Live eventJoin Spotify and Watershed for a briefing on 2027 sustainability strategy.Register now →
[TEST, REMOVE] Cimpress stand-in logo[TEST, REMOVE] IVC Evidensia stand-in logo

How Cimpress and IVC Evidensia turned weeks of manual data work into repeatable workflows

Two sustainability teams turned messy supplier data into faster, repeatable, audit-ready workflows.

Cimpress & IVC Evidensia x Watershed group case study

Every sustainability team knows the drill. Supplier data rarely arrives ready to use. It comes in different formats, units, languages, and levels of quality, leaving sustainability teams to spend hours cleaning and reconciling it.

Christine Trella at Cimpress and Katharina Wirth at IVC Evidensia joined Watershed’s inaugural AI Fellowship to solve similar data-processing challenges. They built workflows that made the work faster, repeatable, and easier to audit.

Cimpress

From months in Excel to an audit-ready pipeline

Cimpress is a global mass-customization company whose portfolio of brands helps consumers and businesses create personalized products. Each year, Christine Trella’s lean team processed a 331,000-row spend file, supplier reports, travel data, and more than a million rows of logistics data in Excel. The work often took months.

In FY25, they hand-built 188 upload files across 14 datasets, then manually cross-checked every output against the source files for their auditor. One missed deduction or duplicated row could mean double-counted emissions—and there was no easy way to catch it. "We want to spend our time analyzing and driving decarbonization—not getting bogged down in spreadsheets to prevent double counting," Trella explains.

So she used agents to do the wrangling. She fed it all source files: the master spend report, supplier-specific emissions reports, generic emissions factor data, site-specific files, and logistics data. The agent scans the spend file, maps rows to Watershed data categories, flags where supplier-specific data exists (routing those rows away from spend-based datasets to prevent double-counting), applies emissions factors, segments data, generates formatted upload templates, and produces a full tie-out document—plus a rules reference so the logic carries forward to future fiscal years.

Trella estimates the agent cut about three weeks from their data processing timeline. It processed and tied out more than 175 upload files across 14 datasets automatically. The report shows how each output traces back to the original source data, giving the auditor a clear paper trail.

“For me, on the assurance side of things, showing the auditor that I've tied out all those output files is a game-changer. ”

Christine Trella,
Sustainability Data Manager, Cimpress

IVC Evidensia

One day instead of five, across every reporting cycle

IVC Evidensia’s problem was similar but structured differently. IVC Evidensia is an international veterinary-care group that operates a network of clinics and hospitals serving pets and their owners. It collects waste data from suppliers in four countries, each using a different format. IVC’s Katharina Wirth had to combine the files, fill data gaps, convert partial-year data into annual estimates, match facilities, and calculate waste KPIs. Each run took about a week and had to be repeated for mid-year reviews and base year recalculations.

Katharina used Watershed agents to automate the full pipeline. The agent transforms varied supplier reports into Watershed's schema, estimates zero values using spend data where weights weren't captured, matches facilities to IVC Evidensia's Watershed facilities list, annualizes partial data, estimates for uncovered facilities, and generates a README audit trail documenting every step. The result: 16 upload files reduced to four.

The workflow cut processing time by about 80%, from five days to one, across year-end, mid-year, and base-year reporting. It also surfaced missing data and enabled comparisons across countries that revealed new reduction opportunities and prompted direct engagement with local teams.

“This process will reduce the amount of time spent to a day from five days—replacing manual work with the click of a button.”

Katharina Wirth,
Sustainability Manager, IVC Evidensia

Conclusion

Why this matters

The work Trella and Wirth automated is where most of the time in sustainability reporting actually goes—and where most of the audit risk lives.

For both teams, the value goes beyond faster processing. Their workflows are repeatable, documented, and easier to hand off, so the logic no longer has to be rebuilt or held in one person’s head each year.

For sustainability teams running lean and facing growing reporting requirements, that's the real value: saving weeks and building a process you can trust and hand off.