A FIELD GUIDE FROM
The Consight team
About this guide
A practical framework for estimating teams adopting AI without abandoning the disciplines that make an estimate defensible.
Who it is for
Preconstruction leaders, chief estimators, and estimating teams deciding where AI belongs in their process.
What you will take away
Eight laws covering scope, quantities, cost, bid leveling, trade interfaces, requirements, benchmarks, and documentation.
01
If the scope is wrong, everything downstream is wrong.
Every estimate starts with understanding exactly what is being built.
Drawings, specifications, narratives, addenda, owner requirements, and site conditions collectively define the work. The challenge is that the scope is rarely written neatly in one place.
Important requirements may be buried hundreds of pages apart.
A strong estimating process turns that fragmented information into a clear definition of responsibility before pricing begins.
HOW AI CHANGES THE WORK
Modern AI can review plans, specifications, RFPs, narratives, and supporting documents together to help teams identify requirements and organize them into structured scopes of work.
Instead of beginning with a blank scope sheet, estimators can begin with a first draft generated from the project documents—and then apply their judgment.
THE LAW
Define the work before you price the work.
02
You cannot accurately price what you have not measured.
Quantities turn scope into something estimators can price.
Square feet. Linear feet. Cubic yards. Counts. Tons. Fixtures. Openings. Equipment.
For decades, producing those quantities has required estimators to manually trace, count, measure, and transfer information between drawings and estimating systems.
That work is increasingly becoming machine-assisted.
HOW AI CHANGES THE WORK
Computer vision and AI-assisted takeoff tools can identify drawing elements, perform measurements, count objects, and compare quantities between drawing revisions.
The estimator still determines what matters and validates the result. But software can increasingly handle the repetitive process of finding and measuring it.
THE LAW
Measure before you assume.
03
A quantity is not an estimate.
Knowing that a project contains 50,000 square feet of concrete tells you very little until quantities are translated into labor, material, equipment, subcontractor pricing, conditions, escalation, and risk.
That translation is where estimating experience matters.
A good estimate is not simply a calculation. It is an informed opinion of what the work should cost.
HOW AI CHANGES THE WORK
AI can connect quantities with historical project data, internal cost libraries, assemblies, regional information, and previous estimates.
It can also help identify unusual unit costs, missing cost components, or values that differ significantly from historical projects.
The goal is not to let a model decide what a project costs. It is to give the estimator more information before they decide.
THE LAW
Build an independent opinion of cost before the market gives you one.
04
The lowest bid is not necessarily the lowest cost.
Subcontractors rarely bid projects the same way.
One includes temporary protection. Another excludes it. One carries premium time. Another assumes straight time. One includes equipment. Another expects the GC to provide it.
Looking only at the number at the bottom of the proposal can hide enormous differences in scope.
Bid leveling exists to expose those differences.
HOW AI CHANGES THE WORK
AI can read subcontractor proposals, extract pricing and scope information, populate comparison sheets, identify exclusions, and highlight differences between bidders.
Instead of spending hours transferring information from PDFs into spreadsheets, estimating teams can spend more of that time investigating the differences that actually matter.
This is one of the problems Consight was built specifically to solve.
THE LAW
Never compare price without comparing scope.
05
Risk lives between the scopes.
Many of the most expensive estimating mistakes are not obvious omissions inside a single trade. They happen between trades.
Who provides housekeeping pads? Who owns trenching and backfill? Who provides embeds? Who patches penetrations? Who carries temporary power? Who removes hazardous materials discovered before demolition?
Every handoff between trades creates an opportunity for scope to be missed, duplicated, or misunderstood.
HOW AI CHANGES THE WORK
AI can evaluate scopes across multiple trades simultaneously and help identify areas where responsibilities intersect.
That makes it possible to systematically review interfaces instead of relying entirely on an estimator remembering every possible handoff.
The estimator still resolves the responsibility. AI helps make sure the question gets asked.
THE LAW
The greatest scope risk often exists between the trades.
06
A bid is only complete if it satisfies the project requirements.
A subcontractor can carry the right quantity and still submit the wrong number.
Specifications may require certain manufacturers, testing procedures, warranties, mockups, certifications, logistics, phasing, temporary work, or closeout requirements.
Those details frequently determine whether two bids are actually comparable.
HOW AI CHANGES THE WORK
AI can search project specifications, addenda, drawings, and contracts for trade-specific requirements and compare those requirements against subcontractor proposals.
Instead of asking an estimator to remember every provision buried throughout the documents, software can help create a structured checklist of items that need verification.
THE LAW
Verify the requirements behind the number.
07
Every estimate should pass a reasonableness test.
A detailed estimate can still be wrong.
That is why experienced estimators step back and look at the project from another level.
Cost per square foot. Concrete per square foot. Steel tonnage. Labor productivity. Equipment counts. Historical subcontractor pricing. Trade percentages. Cost by building area.
These metrics provide a second way to test the estimate.
HOW AI CHANGES THE WORK
AI can continuously compare current estimates against historical projects, internal benchmarks, unit costs, and other available data.
Instead of manually hunting for unusual values, teams can surface anomalies as the estimate develops.
The result is an estimate that is tested from both directions: bottom-up detail and top-down reasonableness.
THE LAW
Never trust a number you haven’t benchmarked.
08
An estimate is only as valuable as the record behind it.
Months after a project is awarded, someone will ask: Why did we carry this number? Which drawing revision did we price? Who agreed to this exclusion? What changed between estimate versions? Which subcontractor clarification led to this scope decision?
Without documentation, those answers disappear into inboxes, spreadsheets, meeting notes, and individual memory.
HOW AI CHANGES THE WORK
Modern systems can preserve estimate versions, organize documents, summarize revisions, maintain decision histories, and make project information searchable long after the original estimate was completed.
That record becomes valuable during procurement, construction, change management, and eventually the next estimate.
Every project should make the estimating organization smarter.
THE LAW
If the decision matters, preserve the reason behind it.
There is a temptation to describe AI in construction as replacing the estimator. That misses the point.
The most valuable parts of estimating have never been copying numbers between spreadsheets or searching through a 2,000-page specification for a paragraph.
They are judgment. Experience. Risk recognition. Knowing when a number does not look right. Knowing which subcontractor needs another phone call. Understanding where the documents leave room for interpretation. And ultimately deciding what the company is willing to put its name and money behind.
AI can handle more of the mechanical work surrounding those decisions. That gives estimators more time to actually make them.
The companies that benefit most from AI will not be the ones that abandon estimating fundamentals. They will be the ones that apply those fundamentals more consistently, across more projects, with less manual work.
That is what the 8 Laws are about.
The 8 Laws of Estimating
01 — Define the Scope of Work
Define the work before you price the work.
02 — Quantify the Work
Measure before you assume.
03 — Develop an Opinion of Cost
Build an independent opinion of cost before the market gives you one.
04 — Level the Bids
Never compare price without comparing scope.
05 — Identify Trade Interfaces and Overlaps
The greatest scope risk often exists between the trades.
06 — Confirm Trade Requirements
Verify the requirements behind the number.
07 — Evaluate Key Metrics
Never trust a number you haven’t benchmarked.
08 — Document Everything
If the decision matters, preserve the reason behind it.
Consight is building the preconstruction workspace around these principles—from understanding project documents and building scopes to collecting bids, leveling proposals, reviewing risk, and documenting the final award.
Put the 8 Laws into practice on every project.
Talk to sales