happily.ai research All research
Decision-making

Decisions move at the speed of trust

A founder decides at the speed of thought. An employee five layers down decides at the speed of the org. This is a model of what happens in between: how layers, trust, and meeting cadence govern both how fast a decision moves and how good it is when it lands.

Layers
Every rung adds a wait
Trust
Accelerates and protects
Cadence
Sets the heartbeat

A founder decides at the speed of thought. They hold the context, they hold the authority, and the decision happens in the same head that needs to make it. With AI, even less stands in the way. Data, context, and a second opinion all arrive in seconds. The gap between having a question and having an answer is close to zero.

An employee five layers down decides at the speed of the org. They have the idea on a Tuesday. They wait for the next 1:1 to show their manager the homework. They wait for a later meeting to get the decision moved up a rung. Then another meeting, and another rung. Each rung is another wait. The decision is not hard. It is just not theirs to make alone, and the people it has to pass through are busy with their own.

Organizational decisions move at the speed of feedback loops, accelerated by trust and throttled by layers.

Speed of thought, speed of the org

The founder and the frontline employee are not separated by talent or by how good their ideas are. They are separated by distance: the number of people who must touch a decision before it can take effect.

For the founder, that distance is zero. The decision is made where it is felt. For the frontline employee, the same decision has to travel. It has to be explained to a manager, carried to that manager's manager, raised in a forum, defended, and approved. None of those steps is slow on its own. The decision is slow because it must be carried, meeting by meeting, by people whose calendars it does not control.

This is the part that is easy to miss. The frontline employee's decision is not slow because it is hard. It is slow because every step waits for a calendar. And that calendar belongs to someone else.

Three forces

Three things govern how a decision travels from where it is made to where it lands. They are simple to name and they interact in ways that are not obvious.

Layers. A layer is every rung between where a decision is made and where its impact lands. A flat team has one or two. A large organization has five, six, more. Every layer is a person who has to be briefed, who has to agree, and who has to find room on a calendar to do it.

Trust. Trust here means psychological safety: the confidence that raising something, disagreeing, or carrying a decision forward will not be held against you. It does two things at once. First, it lets a decision advance between meetings instead of waiting for one: a trusted person can move something on a message, in a hallway, with a quick yes, rather than booking a forum to make it official. Second, it lets critical context survive the trip. In a low-trust org, customer information from the frontline rarely reaches the decision point intact. It gets softened, trimmed, or dropped at each rung, because saying the hard part is not safe. By the time the decision is made, the people making it are working from a cleaned-up version of reality.

Cadence. Cadence is how often a forum exists where the decision can move at all. If the only place a team aligns is a monthly meeting, then a monthly meeting is a monthly heartbeat. The decision can advance one rung, once a month, and not a day sooner. Weekly cadence beats four times as often. Cadence sets the ceiling on how fast anything can travel, no matter how good the idea.

Simulate your org

One slider, four organizations. Drag AI adoption, or press Play, and watch the same decision speed up for a founder and slow down for everyone else.

An illustrative model, not a measurement. Each line is one organization (its real layers, trust, and meeting cadence) plotted across every level of AI adoption. The time axis is logarithmic. See the formulas.

