Is Adoption Latency the same thing as the Coordination Tax scaled down for a specific decision?

No — they measure different things even though both scale with headcount. Coordination Tax measures the ongoing cost of keeping an existing structure functioning. Adoption Latency measures the delay before that structure changes at all — specifically, before an unconventional or unproven capability is tried. A team can have relatively low Coordination Tax and still have high Adoption Latency if trying something genuinely new requires broad organisational buy-in that routine operation does not.

Does this mean larger teams should never attempt to adopt unconventional AI capability early?

No — it means a larger team attempting to occupy the same window needs to deliberately reduce how many people must agree before an unconventional attempt is made, rather than assume its larger resource base compensates for slower consensus. This usually means carving out a genuinely small, empowered sub-team with real latitude to try something without first securing broad organisational buy-in.

How would a business know if it has genuinely low Adoption Latency, rather than just assuming it because the team is small?

The test is not team size directly — it is how quickly an unconventional, unproven idea moves from being proposed to being actually tried. A small team that still requires multiple rounds of justification before attempting something unconventional has high Adoption Latency despite its size. The genuine signal is whether the team can move from idea to attempt without needing to convince anyone who was not already going to be doing the work.

Why does the compensation argument specifically require an ad hoc structure rather than simply a higher standard bonus?

Because a standard bonus, even a generous one, is still calibrated against an expected range for a team or role of that size — it treats the outsized result as an unusually good version of what was expected, rather than a category of result the standard band was never designed to capture, because it depended on a condition — low consensus requirement — the standard structure does not usually create or reward directly.

Is there a point where a team becomes too small to benefit from low Adoption Latency, because it lacks the capacity to execute on what it discovers quickly?

Yes — the advantage this memo describes is specifically about speed to first attempt, not about the capacity to scale a validated approach afterward. A team small enough to try something unconventional quickly but too small to build on a validated result once it works has captured the discovery advantage without being able to capture the full return from it.

Why Small Teams Move First

Adoption Latency is the delay between an AI capability becoming available and a team actually incorporating it into working practice — a delay that scales with team size, because more people require more consensus before an unconventional or unproven tool is tried at all. It is distinct from Coordination Tax — the overhead cost of human-to-human alignment: the meetings, approvals, status updates, and manual handoffs required to keep a traditionally structured business functioning. Coordination Tax is a cost of steady state. Adoption Latency is a cost of change — the specific delay a team pays every time something new appears and has to be tried, and it is paid disproportionately by larger teams, not because they lack talent or resources, but because they have more people who need to agree before the trying starts.

The observation

A small team can ship more with a new AI capability than a larger, better-resourced one, in the specific window right after that capability becomes available — before its use has been normalised, before best practice exists, before anyone else has proven it works for the specific problem at hand. The advantage is not raw output capacity. A larger team almost always has more of that. The advantage is speed to first attempt. Two people can decide over lunch to try an unconventional approach with a tool that shipped last week. Twenty people need someone to champion the idea, a case for why it’s worth the risk, and usually a fallback plan in case it doesn’t work — and by the time all of that is assembled, the specific window in which trying it first mattered has often already closed.

This is not the same claim Coordination Tax already makes. Coordination Tax explains why a larger team’s ongoing operating cost is higher — the meetings, approvals, and status updates required to keep an existing structure functioning. Adoption Latency is about something narrower and more specific: not the cost of running the team as it already operates, but the delay before the team changes how it operates at all. A team with genuinely low Coordination Tax in its steady-state work can still have high Adoption Latency if trying something unconventional requires broad buy-in — the two are correlated, because both scale with headcount, but they are not the same measurement.

Why the window matters more than the eventual capability

Every AI capability eventually becomes conventional. Best practices get written, case studies circulate, and the unconventional bet of one month becomes the standard approach of the next. Adoption Latency is not about who eventually adopts a capability — nearly everyone does, eventually. It is about who tries it first, while it is still unproven, while using it well is itself a competitive advantage rather than table stakes. That window is genuinely narrow, and a team’s ability to occupy it is determined less by its resources than by how many people have to agree before the attempt is made.

A small team operating with real freedom to explore — a stack they control, latitude to pursue an unconventional approach without first justifying it up a chain of approval, and enough pressure on themselves to actually demonstrate whether their instinct is right — can compress that window to nearly nothing. The team decides, tries, and either the approach works or it doesn’t, on a timescale a larger organisation cannot match regardless of how much talent or budget it has, because the constraint was never talent or budget. It was how many people needed to be convinced before anyone was allowed to find out.

What this implies for compensation

A small team or an individual operator who demonstrates outsized returns by occupying this window — genuinely trying something unconventional first, while the capability was still unproven, and being right — has produced a result a larger organisation’s structure made structurally difficult to produce at all, independent of how much that organisation was willing to spend trying. This is close in spirit to Workforce Arbitrage, which measures the cost delta a business captures by replacing human execution with an agentic stack. The result here is a different kind of delta: not the cost saved by fewer people doing the work, but the return created because fewer people were needed to agree before the work was attempted at all.

That result is difficult to compensate through a standard structure, because standard compensation structures are themselves built around the same headcount-scaled consensus this memo describes — a role, a level, a band, all calibrated against what a typical team of that size typically produces. A small team that produced a result a typical team of that size could not have produced, specifically because it was small enough to try something a larger team would not yet have agreed to try, is not well described by the typical band for a team of its size. An ad hoc compensation structure tied to the specific, demonstrated return — rather than to the headcount or seniority that would normally set the number — is the honest response to a result the standard structure was never built to price.

The Operator’s Verdict

Team size is not primarily a cost variable in the age of unconventional AI adoption. It is a consensus variable — and consensus, not capability or capacity, is what determines who tries something first. A business that wants to occupy the adoption window repeatedly should not simply build small teams. It should build teams small enough that the number of people who must agree before an unconventional attempt is made stays low, give them a stack they actually control, and be prepared to compensate the outsized outcome on its own terms rather than the terms a standard structure would have assigned it.

Technology changes what becomes possible to try. Team size determines how long it takes before anyone actually tries it.

KEY TAKEAWAY

What is Adoption Latency and why do small teams have a structural speed advantage in AI tool adoption?

Adoption Latency is the delay between an AI capability becoming available and a team actually incorporating it into working practice — a delay that scales with team size because more people require more consensus before an unconventional or unproven tool is tried at all. It is distinct from Coordination Tax, which measures the standing cost of alignment a team already requires to keep functioning in its current state. A small team’s advantage is not raw output capacity — it is speed to first attempt, because fewer people need to agree before an unconventional bet is tried. Every AI capability eventually becomes conventional; the competitive advantage exists specifically in the narrow window before best practice is established. A small team or individual that demonstrates outsized returns by occupying this window has produced a result standard, headcount-calibrated compensation structures are not built to price, making an ad hoc compensation structure tied to the demonstrated return the honest response. Source: Arco Venture Studio.