The old pyramid, in one chart.
A mid-sized consulting firm in 2020 ran roughly this staffing structure on a typical engagement.
One partner at 10 percent of total hours, billed at $650 an hour. One manager at 20 percent of hours, $425. Two consultants at 40 percent, $275. Two analysts at 30 percent, $175. The engagement leans on juniors for most of the execution, and the seniors review and sell.
This structure was not an accident. It solved a real problem: how to scale expertise across many engagements at once. It also had a known flaw: most of the work was being done by the least experienced people, and the quality of the output depended heavily on the review layers above them catching the problems.
What Claude absorbed, what it did not.
The clearest way to describe what changed is by listing what Claude is good at and what it is not.
Claude absorbed: first-draft research writeups, literature synthesis across 50 papers or 200 forum posts, initial report structures, boilerplate security findings, competitor analysis first passes, pricing comparisons, schema definitions, test harness scaffolds, first-cut threat models, initial positioning briefs, spreadsheet-style data transforms, long structured writing. These were the bulk of the junior workload in 2020.
Claude did not absorb: judgment calls under uncertainty, prioritization with real stakes, relationship work, subtle voice and tone that sells, contested sales conversations, final review, deciding what not to ship, reading the room, exploit validation in security, production incident response, client-side politics. These were the senior workload in 2020 and they remain the senior workload now.
The consequence is that the junior-heavy pyramid stops making economic sense. The juniors' former work is now a Claude task. The seniors are still required. The ratio inverts.
The inverted team.
Our team structure for an engagement is inverted from the old pyramid.
One senior operator leads the engagement at about 60 percent of billable hours. This person does sales, scope, direction, review, final writing, client conversation, and decisions. They are never replaceable by Claude.
One reviewer or peer second-pair at about 25 percent of hours for engagements large enough to need a second set of eyes. This is another senior, usually from a different discipline (a security lead reviewing a product engagement's threat model, for example). Serves as internal challenge and cross-pollination.
Claude at whatever time is needed, not as a "headcount" slot but as the tool both people use. Claude produces drafts, explores corners, searches, synthesizes. It is used everywhere but owns nothing.
One coordinator or operator-assistant at about 15 percent of hours for scheduling, logistics, form work, and the kind of coordination that is lightweight but constant. Not junior in the old sense; more like a chief-of-staff role.
That is the whole team on a $40k-to-$120k engagement. Two senior humans, one coordinator, Claude as the production tool. It is not what a consulting firm in 2020 would recognize as a staffed engagement, but it delivers output that matches or exceeds a five-person team from five years ago.
What a senior operator actually does all day.
The day has shifted, not just the headcount. A senior used to spend most of their time reviewing work. Now they spend most of their time doing work, with Claude as a drafting partner.
A rough breakdown from a tracked week:
Direct client work: 40 percent. Calls, debriefs, interviews, sales pitches, delivery presentations. This hasn't changed.
Production work with Claude: 35 percent. Drafting reports, coding, writing, synthesizing. Most of this used to be delegated; now the senior drives it, with Claude producing first passes that the senior shapes.
Review and critique: 10 percent. Reviewing own drafts, running self-critique passes, reviewing a teammate's work. Down from maybe 25 percent in the old model.
Sales and growth: 10 percent. Proposals, outreach, content. Up from about 5 percent because there is no junior to absorb the sales-support work.
Operations: 5 percent. Invoicing, scheduling, admin that the coordinator cannot absorb.
The headline: seniors spend more time producing than they used to, and less time reviewing junior work. The leverage is in the tool, not the headcount.
The failure mode: juniorless juniors.
There is a version of this model that looks inverted but performs like a pyramid: a single senior plus a team of mid-career people who were promoted because Claude made them look senior.
The failure mode is that mid-career people without true senior judgment plus Claude produces output that looks polished but makes bad calls. Claude writes well. It cannot decide. The final judgment that a senior brings is the last human layer between the tool and the client.
We have met firms that staffed this way. Their output was fluent, plausible, and shallow. They did not catch the wrong call. They did not know what they didn't know. Claude's confidence was a liability, not an asset, because there was no layer above it that would slow down and disagree.
The inverted team only works if the inverted position is actually senior. Promoting someone into a senior role because Claude makes their output look senior is how you fail quietly, engagement after engagement, until a big one blows up.
What this does to margin.
The economic effect is direct.
A $60k engagement under the old pyramid might cost $32k in loaded salary plus overhead for the week team. Margin roughly 45 percent before firm costs.
The same $60k engagement under the inverted team costs about $18k in loaded salary for two seniors plus a coordinator for three weeks, plus $200 in Claude API costs. Margin roughly 68 percent before firm costs.
The margin improvement is not a pricing trick. It is a structural consequence of removing the junior layer whose work Claude can now do. The senior hours cost more per hour, but there are fewer of them, and the total is lower.
The client, meanwhile, is paying for the same deliverable, getting it a third faster, and dealing with two senior humans instead of a rotating cast of five. This is the value proposition that closes the sales cycle: higher quality, lower total cost, faster delivery, better relationships. The inverted team is not a consultant trick. It is what the tool makes possible and what the market is about to expect.
One essay a week. No filler.
Four pillars, one email every Tuesday. If we have nothing worth sending, we skip the week.