The 20% That Matters: AI and the End of Time-Based Work

The Work Went Synthetic. We Just Kept Billing By the Hour.

Tom Cruise spent two years telling everyone Top Gun: Maverick was real. Actual jets. Actual G-forces. No CGI.

The film had more than 2,400 visual effects shots.

He wasn't lying, exactly. "Practical" had just become a marketing word. The work went synthetic years ago — the label was the last thing to fall.

I've been thinking about this a lot lately, because I'm watching the same pattern play out in professional services. Last week, I sat with a team that delivered a sharp financial analysis to a client. Most of the work — data gathering, normalization, initial modeling — had run through AI before a human ever touched it. What the humans added was the last mile: what's right, what's wrong, what context the model missed, what actually ships to the client. Maybe 20% of the total effort.

It was 100% of the value.

And yet we're still billing like we did when every cell in that spreadsheet required human hands.

The Economics Always Move First

Robert Tercek has documented how animation and CGI ate Hollywood over thirty years, one cost-saving decision at a time. By the time "live action" became a phrase studios used to sell tickets, the transformation was already complete. A studio executive gave him the quiet part out loud: "We don't make art. We make money." Art follows economics.

Professional services are no different. The economics shifted the moment AI could handle the first 80% of knowledge work reliably enough that senior judgment could focus on the last 20%. That's already happened. What hasn't shifted yet is how we talk about it — or how we charge for it.

The billable hour was never really about time. It was a proxy. We were buying judgment and using hours as a stand-in for it because we didn't have a better way to measure expertise. We've been paying rent on the chair and hoping a brain showed up in it.

AI just made the stand-in optional.

What the Railroads Taught Us About Labels

Nobody gets fired the day the railroad arrives. The town just slowly empties out.

When railroads bypassed certain towns in the 1800s, the change wasn't dramatic. Businesses didn't close overnight. People didn't flee. But over five years, ten years, the economic gravity shifted. The towns that thrived were the ones that recognized the shift early and repositioned themselves — not as destinations, but as the judgment layer for where goods should go next.

The towns that died were the ones that kept insisting they were still essential, even as the trains rolled past.

I'm watching firms do the same thing now. They're still structuring engagements around hours and bodies when the actual value they're delivering is taste, pattern recognition, and the ability to tell a client "the model says X, but in your situation, that's wrong because of Y."

That's not a time-based skill. That's a judgment-based one. And judgment doesn't scale linearly with hours.

The Uncomfortable Middle

Here's what makes this hard: the work IS still getting done. Clients are still getting value. The firms delivering AI-augmented analysis aren't producing garbage — they're often producing better work, faster, with fewer errors in the foundational layers.

So what's the problem?

The problem is that the business model and the actual work have quietly decoupled, and most firms are pretending they haven't.

If your team is spending 20% of the time on judgment that represents 100% of the client value, you have three choices:

1. Keep billing the same hours and hope clients don't notice they're paying for work that's now machine-generated

2. Cut your fees proportionally and watch your margins collapse

3. Reframe what you're selling — not hours of work, but access to judgment that's amplified by technology

Most firms are stuck on option one, flirting with option two, and terrified of option three.

I get it. Option three requires admitting that the thing you've been selling for decades — labor, time, effort — was never really the thing clients valued. They valued the outcome. The hours were just the only way we knew how to price it.

What Clients Are Actually Buying

I was advising a client last month on a blockchain implementation. They kept asking how many hours the analysis would take. I finally stopped them.

"You're not paying me to recreate thirty years of watching technology adoption cycles fail and succeed. You're paying me because I've already done that, and you haven't. The analysis might take me four hours. The judgment that makes it worth something took three decades."

The value isn't in the hours. It's in the scar tissue.

This is what Hollywood figured out with CGI. Audiences don't pay for practical effects vs. digital effects. They pay for the experience of watching something that feels real. The method is invisible. The outcome is everything.

Clients don't pay for hours of associate-level analysis. They pay for insight that changes their decision. If AI can handle the scaffolding so your senior people can focus entirely on insight, that's better for everyone — except the business model that relies on marking up junior hours.

The Question Nobody Wants to Answer

So here's the uncomfortable part: if that 20% of judgment is worth 100% of the fee, do we actually care how many hours people work?

I'm serious. If your best analyst delivers a game-changing insight in six hours instead of sixty, should they be paid less? Or more?

We built the whole model of professional work around time. Seats filled, hours logged, Fridays accounted for. Utilization rates. Realization percentages. We created an entire infrastructure to measure presence and call it productivity.

But we were never buying the hours. We were buying the judgment. We've been using time as a lazy proxy for expertise because time was easier to measure than wisdom.

What happens to your org chart the day you admit that?

What This Means Monday Morning

I'm not saying burn down the billable hour tomorrow. I'm saying the firms that survive the next decade will be the ones that start experimenting with alternatives now — before their clients start asking why the invoice doesn't reflect the actual value exchange.

A few things to try:

Ask your team: What percentage of this deliverable was human judgment vs. machine output? Not to police them — to understand where the value actually lives.

Pilot a value-based engagement with one trusted client. Agree on the outcome, not the hours. See what breaks. See what improves.

Track quality, not just time. If AI is handling the foundational work, your people should be producing higher-quality insights, not just faster ones. Are they?

Reframe the pitch. You're not selling analysis. You're selling judgment, amplified by technology your competitors aren't using yet.

The work went synthetic quietly. The label — "professional services," "expert analysis," "strategic advisory" — is the last thing to fall.

The question is whether you'll be the one redefining what those labels mean, or whether you'll be the firm still insisting you're "live action" while everyone else has moved on.

What to do this week: Pick one deliverable your team is working on. Ask them to flag which parts are AI-generated vs. human judgment. Don't change anything yet. Just see the split. You can't redesign the business model until you see how the work actually flows.