Jeremy Sexton

AI is here to stay.

That statement creates a lot of different reactions.

Some people are excited.
Some people are overwhelmed.
Some people are honestly worried it is going to take jobs away.

I understand that fear.

But I do not think the real story is simply “AI is replacing people.”

I think the more accurate statement is this:

AI is going to change almost everyone’s job in one way or another.

And in some ways, we have already been living with pieces of this shift for years. Not necessarily full AI, but automation.

Out-of-office replies.
Excel macros.
Bid board notifications.
Template-driven emails.
Workflows that take one action and trigger several others.

Those are simple examples, but they all point to the same bigger idea:

The way we process information has been changing for a long time.

Today, I spent time looking at my own proposal workflow through that lens.

Right now, my proposal process already has multiple steps. I submit through the bid board, and I also send a direct email with the proposal attached. That gives the GC more than one path to receive, review, and respond.

But now there is another layer coming into the picture:

AI-assisted bid leveling.

Whether we like it or not, that is becoming a real part of how proposals may be reviewed, compared, sorted, and scored.

So I started thinking about something different.

What if one proposal is written for the human estimator, project manager, or owner’s rep…

And another version is structured in a way that is easier for an AI-assisted bid leveling system to read?

The first version is the traditional proposal.

Cover sheet.
Project-specific title page when available.
Scope of work.
Line-item pricing with descriptions.
Logistics and communication notes.
Takeoff source page.
Marked-up takeoff documents showing how the numbers were built.

That version tells the story.

It shows the thought process, the scope logic, the documents reviewed, and the estimating approach behind the number.

The second version is different.

It is colder.
More structured.
More direct.
More data-driven.

Strict line items.
Clear inclusions.
Clear exclusions.
Unit pricing.
Quantities.
Cost per unit.
Scope assumptions tied directly to each line item.

That version is not necessarily better or worse.

It just serves a different reader.

And that is where I think this conversation gets interesting.

A proposal written only for a human may not be structured cleanly enough for AI-assisted review.

A proposal written only for AI may miss the real-world construction logic that an experienced estimator would understand immediately.

Because construction is not always clean.

Sometimes the plan says one thing.
The spec says another.
An addendum changes the scope.
A hyperlink gets buried in an old reused document.
A finish schedule conflicts with a detail.
The ITB leaves something out that everyone in the field knows still matters.

That is the part where experience still matters.

AI can help process information faster.

But it does not automatically understand the operational truth of a project unless the information is structured in a way it can read.

That does not make AI bad.

It makes the way we communicate even more important.

To me, this is not about AI taking over construction.

It is about AI forcing us to get sharper.

Sharper scopes.
Sharper exclusions.
Sharper alternates.
Sharper assumptions.
Sharper documentation.
Sharper communication.

And honestly, that is preconstruction.

We are already professional future-preppers.

We read incomplete information, identify risk, build assumptions, price unknowns, and try to tell the clearest story possible before the job ever starts.

AI is not replacing that.

It is changing the way that story may need to be told.

The best tool in construction may not be made by Milwaukee, DeWalt, or Festool.

It may be the tool that helps us process information faster, communicate more clearly, and spend more time on the human relationships that still win work.

Because at the end of the day, relationships still matter.

Trust still matters.

Experience still matters.

The question is:

How are we adapting our workflows so AI makes us better instead of making us easier to misunderstand?