The Judgment Application Framework: How Judgment Works in a Business

The Judgment Application Framework explains how judgment works in a business: what judgment each decision and function needs, who has it, and how to get it where it's needed.

It rests on one premise: judgment is situational. The same leader can read one decision well and the next one badly, and nothing about the leader changed in between. The decision changed. So judgment in a business is less a trait of its people than a fit, decision by decision, between the judgment needed and the judgment that can reach it.

The framework has two halves. The Judgment Model reads the fit through three elements: Judgment Need, Judgment Capacity, and Judgment Alignment. Judgment Systems are the build that improves performance, by getting judgment where it's needed.

Judgment Need

Judgment Need is how much judgment a decision calls for, and what kind. Approving a routine vendor invoice calls for almost none. Repricing a product line that a third of your customers depend on calls for a lot, and for several kinds at once.

Nine dimensions describe the difference. Each one is a question you can put to a decision:

• Complex. Do the parts interact?
• Ambiguous. Is it clear what kind of problem this is?
• Important. How hard does it land?
• Uncertain. How wide is the range of outcomes?
• Expensive. How much does it cost, and can it be undone?
• Novel. Has anyone here seen this before?
• Contested. Who else holds a legitimate claim on the outcome?
• Technical. Does it turn on expertise the decider doesn't have?
• Long-dated. When will you know whether it was right?

Judgment Need gets read one decision at a time. The dimensions have no order and no stages, and most decisions carry only one or two. A decision that carries five asks for far more than one that carries none, even when the two look alike on a meeting agenda.

Functions have Judgment Need too. A function's need is the pattern in the decisions it makes: sales lives with uncertain and contested calls, finance with technical and expensive ones. A function has a signature of judgment that it needs.

Each dimension changes what a good decision requires. The Judgment Need article looks at each dimension in depth.

Judgment Capacity

Judgment Capacity is the ability to read a situation and make a sound call on it. It has two parts. Judgment Range is how many domains someone has calibrated judgment in. These are areas where they've made enough calls, and watched enough of them play out, to know what a good one looks like. Situational Fluency is how well someone reads the context a decision sits in, so their judgment can work on it. A long-tenured CFO might read every financial decision fluently and have little range outside finance. An owner might have range across the whole business because they've seen the decisions before…but still only read each area at the surface and therefore miss important context.

Capacity sits in multiple places. A leadership team holds it. So does an organization, in its people and in whatever it has written down. An engineered AI agent holds some too. AI models bring a generic kind: adequate on common situations, thin on specific ones…and delivered with the same polish either way. That last part is where AI Surface Expertise comes from.

Some capacity sits outside the company and can still reach the decision. A board, a peer group, an industry network, an advisor. That's borrowed capacity, and on the hardest decisions a leader faces, it's often a large share of what's available.

One kind of capacity matters for everything that comes after. Every company has judgment that belongs to it alone: how it prices a project like this particular one, which customers it walks away from, what good enough means on a delivery, which problems go to the owner. Most of it lives in people's heads (what’s usually called “tacit knowledge”), and the owner's head holds the most. House Judgment is that company-specific judgment, stated well enough that someone or something other than its holder can run on it. Plenty of companies have the judgment. Few have stated it and claimed it as their House Judgment.

How range and fluency get built, in an owner and in rising leaders, is the subject of the Judgment Capacity article. If the judgment in question sits inside an AI agent, the Work Agents and Judgment Agents article shows how the Judgment Model applies to AI agents. And getting judgment out of one head, into a form others can use, is the work of the House Judgment article and the AI Logic Engineering article.

Judgment Alignment

Judgment Alignment is whether the judgment that reaches a decision matches what the decision calls for. A leader can be well aligned with a strategic decision one day and badly aligned with one the next day, with nothing about the leader different between them.

Alignment is a reading, the way a gauge is. Improving it is a separate job with two paths: growing people's judgment, or building ways for decisions to reach the people who have it. That work belongs to (growing) Judgment Capacity and (building) Judgment Systems.

Alignment also moves over time. Judgment Drift is the slow version of misalignment: the decisions change, and the judgment applied to them doesn't. Nobody makes one bad call to cause it. As a company grows, it shows up as decisions that arrive more complex, more consequential, and less familiar than the ones that built the business. That's Second Stage Judgment Drift, and the Second Stage Judgment Drift article covers it, along with the quieter form it takes in long-established family businesses. Owners can often feel it before they have words for it.

When the mismatch persists, something builds up. Each decision made at a poor fit looks reasonable on its own, and the cost collects across quarters. That's Judgment Debt, as described in the Judgment Debt article.

When alignment fails, it fails in one of two ways.

Shortfall. Nobody who can be brought into the decision has the judgment it needs. Not in the building, and not among the people the building can call.

Here's an example of judgment shortfall. I worked with a client who grew beyond their leadership team's judgment capabilities. That resulted in several bad hires on the leadership team, which then resulted in those bad hires taking the company in unproductive directions in operations, finances, and talent management. Once I was able to describe what the bad hires should have been doing, the owners of the business were able to remove the leaders and reset the path with better strategy in those areas. The business righted itself…once the right judgment was added to the right decisions.

Misplacement. Someone has the judgment the decision needs, in the company or outside it, but they aren't part of the decision.

Sticking with the example we just looked at, one of the changes that was made by a bad hire was to remove the owners from the hiring process. What ensued was a series of weak hires, and especially ones that weren't a good cultural fit, which made the impact even broader. Once the owners returned to an active role in hiring, the quality of the hires improved back to its historical strength, and the culture of the company was restored.

