How we estimate operational drag.
We want you to be able to check our work, so here is exactly how the calculator arrives at its numbers. The dollar and hour ranges come straight from your inputs, run against published industry research — nothing fudged, nothing guessed. The personalized task analysis is written by Claude (Anthropic) from your own business profile, and we tell you so plainly rather than passing it off as ours. The math is sourced; the analysis is yours to keep.
How your inputs become drag hours.
To get to a weekly drag figure, we multiply five quantities together: headcount × repetitive task hours × tool tax × retrieval friction × resilience factor. The first two come from you. Every multiplier after that is calibrated against published research, so you can trace where the number came from rather than taking our word for it.
-
Tool tax ~9% per tool Productivity lost per platform boundary, top-coded at 11+ tools.
This comes from Asana's Anatomy of Work Index. We read it as a ladder: one platform carries no tax, three to five adds a 15% multiplier, six to ten 35%, eleven or more 55%. We kept the reading on the conservative side on purpose.
Asana · Anatomy of Work Index · asana.com/resources/anatomy-of-work -
Retrieval friction ~1.8hrs / day Hours a day knowledge workers spend searching for information.
This comes from McKinsey's The Social Economy, and it anchors the status-of-X ladder — seconds (no friction), minutes (10%), hours (25%), nobody knows (40%). When friction is high, you end up paying for the same hours twice: once to do the work, then again to go find it.
McKinsey Global Institute · The Social Economy · 2012 -
Resilience factor
When the whole thing rides on one person, you pay for that fragility even on the days they are in.
If work grinds to a halt the moment one person is out, it usually means most of your operating knowledge lives in a single head. That costs you hours every week in re-explaining things and quietly double-checking them — and you pay it whether or not that person ever actually leaves, because it is baked into how the team moves day to day.
Calibrated from project-management literature on bus-factor and workflow continuity.
How much of that drag you can realistically take back.
Once we have your weekly drag, we apply a reclaimable-share band of 45% to 65% to get to the reclaim figure. We are deliberate about this on purpose: it is the part of the drag that structure and careful AI delegation can genuinely take back — not all of it, not even most of it, but a band we can stand behind, grounded in two sources you can look up yourself.
-
Floor · 45% ~30% by 2030 Of work hours automatable by 2030, weighted up for SMB-typical repetition.
McKinsey's 2023 generative-AI productivity report. The 30% headline averages across all knowledge roles; the calculator's 45% floor weights up because SMB operations run more repetitive than the cross-economy mean.
McKinsey & Company · Generative AI and the Future of Work in America · 2023 -
Ceiling · 65% 58% meta-work Of work is “work about work” — coordination, status checks, search.
Asana's State of Work surveys consistently put over half of work time in meta-work. It is the most automate-friendly category, so it is the soft ceiling on reclaim — but some coordination is irreducible.
Asana · State of Work · 2023
Where the per-vertical loaded rates come from.
The default hourly rate for each vertical is built from Bureau of Labor Statistics Occupational Employment and Wage Statistics (May 2025), blended across the two-to-three SOC codes that map to the typical SMB role mix in that vertical, with a ~40% employer-side burden multiplier for benefits, payroll tax, and overhead.
These defaults are starting points, not claims about your specific business. The calculator leaves the rate as a field you can edit, and your own number always wins. The defaults are there to give you a credible place to begin if you have not run a recent loaded-cost calculation — not to tell you what your hours actually cost.
What "loaded" means
Loaded hourly rate = base wage × (1 + employer burden). Employer burden covers benefits, employer-side payroll taxes (FICA, SUTA, FUTA), workers' compensation insurance, and a small allocation of overhead. We use 40% as a conservative multiplier; well-benefited operations can run 50%+, lean shops 30%.
Specific blends per vertical are documented inline in the calculator's source code (calc.js) by SOC code for full transparency.
The boundaries of an estimate.
-
An estimate
This is an estimate, not an audit. A real Structural Audit takes one to two weeks of access to your operation and produces a written diagnosis that is yours to keep, in plain English. The calculator gives you a defensible range in 90 seconds. The two go together; one does not stand in for the other.
-
Industry median
Vertical-level medians hide a lot of variance. A 25-person solar installer in Phoenix and a 25-person solar installer in Massachusetts pay measurably different loaded rates. The calculator's vertical defaults are a useful place to start — but your own loaded rate, dropped into the form, is the number that tells you the truth.
-
AI-generated content
The personalized task analysis shown in the calculator's delegation map is written by Claude (Anthropic) from your inputs. We say so inline on every result, because we would rather you know where it came from than have us present AI-generated work as if it were our own.
-
No guarantees
We do not promise you a guaranteed reclaim. What we stand behind is a defensible range, built from sources we have shown you. The real outcome for any one engagement depends on how the structure gets built, how your team takes to it, and how carefully the agents are managed once they are running — see our Trust Framework for what that managing looks like in practice. We would rather be honest about that up front than sell you a number we can't keep.
Now that you can see how it works, run it on your own numbers.