86% of Large Contractors Expect an AI Advantage. Only 69% of Mid-Size Firms Do

The gap in the 2025 Dodge survey is not about tools. It is about who decides where AI runs and where judgment holds the line.

Industry Trends
Renata Pinheiro
Renata Pinheiro
July 6, 2026 · 6 min read
Two mid-size general contractor leaders reviewing an estimate together at a job trailer desk in warm late-afternoon light

On June 30, 2026, Trimble shipped AI takeoff for MEP estimating that it says cuts manual takeoff setup by as much as 60 percent, and it reports more than 4,000 contractors already using AI inside its estimating tools (Trimble, 2026). The same day, Higharc raised a $95 million Series C led by Insight Partners to scale AI that turns a floor plan into priced material quantities (PR Newswire, 2026). Fast estimating is no longer a bet. It is arriving, funded, at scale.

Here is the number that should hold a mid-size GC's attention. In the 2025 Dodge Construction Network survey, 86 percent of large contractors said they expect AI to give them a competitive advantage. Among small and mid-size firms, only 69 percent did (Dodge Construction Network, 2025). That 17-point gap is not about who has the tools. Trimble and Higharc will sell to anyone. It is about who has decided how to use them.

Speed is becoming the baseline, not the edge

The same Dodge survey found 87 percent of contractors expect AI to have a meaningful impact on their business, yet only 19 percent have actually adapted their workflows (Dodge Construction Network, 2025). The tools are ahead of the decisions. When 4,000 contractors are already running AI takeoff inside one vendor's product alone, the speed those tools deliver stops being a differentiator and becomes the floor everyone stands on.

"

When the fast tool is standard issue across the field, speed stops being an advantage and starts being the floor.

That is the trap in the 86 versus 69 gap. Large firms are not pulling ahead because they bought faster software. They are pulling ahead because they have the staff and the process to decide where that software is allowed to run and where a person still has to sign. A mid-size firm that buys the same tool without making those decisions gets the speed and none of the edge.

The part the tool was never going to do

Read what contractors themselves say they are afraid of. In the Dodge survey, 57 percent named a lack of reliability or accuracy in AI output as their chief concern, 54 percent named data security, and only 26 percent rated their current data quality as high (Dodge Construction Network, 2025). Those are not complaints about speed. They are a warning that the output needs a human check, and that most firms know their data is not clean enough to trust blindly.

That is where the real work moves: from counting to judging. A takeoff tool can now surface the places where no subcontractor bid covered a piece of work, or where two subs each assumed the other would. That is a genuine gift. But a flag is not a decision. Which gaps are real risk, how much to carry, whether the addendum revised the assumption three times over, whether this owner has a history of disputing exactly this line: none of that is in the takeoff. It is in the estimator.

The 57 percent who worry about accuracy are pointing straight at that gate. The question is not whether to trust the AI. It is where a person is required to look before the number leaves the building.

Where the recovered hours should go

Whatever time AI hands back, most firms have never had it to spend on their own most important question. Construction business author George Hedley, who has surveyed thousands of contractors, has long reported that fewer than 6 percent know and track their bid-hit ratio, the rate at which bids actually convert to wins (Hedley, cited by For Construction Pros). The manual takeoff load consumed the hours that measurement would have taken.

Spend the recovered time on judgment and measurement, not on more volume. The temptation is to point freed capacity at bidding everything that moves. That is how a firm turns a faster tool into faster margin erosion. The firms closing the advantage gap treat that time as Presence Time: estimators and preconstruction leads applying attention to go or no-go calls, scope-gap risk, and which jobs are actually worth winning.

"

The bid you win is the one you had time to read.

This is why the boundary matters more than the tool. Inside estimating, AI handles the counting and cross-checking on its own. It recommends quantities, exclusions, and bid comparisons for a person to approve. A person leads the scope-gap risk call and the go or no-go decision, with AI supporting. Those three functions are not interchangeable, and a firm that buys the software without naming which function sits where has bought speed with no steering.

That naming is the work we do before any tool selection. We start with the questions the estimating team actually needs answered, set the boundaries on what AI decides versus what a person leads, and only then draw the Workflow Map that shows the whole path. The map is the artifact. The judgment about where humans hold the gate is the point. We covered the deeper version of this question in what should humans still own, and estimating is where the answer gets expensive if you get it wrong.

Closing the 17-point gap

The Dodge number is a choice, not a verdict. Large firms expect an advantage because they have decided where AI runs. Mid-size firms can decide the same thing, and they can do it faster because they have fewer layers to move through. The tool is available to everyone on the same day. What is not evenly distributed yet is the discipline of naming where a person holds the decision.

If you are deciding how to bring AI into estimating, the first move is not a demo. It is a clear map of where your people hold the decision, and where the machine is allowed to run. We build that map with you before a dollar goes toward a tool.

Sources

Common questions
Does AI estimating give general contractors a competitive advantage?
Not from the tool alone. In the 2025 Dodge Construction Network survey, 86 percent of large contractors expected AI to give them a competitive advantage, but only 69 percent of small and mid-size firms did. The advantage comes from deciding where AI runs and where human judgment holds the line, not from owning the software.
How much time does AI save on construction takeoff?
Trimble, which added AI takeoff to its MEP estimating tools on June 30, 2026, reports the features cut the time for manual pre-takeoff and takeoff tasks by as much as 60 percent, with over 80 percent savings on specific queries like historical pricing lookups. These are vendor-reported figures from Trimble; independent, audited productivity benchmarks across the industry are still thin.
What is the biggest risk with AI in construction estimating?
Reliability and data quality, not speed. In the 2025 Dodge survey, 57 percent of contractors named a lack of reliability or accuracy in AI output as their chief concern, 54 percent named data security, and only 26 percent rated their current data quality as high. Those numbers say the human check on AI output is the part that matters most.
What is a scope gap in construction bidding?
A scope gap is work that no subcontractor bid included, or that two subs both assumed the other would cover. Unpriced scope gaps are a common source of margin loss on won jobs. AI tools can now flag gaps across multiple sub bids quickly, but deciding which gaps are real risk and how to carry them is a judgment call that stays with the estimator.
Can AI replace a construction estimator?
No. AI replaces the manual counting and cross-checking inside estimating, not the judgment. It can recommend quantities, flag exclusions, and compare bids. A person still decides what to carry, what to exclude, what the addendum changed, and whether the number is one the firm can stand behind. The role shifts from data entry to decision-making.
Why do most contractors not track their bid-hit ratio?
Construction business author George Hedley, who has surveyed thousands of contractors, has long reported that fewer than 6 percent know and track their bid-hit ratio, the rate at which bids convert to wins. Most estimating teams have been consumed by manual takeoff volume. Recovering estimator hours through AI is a chance to finally measure and improve that ratio.
What should a mid-size GC do first before buying AI estimating software?
Map the estimating workflow and decide where the human decision gates sit before selecting a tool. Name what AI handles on its own, what it recommends for a person to approve, and what a person leads with AI support. Buying speed without those boundaries means you lose money faster, not less.
Is fast bidding enough to win more construction work?
No. In the 2025 Dodge survey, 87 percent of contractors expected AI to have a meaningful impact, but only 19 percent had actually adapted their workflows. Once AI is available to every bidder, speed becomes the baseline. The durable advantage is what a firm does with the recovered hours: tighter scope review and better go or no-go decisions.