Tyler Eiland is the managing partner at Axsus Civil Development Company in Boise. He spent years tracking large industrial projects in oil and gas, chemicals, and semiconductors, and he is entirely self-taught.
In this conversation, he talks about what that work taught him, and why he thinks most companies chasing AI are solving the wrong problem.
What's the problem nobody in construction is actually solving?
The gap between the field and the office. Work gets done faster than it gets documented, and by the time anyone upstream sees it, decisions have already been made without the data behind them.
“You could walk around this valley and you can ask people how much money they made yesterday and nobody could tell them,” Tyler said.
It doesn't have to be that way. On his projects, earned value gets tracked from the field to the office in real time. Crews log progress against unit rates, so planned production and actual production sit side by side. If the plan was 500 feet of water main and the crew installed 400, that shows up the same day, not at the end of the month.
“You could walk around this valley and you can ask people how much money they made yesterday and nobody could tell them.”Tyler Eiland, Managing Partner, Axsus Civil Development
How far behind is the industry?
Further than people assume. Tyler described working a power plant project in Mountain Home for a company doing $30 billion in revenue. They still run on paper.
He spent years holding large industrial projects together with Excel and VBA. Fifteen spreadsheets, fifteen people updating them fifteen different ways. That's where the conviction came from. “Everything should be streamlined and it's not.”
So where does AI actually fit?
This is where he pushed back hardest.
“AI is a buzzword,” Tyler said. “If you have no architecture, if you have no infrastructure, if you have no data, what is AI? People throw this word AI around. Well, if you don't have data for the AI to process, AI is worthless. AI is a chatbot.”
His sequence is the opposite of how most companies approach it. Aggregate the data first. Capture it at the point of work, structure it, get it into one place. Only then does anything intelligent sit on top of it and mean something.
“If you don't have data for the AI to process, AI is worthless.”Tyler Eiland, Managing Partner, Axsus Civil Development
A lot of contractors say they wish AI could do things it can't. What's your read on that?
He disagrees with the premise.
“There is not a limitation. You can make it do whatever you want. It's just a matter of you implementing those workflows.”
The gap, in his view, isn't capability. It's that nobody built the system underneath. “They're talking about software. You have to have architecture for the AI to do any of those things.”
That's the line worth sitting with. Most of the wish list contractors have, searchable knowledge from past jobs, automated field reports, estimating that pulls from real cost history, isn't waiting on better AI. It's waiting on somebody to build the plumbing.
Tyler Eiland is the managing partner at Axsus Civil Development Company in Boise, Idaho.
Part of AIdaho Conversations, a series with Idaho business leaders on how they’re actually using technology and where it falls short.