Interactive Assessment

Should you automate this role?

Walk through the Judgment Architecture framework in 5 minutes. Get a structured recommendation based on judgment type, risk exposure, and governance readiness.

Step 1 of 6

Define the role

Start with the role or function you're evaluating for AI automation. Be specific — "customer support agent" not "operations."

Role or function name

Department / Business Unit

How many people currently perform this role?

Primary responsibilities (brief)

Step 2 of 6

Decompose the judgment layers

Estimate what percentage of this role's judgment falls into each layer. The three sliders must total 100%.

Visible judgment 40%

Data in systems, pattern recognition, structured decisions with clear feedback loops

NoneDominant

Contextual judgment 35%

Interpretation, calibration, reading the room, understanding what data doesn't say

NoneDominant

Invisible judgment 25%

Relationships, institutional memory, informal signal networks, things you don't know exist until they're gone

NoneDominant

Step 3 of 6

The Volume Trap check

Volume and consequence are different distributions. AI might handle most of the volume while none of the consequences.

What % of task volume could AI handle? 70%

Routine, repetitive tasks that follow clear rules

What % of consequence weight sits in that automatable volume? 30%

If AI gets these tasks wrong, how much of the real damage comes from this group?

Step 4 of 6

The Bottleneck Principle

One load-bearing invisible component makes the whole role unsafe to fully automate. Identify if any exist.

No bottleneck identified

All components are either visible or contextual. No single invisible element is load-bearing.

Possible bottleneck

There may be invisible components (relationships, institutional memory) but they're not clearly critical.

Clear bottleneck exists

At least one invisible judgment component is load-bearing. Removing it would cause downstream failures.

Describe the bottleneck (if any)

Step 5 of 6

Three governance gates

Rate your organization's readiness on each gate. Be honest — skipping any gate has produced predictable failures.

Gate 1: Values alignment 5/10

Have the values governing this role's decisions been articulated in a way AI can operationalize?

Not at allFully documented

Gate 2: Liability exposure 5/10

Has worst-case damage been mapped, priced, and assigned ownership?

No analysisFully mapped

Gate 3: Escalation path 5/10

Is there a clear human escalation path for cases AI can't handle?

No path existsFully designed

Step 6 of 6

The confidence problem

AI doesn't just get things wrong — it gets things wrong with certainty. How much does that matter here?

Low stakes if AI is confidently wrong

Errors are easily caught, reversible, and low-cost. A wrong answer doesn't cascade.

Moderate stakes

Some errors are costly or hard to reverse. Confident wrong answers could mislead decision-makers.

High stakes — false confidence is dangerous

Legal, health, safety, or financial decisions where a confidently wrong answer could cause irreversible harm.

Judgment Architecture Assessment

Judgment layer decomposition

The Volume Trap analysis

Governance gate readiness

Recommended approach

Want expert guidance on this assessment?

Get a deeper analysis of your judgment architecture and a custom automation roadmap.

Start a Conversation →