Good AI Task

AI compatibility

Diagnosing a foundation crack is hands-on work no AI can safely do remotely.

Human required

A human should do this one.

Average across 1 submission.

8
avg / 100

The honest read

Evaluating a foundation crack requires physical inspection, tactile assessment, and expert structural judgment that no AI agent can perform remotely. The stakes are high — a missed structural crack can lead to catastrophic failure or costly delayed remediation — and the data an agent would need simply cannot be captured through text or even photos alone. This is a job for a licensed structural engineer or foundation specialist.

Aggregated across 1 submission.

The five dimensions

Repeatability

Low

Every crack is unique — width, length, orientation, location, moisture presence, and surrounding soil conditions all vary. There is no repeatable structure an agent can reliably apply without physical access to the specific crack.

Ambiguity Tolerance

Low

The success criterion — 'cosmetic vs. structural' — sounds binary but is deeply ambiguous in practice. Distinguishing a harmless shrinkage crack from an active structural failure requires expert interpretation of subtle physical cues that resist crisp definition.

Data & Tool Availability

Low

An agent has no access to the physical crack: no ability to measure width with a gauge, probe depth, assess moisture, or observe the surrounding foundation context. Even if the user uploads photos, image quality and angle are rarely sufficient for a reliable diagnosis.

Error Cost

High

A false 'cosmetic' verdict on a structural crack could lead the homeowner to ignore a serious problem, resulting in foundation failure, flooding, or collapse — outcomes that are expensive, dangerous, and irreversible. The downside risk is severe.

Human Judgment Required

High

Structural assessment requires licensed engineering judgment, physical inspection, and accountability that AI cannot provide. Even experienced professionals sometimes disagree on borderline cases, underscoring how much expert intuition and on-site context matter here.

What an agent would need

  • High-resolution, multi-angle photos of the crack with a ruler or reference object for scale
  • Information about crack type (horizontal, vertical, diagonal, stair-step), age, and any recent changes
  • Knowledge of local soil conditions, foundation type, and home age
  • Access to structural engineering databases or building code references for the region
  • A clear disclaimer framework, since any AI output here carries serious liability risk

Best-matched agent type

Research Agent

The kind of agent this work would call for if it were a fit. For this task, it isn't.

Run your own fit check

Get a calibrated read on your specific task in under a minute.

Check a task