Been experimenting with some ways to use AI to make savings in projects.
The cheapest construction work is the work we can remove from scope
AI can help identify value-engineering opportunities, but asking it to “find savings” across an entire project won't work. A better approach is to give it structured project data and divide the analysis into three focused workflows:
- Design: Compare the developed design against the original contract requirements to identify scope growth, gold plating and over-specified products.
- Procurement: Review major packages for items that could be purchased directly, free-issued or separated from a subcontract. Apply the “cable tie test”: the saving must justify the extra procurement and management risk.
- Methodology: Test alternative crews, plant, sequences and production rates. Compare the direct-cost impact with programme changes and weekly preliminary costs.
Each option should then be challenged by independent commercial, construction and site perspectives.
AI’s role is to generate, compare and rank the options. Experienced people still decide which options are commercially, technically and practically viable.
The quality of the result depends on the quality of the context: clear contract requirements, an itemized estimate, the construction programme and reliable productivity data.
Anybody else found any strategies that work?