Why is a broad service offering a liability for an engineering firm now?
AI supplies breadth faster than you can hire for it. Depth in one lane is what a buyer can check, and what wins the work.

Because a machine now supplies breadth faster than I can hire for it, and depth in one lane is the one thing it cannot supply on my behalf. That is my read of what changed. When a buyer asks an AI assistant which firm does a specific kind of work, the answer is built from evidence of that work. A firm listing twelve services has little to be ranked on. A firm with one deep lane has years of finished work to be ranked on.
Why did breadth used to work?
In my experience breadth was survival before it was strategy. You take the pump station, then the subdivision, then the small treatment plant, because taking the work is how the first two years of payroll get paid.
It was also a signal a buyer could read. A wide service list said this firm has seen enough not to be surprised. When checking a firm was slow and expensive, that signal carried deals.
My read is that checking got cheap. A buyer can see what a firm has finished in the lane they are buying, and see the same about the firm down the road, without a phone call.
The signal changed hands. It belongs now to whoever can show finished work in one lane, not to whoever can name the most lanes.
What did AI actually take over?
The first draft. The boilerplate permit narrative. The standard specification section. The coordination memo saying what everyone already agreed to. That is the work that used to pay for a junior engineer's week and for a generalist firm's retainer.
It did not take over judgment. It cannot tell you which of two options the reviewer will accept, which number in the model is wrong, or which client instruction will cost three weeks in review.
Judgment is not spread evenly across twelve services. It concentrates in the work you have done often enough to be right about, and that concentration is the whole asset.
The same tools also make breadth cheap to claim. Any firm can turn out polished documents across a dozen service lines, so the length of the list stops telling a buyer anything. My read is that this is the change few owners have priced yet.
What does depth look like when it is working?
It looks like a library. A firm working one lane takes the jobs it has already finished in that lane and loads them into the tool the team uses, so the finished work carries the assembly and the engineer spends the day on judgment.
The firm does not get faster because it bought software. It gets faster because it had the work to put in the software. Years spent on one kind of work is a different asset from the same years spread across twelve.
I think the list is the liability now and the library is the asset. The tool is available to everyone in the market. The library is yours, and no vendor can sell it to your competitor.
Why does breadth cost so much in an hours model?
Every service on the list needs a senior person who can hold it. That cost never appears in the proposal. Add a service and you add one more thing only a few people can sign off on, one more set of reviewers to keep up with, one more kind of mistake to catch.
Under the hours model that cost lands on the owner. The founder is the one who learned all twelve, so the founder stays in the middle of the work because nobody else holds the knowledge. Growth means hiring more people to feed the founder's review queue, which is not growth. It is a longer week.
Modern delivery does something different with the same list. It cuts it down and puts what remains into a library and a system, so the work carries itself and the owner is not the bottleneck. The point is not to look narrow to a buyer. It is to build a firm where the knowledge lives somewhere other than one person's head.
A generalist is rarely losing on talent. It is losing on evidence: the buyer is not asking who could do the work, but who has done it, and how many times.
How do you know whether breadth is running your firm?
Four questions, answered honestly, separate a deliberate offer from a habit.
- Can you name your lane in one sentence that includes the client type? Not water and wastewater, but discharge permit renewals for industrial dischargers.
- If a buyer asked for your last five projects in that lane, could you send them today without leaving any out?
- Does your tooling know that lane, or does it only know your document templates?
- When work outside the lane arrives, who decides to take it, and on what basis?
If the answers are mostly no, breadth is not insurance. It is overhead you pay every month in slow onboarding, thin margins and proposals you should never have written.
Where is the lane already hiding in your numbers?
You do not have to invent a lane from a list of markets. Read your own revenue instead.
Sort two years of invoices two ways, by type of work and by client type. Most owners have never done it and are surprised by the cell that repeats. It is usually the work that came back without being chased, and the work your steadiest engineer finished early.
Then check margin. The lane you want is where you were not the cheapest and still won. If you won on price, you were not the specialist that day. You were the fallback.
What about a principal starting a firm, and do you have to turn work away?
If you are leaving a firm to start your own, breadth is the trap in the doorway. You will be tempted to list every service you touched over a career, because the list feels like proof of experience. To a buyer it reads as one more firm that will take anything.
A narrow lane does something else for a new firm. It lets you hire a first engineer and point them at a library instead of asking them to be good at everything in month one. That is the difference between a small firm that gets faster and a small firm that gets tired.
None of this means turning work away. It means you stop calling the outside work part of your offer. Take it when it pays and when you can staff it well, deliver it, and then stop presenting it as what you are.
The offer and the work are two different things. Every firm takes some odd work. The firms that stay slow are the ones that keep advertising it on the front page.
What to do next week
- Pull two years of invoices and sort them by type of work and client type. Count, do not estimate.
- Write the one sentence. If two candidates sound right, pick the one with the most repeat clients.
- Gather the finished work in that lane into one folder and load it into the tool you already use. Finished work beats a new subscription.
- The next inquiry outside the lane, decide before you answer: take it as an exception, or pass it to someone who lives there.
Then cut your service list down to what you would defend in a room with a buyer who has already checked. That list is shorter than the one on your site, and it is the one that wins.
From the book Engineer Your Freedom
Michael Groselle, P.E., is the founder of Modern Engineering Solutions, a multi-state water and wastewater engineering firm. He left salaried engineering in 2017 and built the firm from one resignation letter and zero clients to more than $2M in annual revenue, with a team of 37 and 525+ completed projects.
Sold on Amazon. Prefer the Kindle edition? Buy it on Amazon ($9.99).