AGRA turned a black box of PDF invoices into negotiation power.
Nils Simonsen inherited indirect spend at AGRA, a family-owned Nordic food group with 11 factories, and could only "click through PDFs to find what we'd actually bought." Once Mithra gave him line-item visibility, he started walking into supplier negotiations better prepared than the suppliers.
Group Procurement · Indirect spend
Norway · Sweden · Denmark · Poland
"It became the colleague I spoke to most."
Nils on inheriting indirect spend as a black box of PDF invoices, and what changed once he could ask his data a question and get an answer.
Indirect spend was a black box.
AGRA prizes factory autonomy, so each of its 11 sites chose its own suppliers. For central procurement, that meant indirect spend lived as PDF invoices outside the ERP. The only option was clicking through PDFs to find what the group had actually bought.
- Indirect invoices arrived only as PDFs, outside the ERP
- No line-item visibility, "click through PDFs to find what we'd bought"
- Taxonomy built on ledger codes, not what was actually purchased
- 11 autonomous factories, each choosing its own suppliers
- Suppliers owned the data negotiations, which started on the back foot
"I was quite shocked when I started. The only data I could reach was invoices in PDFs. We more or less knew nothing about what we bought."
Hand over the invoices. Then ask the data anything.
AGRA exported its PDF invoices to Mithra, which extracted, classified, and harmonized the spend into a queryable view, then made it conversational, so the team could find opportunities by asking, not by building VLOOKUPs.
Collect & classify
PDF invoices became 250,000+ line items with a procurement-native taxonomy based on what AGRA actually buys not how factories happened to book it in the ledger.
Ask the data
No coding, no spreadsheets. The team asks plain questions "where do we buy the same item at different prices across countries?" and gets reports back in minutes.
Find & cluster
A couple of hours surfaces a list of opportunities. AGRA started where the value was densest, MRO, with ~1,000 suppliers, and began building supplier clusters.
From one long project to many fast ones.
In his first session with the data, Nils found 9 to 10 opportunities. One was lab equipment bought from a European distributor in one country but local distributors elsewhere. A single one-hour call later, AGRA had a new price list for every country: a €40K annual saving from about three hours of work.
See more customer stories- €40K annual saving landed from a single ~3-hour project
- 9–10 opportunities surfaced in the first questioning session
- Many parallel sourcing projects instead of one slow tender
- Negotiations entered better prepared than the supplier
- Compliance visibility see which factories actually switched
"I can spend half an hour the day before a negotiation, talk to the system, and walk in better prepared than my opposition. They don't expect procurement to have that data depth and it changes the whole balance of power."
It didn't just save money. It changed how the team works.
Out of the cubicles
Instead of one analyst surfacing a single find a month, the team trades questions and opportunities in real time and works far closer to operations.
Credibility with engineers
Data-driven detail "these photoelectric sensors look identical across suppliers" opened doors with maintenance teams that used to stay shut to procurement.
Time back for the real work
No more reading invoices and typing into Excel. The team spends its time on negotiation and strategy what it's actually good at.
"Our biggest mistake was waiting a year or two to start, trying to get our data perfect first. Start with what you have, know it isn't perfect, and adapt as you go. We should have started earlier."
See what Mithra could find in your spend data.
Bring a representative sample, even a folder of PDF invoices, and we'll show you your classified spend, normalized suppliers, and a prioritized savings analysis, on your own data, in weeks.