The AI Playbook for Construction CFOs in 2026

A Practical Approach to the Hype — and How to Focus Your Organization

Brown Edwards & Company | Winter Construction Webinar Series

Why We Hosted This Conversation  

AI is making news every single day, and the construction industry is no different. From AI-assisted bidding and scheduling software to autonomous dozers and robotics, the technology is already reshaping how we build. At our Winter Construction Webinar, we had the pleasure of welcoming Geoff Marsh, Managing Partner of Amend Consulting out of Cincinnati, Ohio, to cut through the noise and give CFOs and financial leaders a practical framework for 2026.

Amend Consulting is a management consulting firm with roughly 100 employees that works with organizations ranging from $20 million to a couple billion in revenue — right in the wheelhouse of the construction companies we serve. Geoff's focus on enterprise technology selection, data architecture, and digital transformation made him exactly the right person to help us answer the question every construction executive is wrestling with: how is our organization going to use AI?

First, Let's Be Clear About What AI Actually Is  

One of the most valuable things Geoff did was establish a real baseline. AI — artificial intelligence — is the ability of machines to perform tasks that simulate human thinking. It has been around since the mid-1900s. What changed recently is generative AI: the use of large language models (LLMs) to predict and contextualize written language in a way that mimics human interaction. The underlying technology isn't new; what's new is the capacity and resources available to run it.

Here's something worth understanding: when an AI gives you an answer, it is not thinking like a human. It is predicting the most statistically likely response based on the data it was trained on. It doesn't truly 'know' the answer the way an auditor or attorney knows something. It is making a mathematically informed guess. That distinction matters enormously when you are thinking about how to deploy AI in your organization.

Geoff also demystified 'agentic AI,' one of the buzziest terms in the market right now. Agentic AI is simply the linking together of small automated tasks into a workflow. Think of the sequence of steps that happen after you receive an email — reading it, routing it, responding. Connecting those steps through automation is the essence of agentic AI, and it is where organizations are seeing the most practical value today.

The Circumstance Construction Finds Itself In  

Every software vendor is pitching AI. The media is hyping it. But construction leaders are not getting clear, practical guidance on how to actually build a plan. Geoff identified exactly why that gap exists: AI has a deep dependency on quality data and quality underlying systems, and construction is an industry notorious for running projects off of manila folders, printed spreadsheets, and handwritten invoices.

At a recent CFO roundtable Geoff attended, the unanimous answer when executives were asked what they are doing with AI was this: for the past six to eighteen months, they have been re-implementing or bringing on new operational systems so that their data quality is good enough to support AI tools. That is the real starting point — not buying an AI tool, but getting your core systems in order first.

The AI maturity model makes the timeline concrete. Getting core systems into a place where AI can leverage them: one to two years. Going from early AI exploration to AI embedded in operations: another twelve to twenty-four months. That means most construction organizations are realistically two to three years away from having AI woven into their day-to-day practices. Setting realistic expectations is not pessimism — it is responsible planning.

Where AI Is Creating Real Value in Construction Right Now  

Despite the longer journey ahead for enterprise-wide AI, there are isolated areas where construction organizations are finding real wins today. Geoff highlighted four current trends.

Design and BIM tools are integrating AI to streamline engineering design, with platforms like Transcend (which helps engineers design water treatment infrastructure using decades of embedded institutional knowledge) and Augmenta being early examples. These tools require very specific, narrow use cases to work well, and building your own from scratch is not realistic for most construction organizations — you need to find the right third-party software that matches your specific need.

Modular and prefab construction is another area of progress. Because prefab moves construction closer to a manufacturing model, AI can better connect field management systems with enterprise resource planning systems to improve logistics, inventory planning, and scheduling decisions.

Cloud and mobile technologies are enabling smarter data capture on job sites — drones for site monitoring, tablets for time and production tracking — with AI helping connect that captured data back to core systems automatically, reducing manual re-entry.

As for the 'AI buzz' in every software platform? Vendors have broadly reclassified their products as AI-powered. Some of that is legitimate; some is marketing. The key question to ask is whether the AI feature actually solves a specific problem you have identified, not whether the tool has an AI label on it.

