Hi Reader,

The most visible problem and the most expensive problem are almost never the same answer.

I learned this early in advisory work— and I’ve had it confirmed in nearly every firm I’ve worked with since.

The Session

A proposal manager at an engineering firm called me with a familiar problem.

Her team was using AI for occasional tasks— drafting emails, summarizing documents, cleaning up formatting. The tools were good. The results were fine. Nobody felt like they were getting ahead.

We spent the first 30 minutes of session one on a single question:

*What’s the most expensive problem in your AI situation right now?*

Not the most frustrating. Not the most visible. The most expensive.

Her answer surprised her.

Proposals. 200 a year. Each first draft taking 12 to 16 hours— compliance checking, reformatting, pasting from past projects, rebuilding the same structure from scratch every time.

She’d been solving smaller problems. The expensive one had been sitting unaddressed for years.

The Workflow

We looked at what the work actually was.

Most of it was copy-paste, reformatting, compliance language pulled from templates, and section structures that were nearly identical from proposal to proposal.

AI handles all of that now.

Her first drafts come out in 3 to 8 hours. The savings run 4 to 13 hours per proposal depending on complexity.

200 proposals at 8 hours saved on average is 1,600 hours.

At conservative burdened rates for an AEC firm, that’s $160,000 in recovered labor.

One role. One workflow. 30 days in.

What This Is and Isn’t

That number is real. It’s also not magic.

It came from someone who knew her firm’s proposals well enough to tell me exactly where the friction lived. It came from building a workflow around her templates, her compliance language, her section structure— not a generic tool pointed at a general problem.

The AI didn’t find the $160,000. The audit found it.

AI is the mechanism. The strategic question— where is the most recoverable time, right now— is what unlocks the number.

Most firms never ask that question clearly. They run experiments. They run training days. They watch the tools sit underused.

The gap between “we have AI tools” and “our AI tools are returning value” is almost always a strategy gap, not a capability gap.

What Advisory Actually Looks Like

One hour a week.

We start with the diagnostic: where is the most recoverable time in your current workflows?

The most visible problem and the most expensive problem are almost never the same. That gap is usually where the first 30 days of advisory goes.

Then we work through it. Department by department. Workflow by workflow.

I work with engineering and construction firms specifically. I know the proposals, the RFIs, the staffing plans, the change orders. I know where generic AI tools don’t know your firm yet and where that gap costs you.

If your AEC firm is past the experimenting stage and ready to find out where the real value is— let’s talk.

​https://book.dancumberland.com/ai-strategy​

Keep building,

-Dan

P.S. Last week I wrote about why the AI skills your team installs should be built for your firm, not borrowed from a stranger’s library. Same principle, different scale.

​https://go.dancumberlandlabs.com​

P.P.S. Next week I’m back to the practical side— a skill your team can build and deploy this week.