Your procurement data contains the answers. You just cannot read it yet.
Most FM and construction companies have years of purchase order data in SAP or Oracle. Without UNSPSC classification, it is a pile of free-text line items. With it, you can see exactly what you spend by category, benchmark suppliers, and build tenders from actual cost data. Pearstop classifies invoices at any scale, from a pilot of a few hundred lines to full production volume - it scales with you.
The four-level code that makes procurement data comparable.
UNSPSC (United Nations Standard Products and Services Code) is a global four-level hierarchy used to categorise every product and service a company buys. Segment and Family give you broad spend visibility. Class and Commodity give you the precision to negotiate, benchmark, and manage categories effectively.
Without it, the same physical work appears under dozens of different strings. At SPIE, invoice lines like 'HVAC unt' were enriched to 'HVAC Unit, Air Handler, Climate Control, Carrier' with a consistent UNSPSC commodity code. That is the difference between data you can act on and data you cannot.
Most lines classify themselves. The rest get a second look.
Rules and machine learning classify high-confidence lines immediately. The layer that used to need a dedicated team now runs in the background.
Rules and machine learning
Supplier and GL-based rules combined with a machine learning layer trained on your own spend classify the high-confidence majority of lines immediately.
LLM layer for the edge cases
Field engineer shorthand, part numbers, and multilingual descriptions are resolved by a large language model with deep product and industry knowledge, the layer that handles what rules and ML cannot.
Human review, then it remembers
The five to ten percent the engine is uncertain about is flagged for your team. Each decision feeds back into the model, and the review queue drops over successive months.
What a classified spend baseline makes possible
Category Management
Supplier Benchmarking
ERP and BI Integration
“We had thousands of product lines that needed to be categorised before we could even begin to understand our costs. Pearstop classified them in under a week. That would have taken our team six months and still would not have been this accurate.”
How does UNSPSC classification work for hard services and FM companies?
Hard services FM companies typically buy from 500 to 3,000 suppliers across multiple sites. Invoice descriptions are written by field engineers, not buyers, which means the same work appears under hundreds of different strings. Manual UNSPSC classification at 5,000 to 35,000 lines per month requires one or two dedicated staff working continuously just to stay current. Pearstop's automated engine classifies 90-95% of those lines without human input, with the remainder flagged for review. The review queue shrinks each month as the engine learns from your team's decisions.
Frequently asked
What is UNSPSC classification?
UNSPSC (United Nations Standard Products and Services Code) is a hierarchical classification system used worldwide to categorise procurement spend. It has four levels: Segment, Family, Class, and Commodity. UNSPSC enables consistent spend analysis, supplier benchmarking, and category management across contracts, sites, and ERP systems.
How accurate is automated UNSPSC classification?
Pearstop's classification engine, combining rules, machine learning, an LLM layer, and human review, achieves 90 to 95% automatic classification on typical procurement datasets. The remaining 5 to 10% is flagged for human review. Each reviewed decision feeds back into the engine, shrinking the review queue over time until it reaches near zero.
Can UNSPSC classification identify the manufacturer's part number behind a supplier's own code?
Yes. Where a manufacturer's original part number exists in the data or a connected reference source, the engine resolves it and returns it as a separate end manufacturer MPN field, rather than leaving the supplier's own code as the only reference on the line.
Does Pearstop integrate with SAP for UNSPSC classification?
Yes. Pearstop receives data via CSV export or direct API connection from SAP, Oracle, and other ERP and P2P platforms. In practice, many clients find that CSV export is the simplest way to start. Classified data is returned in the same format, ready to load back into SAP or feed into BI tools.
How long does UNSPSC classification take?
Most clients have a clean, classified dataset ready within four to six weeks of starting. The first engagement begins with a Data Stability Baseline so you can assess the output quality before committing to ongoing classification.
Can UNSPSC-classified spend be used for Scope 3 carbon reporting?
Yes. The spend-based method for Scope 3 multiplies the spend in each category by an emissions factor, so spend has to be classified by category first, and UNSPSC gives you consistent categories across suppliers, sites and years. Pearstop provides the classified spend; the emissions calculation happens in your carbon reporting tool.
Ready to automate your procurement classification?
Talk to sales. We will show you exactly how the classification engine would work with your data.
