ProdRatesIQ: 15,500 production rates, one queryable library

Context

Every estimate, schedule, and claim in heavy industry ultimately rests on a production rate: how many workhours to fix a tonne of rebar, weld a metre of pipe, terminate a cable, or excavate a cubic metre of rock. These rates are the quiet foundation of project controls, and in most consultancies they live in scattered places: commercial estimating references, old tender workbooks, PDFs on a shelf, and the memory of senior estimators.

This case study is FaolanIQ proving its platform on Faolan Consulting's own core asset: the production-rates knowledge that underpins estimating, scheduling, and benchmarking work for mining and infrastructure clients.

The problem

When rates live in documents, three things go wrong. First, retrieval is slow: an estimator under tender pressure digs through workbooks to find a defensible rate, and often re-derives one that already exists two folders away. Second, provenance evaporates: a rate gets copied into a tender, the source is not recorded, and a year later nobody can defend where the number came from. Third, the knowledge does not compound: each tender's judgement calls stay in that tender's file instead of strengthening a shared asset.

The approach

The Build pillar tasked Forge, FaolanIQ's production-rates agent, with building ProdRatesIQ: a single, structured, queryable library of production rates on the fleet's cloud data platform.

Forge read the source references, normalised every item into a common schema, and loaded the result as structured data rather than documents. Each rate item carries its discipline, item code and description, unit of measure in both metric and imperial, workhour values with slow, normal, and fast performance bands, the rate format and item type, a confidence flag, and full provenance down to the source workbook and worksheet it came from. Full-text search indexing makes the library queryable in natural language, and a match cache accelerates repeated fuzzy lookups so agents and estimators find the right item quickly.

Governance ran the same way as every FaolanIQ build: Forge worked to a brief from Thor, the Build pillar head, with schema-validated hand-offs and an audit trail, and the library's contents were verified against the live database rather than taken on trust.

The outcome

• 15,515 production rate items normalised and live in the cloud library, organised across 49 disciplines, from civil and structural through mechanical, piping, electrical, and instrumentation. (Counts verified against the live database on 2026-06-07.)
• Every item is traceable to its source workbook and worksheet, so any rate used in an estimate can be defended line by line.
• The library is versioned under a rate-library version contract, agreed with FaolanIQ's estimating product on 2026-07-06, so every estimate pins the exact library release it priced against. When the library improves, prior estimates keep their audit trail.
• Forge maintains a live gap register against the library, naming the areas still to be sourced (for example underground mining development rates and Southern African labour norms), so acquisition is targeted instead of accidental.
• The library now feeds the wider FaolanIQ Build pillar: EstimatingIQ draws rates for first-principles tender build-ups, and CostIQ uses the same items as benchmarks when checking contractor claims.

What this demonstrates

Database-first beats document-first. A rates library as structured, versioned data compounds in value with every project, keeps its provenance, and can be queried by both people and agents in seconds. It also shows FaolanIQ agents doing sustained curation work at a scale no individual would attempt by hand, under the same governance and verification discipline as client-facing work.

Controls. Intelligence. Delivered.