I’m often asked about the benefits of using AI in the Takeoff process. At the surface level, it is an easy calculation: Do more in less time to produce the same bids at less cost or more bids at the same cost.
However, there are several nuances to the benefits worth exploring.
Our product is intended to act as an assistant to estimators. It’s not an automatic estimator where you press a button, and it creates a bid ready for you to submit. And it’s not going to convert estimating into exacting. It saves time and can certainly improve quality. Let’s unpack quality improvements first and then delve deeper into time savings.
As an assistant, AI handles tasks like counting and routing with minimal input. However, the results should still be reviewed and refined. AI won’t be 100% accurate, but it can catch mistakes that might be overlooked in a rush to meet deadlines. The approach is similar to having two estimators create takeoffs independently and compare results. If discrepancies arise, further review is needed. The key benefit is more accurate results at a lower cost than traditional double estimating.
The time saved by AI directly contributes to improved quality. With AI performing the more mundane counting and routing tasks, the estimator gains time to analyze the specs, contracts, and drawings more thoroughly to identify unusual requirements.
The time savings are easier to analyze. The combination of AI counting and routing plus QA by the estimator must take less time than having the estimator do the same tasks manually. The key factor is the accuracy of the AI. As Den, our CEO and founder, often points out, the benefits of AI do not increase linearly as accuracy improves. The benefits grow on a curve. If AI did not require QA, benefits would increase linearly. With QA, the situation changes, as illustrated in the diagram below.
With an accuracy of 20%, the time to QA and fix things produces very little savings and more frustration than it is worth. At 50%, the situation is better, but there is still too much QA time required to create a lot of value. Somewhere past 70% accuracy, the situation starts to change. At 90% accuracy, the time savings are significant. Of course, improvements in QA tools and tools for manually completely missing data change the equation with more benefits accruing even with less accuracy.
Our goal for 2025 is 95% accuracy on 80% of all projects. Depending on the quality and standardization of the construction drawings, some projects will perform much better and others worse. We’re tackling the problem from three directions: first and most importantly, constant improvements in AI accuracy to support more standards and lower-quality construction drawings; second, by investing in ever-improving QA tools; and third, by producing better tools for manual takeoff.
AI will drive major changes in our industry over the next few years. Like any new technology, it requires adjustments in processes and, to some extent, culture to fully leverage its potential. Given the rapid pace of technological advances and the tangible value AI offers, these changes will happen faster than many other innovations in the industry. We invite you to be part of this journey with us.
Curious about how other companies are already leveraging AI in their takeoff process? Check out our Starr Case Study to see real-world examples and learn how AI transforms the industry.