Misplacement has two fixes. Grow judgment where the decision already sits, which is how rising leaders rise. Or build a way for the decision to reach the people who already have the judgment, which is the work of Judgment Systems. Most companies need both. Misplacement is why this framework covers the whole organization and not only the leadership team.

Judgment Systems

Judgment Systems are the processes or tools (more and more, using AI) that put judgment where decisions are. In the hiring example, the fix was a series of processes that ensured the right assessments got done, and brought the owners into some of them.

Other Judgment Systems are more involved, and AI makes a new kind of Judgment System possible: one that encodes a company's judgment into tools people and AI agents can work from. The most advanced of these are AI agents engineered to work through judgment calls, with the logic to show their work - like the 45,000-word system I built to model AI's impact on companies’ markets.

Every AI-based Judgment System rests on the reasoning inside its AI tools, and AI Logic Engineering is the practice that builds that reasoning. The Judgment Systems article describes what the systems look like tier by tier.

Judgment Systems don't create judgment. They help apply it - putting the judgment that exists where the decision is. That's why the framework is called the Judgment Application Framework.

Frequently asked questions

What is the Judgment Application Framework?

The Judgment Application Framework explains how judgment works in a business: what judgment each decision and function needs, who has it, and how to get it where it's needed. It rests on the premise that judgment is situational, so the same leader, team, or AI agent can fit one decision well and the next poorly. The framework has two halves. The Judgment Model reads the fit through three elements: Judgment Need, Judgment Capacity, and Judgment Alignment. Judgment Systems are what a company builds to get judgment to the decisions that need it.

What is Judgment Need?

Judgment Need is how much judgment a decision calls for, and what kind. Nine dimensions describe it: complex, ambiguous, important, uncertain, expensive, novel, contested, technical, and long-dated. It belongs to the decision and gets read one decision at a time, with no order or stages. Most decisions carry one or two dimensions, and the ones that carry several are where judgment gets tested. A function's Judgment Need is the pattern in the decisions it makes, its signature of judgment. Judgment Need covers each dimension in depth.

What is Judgment Capacity?

Judgment Capacity is the ability to read a situation and make a sound call on it. It has two parts. Judgment Range is how many domains someone has calibrated judgment in: areas where they've made enough calls, and watched enough of them play out, to know what a good one looks like. Situational Fluency is how well someone reads the context a decision sits in, so their judgment can work on it. A person, a team, an organization, or an engineered AI agent can hold Judgment Capacity. So can people outside the company, such as a board, a peer group, or an advisor, as long as they can be brought into the decision. Judgment Capacity covers how range and fluency get built.

What is Judgment Alignment?

Judgment Alignment is whether the judgment that reaches a decision matches what the decision calls for. It works like a gauge: it reads the fit, decision by decision, and it can change from one day to the next for the same leader. Improving it takes one of two kinds of work. Growing people's judgment is the work of Judgment Capacity, and building ways for decisions to reach the people who have the judgment is the work of Judgment Systems. When alignment fails, it fails through shortfall or misplacement. When it slips gradually over time, that's Judgment Drift.

What is the difference between shortfall and misplacement?

Both are ways a decision ends up without the judgment it needs. In a shortfall, nobody who can be brought into the decision has that judgment. In a misplacement, someone has it, inside the company or outside it, and they aren't part of the decision. Shortfall calls for building judgment or borrowing it. Misplacement calls for growing judgment where the decision sits, or building a way for the decision to reach the people who have it, which is the job of Judgment Systems.

What is a Judgment System?

A Judgment System is a process or tool that puts judgment where decisions are made. Some are simple, like a hiring process that brings the right people into the right assessments. Others use AI to encode a company's judgment into tools that people and AI agents can work from, up to AI agents engineered to work through judgment calls and show their reasoning. Every AI-based Judgment System depends on the reasoning built into its tools, which is the work of AI Logic Engineering. Judgment Systems covers what they look like tier by tier.

What is House Judgment?

House Judgment is a company's own judgment, stated well enough that someone or something other than its holder can run on it. It covers the calls that are particular to one business: how it prices, which customers it turns away, what good enough means, which problems go to the owner. It has two conditions. It's specific to this company, and it's explicit. Most companies meet the first and not the second, so their most distinctive judgment sits in one head, where it can't be taught, checked, delegated, or built into an AI agent.

What is Judgment Drift?

Judgment Drift is the gradual mismatch that builds as the decisions in front of a leader or team change and the judgment applied to them doesn't. No single bad call causes it. Second Stage Judgment Drift is the form it takes as a company grows out of start-up: the owner’s or leadership team’s judgment stays calibrated to the decisions that built the business, while the decisions get more complex, more consequential, and less familiar. It's usually the first part of the framework an owner feels.

Does the framework apply to AI agents?

Yes. An AI agent carries judgment of a generic kind: adequate on common situations, thin on the specific ones a business depends on. The same reading applies. What judgment does the agent's task call for, what judgment does the agent carry, and how well do the two fit? Work Agents and Judgment Agents applies the Judgment Model to agents, and the five tests let a leader check whether an agent is engineered to judge well.

Where to start

If you're reading this because a decision recently felt different and you couldn't say why, start with Second Stage Judgment Drift. If you're looking at where judgment sits in your company and where it doesn't reach, that's a conversation I have with owners most weeks. Let's talk.