Know what you buy, from whom, at what price.

Your invoice data is full of answers. Pearstop makes them usable. Your data doesn’t need to be cleaned first. That’s the work we do.

Send us a sample

Send a representative sample of your invoice lines — up to 200 lines, or a handful of invoices (max 10) — and we’ll classify them to show you exactly how it works. No cost, no commitment.

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.

Trusted by leading organisations

StruktonFMOFAROKelpSPIELemTech
Pearstop sample classification

See it work on your own data

Send a representative sample of your invoice lines – up to 200 lines, or a handful of invoices (max 10) – and we'll classify them and send them back labelled. No cost, no setup, nothing to install. It's the fastest way to see exactly what Pearstop does before you commit to anything.

Send us a sample

Send a representative sample of your invoice lines — up to 200 lines, or a handful of invoices (max 10) — and we’ll classify them to show you exactly how it works. No cost, no commitment.

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.

See case studies first

Most companies are flying blind, and only find out when someone asks a question they cannot answer.

Procurement Lead, Infrastructure

You cannot negotiate what you cannot see.

Head of Procurement, Facilities Management

If you cannot defend your cost baseline, you are bidding on a feeling.

Commercial Director, Construction

How it works

1
Stage 1

Send us what you have

PDFs, ERP exports, portal downloads, spreadsheets. API connectors are available for most ERPs.

3
Stage 3

Your spend stays classified

As new invoices arrive, they're classified the same way - no re-cleaning project every year.

As it arrivesWhat Pearstop adds
Invoice descriptionSupplierUNSPSCCategoryUnified supplierConfidence
PMP FLTR 20X24 CS/6AAF INTL40161505Air filtersAAF InternationalHigh
FILTER,AIR,20X24,MERV8AAF FLANDERS40161505Air filtersAAF InternationalHigh
Air Flt 20x24x2 (6/box)AAF40161505Air filtersAAF InternationalHigh
GEN PURP CLNR 5LISS FACILITY47131805General purpose cleanersISS Facility ServicesHigh
Multi-surface cleaner 5ltrISS47131805General purpose cleanersISS Facility ServicesHigh
Additional hoursM&P CONTRACTING80111613Temporary manual labour*M&P Contracting LtdMedium

Illustrative example: six invoice lines from three suppliers. The original description and supplier are kept exactly as they arrived – Pearstop only adds the columns on the right.

* Category inferred from additional data (purchase history and contract context), not stated directly on the invoice — flagged at medium confidence rather than high.

Who it's for

Construction

You're estimating bids from memory because nobody can retrieve historical actuals.

See how it applies →

“Our data isn’t clean enough for this.”

That's the most common thing we hear – and it's the reason to start now rather than wait. Cleaning the data is the first thing Pearstop does, not a prerequisite for working with us.

“Won’t it be cheaper to build this ourselves?”

In-house gives you complete control, but expect to dedicate two data or AI specialists to build it properly and keep maintaining it as your data changes. We've seen an internal build start as an IT team's side project, ship without procurement involved, and turn into a spreadsheet nobody owns within months. Pearstop is that system already built – with a team whose only job is to keep developing it.

Cases

Infrastructure Contractor

A named savings target. The data couldn't locate it. Pearstop made it visible.

We used to have two full-time staff working on category assignment. Now the system does this for us – which has unlocked margin estimations further down the line too. It's more reliable at a fraction of the cost.Head of Procurement, Infrastructure

Read the case →

Facilities Management Provider

100,000+ assets cleaned and resolved. 93% automated. The remaining 7% routed to domain experts.

Our asset lists worked for mechanics on-site, but did not allow us to plan smart maintenance or manage bid risk in a data-driven way.Asset Manager, Facilities Management

Read the case →

Cleaning Services Company

Extraction accuracy: 70% → 99%. Moved from manual review to an ongoing subscription.

It would have taken five engineers and a full year to clean this up. So we decided to look for a better solution.Head of Operations, Cleaning Services

Read the case →

Send us 200 lines. We will send them back labelled.

Send us a sample

Send a representative sample of your invoice lines — up to 200 lines, or a handful of invoices (max 10) — and we’ll classify them to show you exactly how it works. No cost, no commitment.

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.

Frequently asked questions

What is UNSPSC classification?

UNSPSC is a global standard for categorising products and services into a consistent hierarchy. It lets you compare spend across suppliers, sites, and time periods using the same categories, instead of whatever free-text description each invoice happened to use.

How do you classify spend data that isn't clean?

Your data doesn't need to be clean first - that's the work we do. We take invoices, ERP exports, portal downloads, or spreadsheets in whatever state they're in, and extract and classify every line automatically.

Can you extract data from PDF invoices?

Yes. Pearstop reads PDF invoices, scanned documents, and portal exports directly, so you don't need a clean digital export to get started.

How accurate is automated spend classification?

Around 95% of lines are classified automatically. The rest are flagged and routed to a human reviewer, so accuracy stays high without every line needing manual checking.

How long does spend classification take?

A first sample of 200 lines is typically turned around within a few working days. Ongoing classification runs continuously as new invoices arrive.

What does Pearstop do?

Pearstop helps facilities management, infrastructure, and hard services companies clean procurement and asset data so teams can see what they are buying, plan maintenance more reliably, and feed trustworthy data into AI and reporting tools. If you need one plain answer, it is this: we turn messy operational data into something your business can actually use.

Latest Insights

Procurement

UNSPSC vs eCl@ss: when your ERP and reporting disagree

UNSPSC and eCl@ss answer different questions. Here is how to pick when your ERP was built for one an…

Read more
AI & Digital

Why a data-cleaning AI project needs a project manager

Most AI data-cleaning projects fail from a governance gap, not a model gap. Treating the software li…

Read more
Data Quality

When manufacturer and type fields swap in your spend data

A silent field swap between manufacturer and type corrupts classification without a single error mes…

Read more