AI for Estimators:
How to Strengthen Your Estimating Process

Part 2: How ChatGPT Can Strengthen Your Estimating Process

In Part 1 of this blog, we discussed why AI belongs in the construction and technology industries, and now it’s time to talk about how that actually works in practice.

Recently, at Clark Building Technologies, we responded to a large, thousand-page RFP for a major technology installation. The scope included low-voltage cabling, access control, video surveillance, wireless, and audio-video systems, with a long list of requirements. It was the kind of bid that can overwhelm even an experienced estimating team, and anyone who has been through this process knows the challenge. The documents are massive, the deadline is tight, and one missed requirement can cost you the job or come back to haunt you later as an expensive problem.

We’ve dealt with RFPs like his before, but this time, we brought ChatGPT into the estimating process.

How ChatGPT Helped Us Tackle a Massive RFP

As you might expect, the first step involved feeding the RFP documents into ChatGPT and asking it to extract every requirement tied to Divisions 27 and 28. This part is important; we took the time to carefully review the initial extraction from ChatGPT to verify that the AI didn’t miss anything. From experience, I can tell you with confidence that AI is not something you trust blindly, especially on a project of this size.

That said, ChatGPT dramatically reduced the time this step normally takes. It completed what usually requires hours of manual reading and cross-checking in a fraction of the time. Just as important, the structured breakdown of the output gave us more confidence that no requirements were overlooked, because everything was laid out clearly and systematically. That early clarity allowed us to quickly determine whether the project was a good fit, where we needed partners, and where potential risks existed. So, instead of spending days just trying to understand the job, we were able to make a bid/no-bid decision quickly.

Building a Requirements Matrix That Eliminated Guesswork

Once we confirmed the project was worth pursuing, we used ChatGPT to build a detailed requirements matrix. This process involved tying each requirement back to its source, including page numbers and descriptions, which matters more than most people realize. Anyone familiar with it knows that estimating mistakes often come from small details buried deep in specifications or other RFP documentation. Hunting through PDFs, relying on memory, or re-reading sections under deadline pressure is where errors creep in.

For us, the requirements matrix removed that friction, keeping everyone aligned on what the job actually required, reducing second-guessing throughout the estimating process.

Speed Without Sacrificing Accuracy

Having dealt with RFPs for a while now, I feel confident stating that most companies face a different set of challenges today. Competition is higher than ever, many estimating teams are overloaded, and finding experienced estimators can be challenging, to say the least. Even strong teams are asked to do more work with fewer people, all while bids are becoming more complex, and deadlines remain unforgiving.

For us, ChatGPT helped close that gap, but not in the way you might think. It did not replace our need for qualified estimators; instead, it handled time-consuming work, allowing our team to focus on pricing the work correctly, thinking through logistics, and applying real-world judgment. That balance between speed and accuracy is difficult to achieve without help, and this approach allowed us to move faster while still improving quality.

Using ChatGPT to Support the RFI Process

An important part of any RFP is developing RFIs, or questions that clarify gaps, conflicts, or unclear requirements in the specifications and drawings. This step serves two critical purposes:

First, the answers help you fully understand the scope of work. Clear answers reduce assumptions and risk, and they ensure you price and deliver exactly what the client is asking for.

Second, well-written RFIs demonstrate to the client or General Contractor that you understand your trade. Good questions show experience and foresight, and credibility matters when it comes time to select the contractor who will be awarded the work.

ChatGPT supported this process by helping us to identify areas in the documents that needed clarification, but we still relied on our own judgment to decide which questions mattered and how to frame them. In this way, the tool helped us work through large volumes of information far more efficiently, and once the client issued RFI responses, we uploaded those answers back into ChatGPT along with the rest of the RFP documentation. I can’t stress the importance of this critical step enough. When you upload them, those responses become part of the contract documents and often override or clarify the original specifications and drawings. By incorporating the RFI answers into the same workspace, ChatGPT can work from the most current and accurate information available, allowing the summaries, requirements lists, and proposal language to reflect the clarified scope, not the original assumptions.

This approach of adding RFI responses to the RFP documentation doesn’t replace experience; it reinforces it.

Sharper Proposals Written in Our Voice

When it came time to write the proposal, ChatGPT helped us tighten the language and structure, providing past proposals so it could match our tone and writing style. From there, it helped tailor the new proposal to the specific requirements of the RFP, including the clarifications issued through RFIs. When we finished, the final proposal was clear, complete, and aligned with what the client actually asked for, reflecting a strong understanding of the scope and a disciplined approach to execution.

And most importantly, we won the contract.

Why This Matters When You Are Estimating Real Work

When you look at the full process, RFP review, requirements extraction, RFIs, RFI responses, and proposal writing, it becomes clear that estimating accuracy is all about managing information better, not working harder. Most estimating problems happen because information is fragmented, assumptions change, or key details get lost as documents evolve. By keeping the entire document set aligned in one place and using AI to help manage it, we significantly reduced that risk. Our estimators didn’t chase documents or rely on memory; they continually worked from a consistent, up-to-date understanding of the job.

Not surprisingly, we have found that this process leads to better bids, fewer surprises, and more confidence in what you are committing to deliver.

Why Clients and General Contractors Notice

Before we finish up, there is another benefit to this process that often gets overlooked. When your RFIs are thoughtful, and your proposal reflects a clear understanding of the clarified scope, clients and General Contractors notice. It signals professionalism, showing that you understand your trade and pay attention to details others may miss. When it comes to an award decision, that confidence can be the difference between winning the job and losing it to a competitor.

It’s important to remember that AI is a tool; on its own, it does not create that credibility. Experience does, but AI helps support a disciplined process that makes that experience visible.

If Your Estimating Process Feels Heavy, We Can Help

If your team is overwhelmed by large RFPs, tight deadlines, and document overload, this approach can make a real difference. Having better tools with a clear way to use them is a heck of a lot more efficient than replacing estimators or overhauling your department. We know where AI helps, where it does not, and how to apply it without introducing risk because we’ve already built this process into our own estimating workflow.

If you want help strengthening your estimating process, improving accuracy, and giving your team more confidence when bidding complex work, reach out to us at 301-456-6931 or email [email protected]. This is one of the fastest ways we know to reduce mistakes, speed up turnaround time, and put your business in a better position to win more work, and we’ll be happy to walk through how it works in the real world.