations on SAP. Every month its accounts payable (AP) team hit the same wall: invoices arrived from six separate vendor portals and a shared inbox, each one had to be keyed into SAP by hand, and GRNs (Goods Receipt Notes) were posted manually against purchase orders. Mismatches — wrong quantities, wrong rates, invoices billed against POs that were never fully received — usually surfaced only after the payment had already gone out, and recovering a wrong payment meant weeks of vendor follow-up.

Monday mornings were entirely consumed clearing the weekend's invoice backlog, and month-end close routinely slipped because the team was still reconciling AP.





What We Built

We designed and deployed an AI agent that runs live inside Plum's own SAP environment and owns the full invoice-to-GRN workflow end-to-end:

  • Ingests invoices automatically — pulls PDFs and line-item data from all six vendor portals and the shared inbox, with no manual downloading or forwarding.

  • Reads and extracts line items — parses invoice line items, quantities, and rates directly from the source documents.

  • Runs a three-way match — cross-checks each invoice against its purchase order and the goods actually received, the same check an AP clerk does by hand.

  • Posts clean GRNs on its own — where the invoice, PO, and goods receipt line up, the agent posts the GRN directly into SAP with zero human involvement.

  • Flags exceptions with the reason why — where something's off, it holds the document, states the specific discrepancy, and routes it to the AP team for a one-click approve or reject.

Every action the agent takes is logged and auditable, and no invoice data ever leaves Plum's systems — the agent runs inside their existing SAP and email stack rather than exporting data to a third party. We took this from kickoff to live in production in 1–2 weeks, with no custom middleware and no changes to Plum's existing SAP setup.



Outcome

The AP team no longer spends month-end clearing a backlog — clean invoices post themselves the same day they arrive. Four team members each get back 4 hours a day, 16 person-hours saved daily, and every mismatch is now caught before payment leaves Plum's account instead of after. What used to be a full morning of manual reconciliation is now a 15-minute review of whatever the agent has flagged.



"Monday mornings were just for clearing the invoice backlog. Now I check what the agent flagged and that takes 15 minutes."

-Rohit, Operations Manager, Plum

FAQ

Frequently asked questions

What exactly is an AI agent

An AI agent is an autonomous system designed to handle specific business tasks end-to-end. Unlike simple chatbots, AI agents can reason, take actions, integrate with tools, and follow defined workflows.

Can agents integrate with our existing tools and systems?

How reliable are AI agents in production?

How secure are AI agents?

How does an engagement work?

What do you need from our team to get started?

How long until we see results?

What happens when an agent isn't sure?