Will AI Replace CFOs and Accountants?  

This question came in live during the webinar and it deserves a direct answer: no, not in any near-term horizon — and here is why.

There are specific tasks that AI can automate for finance teams. Invoice verification against purchase orders, three-way matching between invoices, purchase orders, and project manager approvals — those are examples of simple, rule-based tasks that AI handles well. If your team can process more invoices with less manual effort, that frees your financial professionals to focus on higher-value work.

But cash flow strategy, project profitability analysis, understanding how your organization takes on new work and manages vendor relationships — those functions require strategic reasoning and contextual judgment that AI is nowhere near replacing. The role of the CFO and accountant should shift toward more strategic planning and less manual entry, not disappear. Organizations that assume AI can fully replace those functions should be very cautious about where that thinking leads.

The Two Pitfalls to Avoid  

Geoff was direct about the two most common mistakes he sees organizations make as they try to adopt AI.

The first is buying tools before understanding the problem. When organizations without strong IT leadership hear that a tool can do something exciting, the temptation is to demo it, buy it, and figure out how to fit it into the organization later. That is exactly backwards. A strong technology leader starts by understanding what problems the organization needs to solve, then identifies whether and how technology can address those problems. The construction industry's history of failed ERP implementations came from the same root cause: buying before designing.

The second pitfall is overestimating AI. Construction organizations that assume AI will autonomously replace entire functions are setting themselves up for disappointment and risk. AI in construction is not autonomous. It needs humans to review its outputs. The concept of 'human in the loop' — having people review AI recommendations at critical decision points — is essential, not optional. AI hallucinates. It predicts. It needs oversight. Designing your systems around that reality will put you far ahead of organizations that ignore it.

The Principles for Building Your AI Playbook  

Geoff closed with four foundational principles every construction organization should build their AI roadmap on.

Treat data as a competency. What counts as data is changing every six months. Your business leaders need to understand the difference between data that belongs in quality source systems and information that is just informational. That clarity has to be defined, trained, and maintained.

Establish quality source systems. If your project managers are still running reviews from a manila folder rather than entering everything into your core financial management system, you have a foundational problem to solve before AI can help you. Getting all project data into modern, structured systems is the prerequisite.

Understand that AI is a feature, not a replacement. In 2026, AI is a scalpel — precise, useful for specific tasks. It is not a chainsaw that can replace Procore or Viewpoint. Use it for the specific manual tasks where it creates the most efficiency, not as an overarching transformation of how your whole operation runs.

Establish an AI use policy — now. This was Geoff's most urgent call to action. Around 60% of webinar attendees reported using AI tools at work already, yet the majority of their organizations have no official policy or plan. That gap is the biggest risk you face. An AI use policy does not need to be complicated: define the purpose, the guiding principles, the approved tools, the prohibited uses, training expectations, enforcement, and governance. Who owns the policy? How often will you review it? Starting there removes the greatest organizational risk immediately.

And if you are a growing construction organization looking ahead to the next three to five years — Geoff offered one more challenge. Some of the most tech-forward construction organizations Amend works with have broken the traditional model and brought in CIOs and executive-level technology professionals from outside the industry. The organizations doing that are head and tails above everyone else. If technology is going to be a genuine competitive advantage for your firm, it may be time to consider whether your current leadership structure can get you there.

The Bottom Line  

AI is not going away, and construction organizations that ignore it will get left behind. But the path forward is not about chasing the latest tool or buying into the hype. It is about building a solid foundation: modern core systems, quality data, a clear AI use policy, and a realistic multi-year roadmap. The organizations that take that practical, deliberate approach are the ones that will see real, lasting value from AI — not just the organizations that bought the most tools.

We are grateful to Geoff Marsh and Amend Consulting for bringing this level of clarity and practicality to our clients. If you would like to learn more about Amend Consulting, visit amendllc.com or reach out to anyone on our team at Brown Edwards and we will make the introduction