The model
What this is
An illustrative model, not a measurement. It expresses a structural argument as arithmetic so the shape of the relationship is visible. The numbers are not drawn from a study.
Time to decision
time-to-decision = own_time·(1−AI) + layers · per-layer wait
Per-layer wait
per-layer wait = (cadence ÷ 2)·(1 − 0.85·trust) + AI·cadence
Decision quality
quality = ρ^layers, where ρ = 0.55 + 0.45·trust. Quality decays with every layer the decision passes through; trust slows the decay.
Stall rate
stall = 1 − P(survive every forum cycle)·P(keep the sponsor). The chance the decision never lands at all: it falls off an agenda, or loses its champion.
How to read it
This is a thinking tool. The numbers show shape and direction, not a forecast for any real company.
Time to decision longer is worse
Founder Instant
Startup 12 days
Scale-up 45 days
Enterprise 6.2 months
Context that survives higher is better
Founder 100%
Startup 79%
Scale-up 40%
Enterprise 15%
Chance it stalls out higher is worse
Founder 0%
Startup 41%
Scale-up 72%
Enterprise 98%
Figure 1 One decision, four organizations. As management layers multiply and trust thins, the same decision gets slower, sheds context, and grows far more likely to die before it ever lands. The founder is not better. They are just closer to the work.
Four organizations, one decision, modeled
Organization Layers Trust Time to decision Context that survives Stall rate
Founder 0 85% Instant 100% 0%
Startup (~30) 2 75% 12 days 79% 41%
Scale-up (~300) 4 55% 45 days 40% 72%
Enterprise (~3000) 6 40% 6.2 months 15% 98%

The AI paradox

AI is not helping as much as leaders think. It removes friction at the top. The decider gets data instantly, gets context instantly, gets a draft instantly. For the person at the top of the chart, the time between a question and a usable answer has collapsed.

But the layers in between are now inundated with more decisions to weigh. There is more code to review, more designs to evaluate, more content to proof, more proposals to assess. AI did not reduce the number of decisions an organization has to make. It increased it. And each rung that has to approve something now has more somethings in the queue.

And leaders get a false read. They feel the least decision friction of anyone, because AI cleared their own path first. At the same time, more decisions flow to them, because the layers below are generating more work that needs a sign-off. More decisions arrive, each one easier to make. That feels like productivity, and it feels like importance. It is neither. It is a bottleneck moving toward the person least able to see it as one.

A common mistake

Connecting your AI to the org's meeting notes will not fix this. The bottleneck was never a lack of notes. It is the number of rungs a decision has to climb and the number of days between each one.

Optimize for quality, not speed

Speed is easy to feel. You can sense whether a decision is moving. You can count the days. Quality is not like that. The quality of a decision can only be known through a reliable feedback loop between where the decision is made and where it creates impact. You find out a decision was good months later, from the result, and only if that result travels back at all.

This is why organizations drift toward optimizing speed. Speed is what they can feel, so speed is what they optimize, or worse, its appearance. Meetings get shorter. Agendas get fuller. Decisions get made faster. And nobody can say whether any of them were right, because the loop that would tell them was never built.

CEO EXECUTIVES DIRECTORS MANAGERS FRONTLINE DECISION MADE HERE IMPACT FELT HERE FEEDBACK LOOP
Figure 2 The distance a decision's signal must travel to tell its maker whether it worked. Quality is what survives the round trip.

It was all over before it started

Someone at a company is excited about a product. It is a good idea, and they know it. To move it forward, they need to take it to their manager. But the team aligns once a month, so they wait for the monthly meeting.

The meeting comes. There are too many decisions on the agenda, and theirs never comes up. It rolls to next month. Next month, the manager is traveling, and the meeting is light. Theirs does not make the cut again. Months pass. Eventually the person mentions, almost in passing, that they are leaving the org, so they will not be the one to move it forward.

It was all over before it started. The idea never failed on its merits, it just never got far enough to be judged. It stalled silently across a few skipped agenda slots and one departure.

That is the stall rate. It is the most expensive number in the model because nobody ever sees it, and it is packed with opportunity cost. A slow decision shows up as a slow decision. A stalled one shows up as nothing at all: no meeting, no debate, no record. The org never learns what it lost.

The way forward

The fix is more trust. Without it, every decision must climb every rung before it can be made.

Build trust between layers. Create feedback loops between your decisions and their impact. Turn your information carriers into decision makers, and let the decision be made where the information already lives.

References

  1. Happily Research (2026). Decisions Move at the Speed of Trust. Illustrative model; formulas in the methodology box above.
See where your decisions stall

Happily measures trust, feedback loops, and how work actually moves between layers, so the stalls nobody sees become visible before they cost you a good idea.

Get in touch
Free pilot for qualifying teams