July 12, 2023

AIA Contract Documents powered by Catina has launched a large-scale text analytics effort to extract, validate, and analyze text data from their database of 11M+ construction contracts. The resulting Construction Benchmark Database is a comparative and predictive integrated dataset of standardized, anonymized construction contract data from a variety of sources.

Text analytics processes used were –

• Natural Language Processing to extract unstructured text from documents and convert it to structured data tables.
• Machine Learning and statistical analyses to clean and validate the resulting data.
• Artificial Intelligence to identify inconsistencies, errors, and omissions that increase contract risk.

Analyses addressed the questions –
– What is the average duration for a construction project?
– How many change orders are typical over the life of a project?
– When in the life of a project do change orders occur?
– What is the normal range of cost variation due to change orders?
– What is the normal range of duration variation due to change orders?
– What is the combined effect of change orders on duration and cost of a project?

Change orders are generally considered to be a major driver of over-budget, over-time construction problems. This Research Note analyzes data from the Construction Benchmark Database to investigate the reality of that belief and to delve into trends, norms, and variations of change orders. Analyses were variously partitioned by building size, building type, and building location to ascertain if there are differences among those profiles. Learn more here:

The Truth About Change Orders