What Work to Delegate to AI: Three Decision Criteria
What you’ll learn
- How reversibility, scope of impact, and verifiability determine whether to delegate work to AI
- Which decisions people should retain, including direction, permissions, published display, and security
- How to apply the three axes to examples and define both delegation scope and verification steps
Evaluate AI Delegation by Reversibility, Impact, and Verifiability
Assess work delegated to AI by whether a mistake can be undone, where the change reaches, and whether the result can be verified. Rewriting a paragraph and sending an email to an external recipient leave different options for dealing with a problem afterward.
These three checks help identify approval needed before delegation and verification that remains afterward. The criteria below come from this site’s operating experience and help define the boundary between delegated work and decisions people retain.
Set Delegation Boundaries with Three Decision Axes
If a problem arises after delegation, I need to know how to reverse the change or verify the result. Treating every task with excessive caution also holds up the work. I organized that decision around reversibility, scope of impact, and ease of verification.
These three axes are a practical decision aid informed by general risk-management thinking and this site’s operating experience. They are not a standardized official framework.
Reversibility (Can It Be Undone?)
Reversibility means being able to restore the state before a change. Check what method would be available after a mistake.
Reversible → easier to delegate
Button colors, paragraph text, and added sections have a clear way to undo the change. Git, a tool that records file changes and can restore an earlier version, lets you compare and revert them.
Difficult to reverse → verify first
Rewriting database contents, changing a published URL, or sending an email to an external recipient can be difficult to undo afterward. A person reviews what will be done and its impact before delegating execution to AI.
Scope of Impact (Where Does the Change Reach?)
Scope of impact means which pages or functions a change reaches. It determines whether checking the edited location is enough or whether other areas also need review.
Localized change → easier to delegate
A change such as “increase the font size of this heading on this page” has a single target. You can check that location to confirm the intended result.
Wide-reaching change → proceed with care
Changes to site-wide navigation, authentication, or styles shared across all pages can affect many locations. Before execution, define both what will change and where the results will be checked.
Also check whether the work involves confidential information, permissions, legal or contractual constraints, or sending data outside the system. The three axes alone cannot establish that such work is easy to delegate.
Ease of Verification (How Can the Result Be Confirmed?)
Ease of verification means whether a person can establish that the result matches the intent. Consider whether the verification method is understood and whether it requires specialized knowledge.
Verifiable by appearance or behavior → easier to delegate
Design changes, text edits, and link placement can be checked in a browser by viewing their appearance or behavior. Some visible results can be assessed without specialized implementation knowledge.
Requires specialized knowledge to verify → human must confirm
Security settings, improvements to processing speed, and connections to external services are difficult to assess merely because a page appears. Understand the verification method before delegating, or work with someone who can verify the result.
Decisions Humans Should Keep
Even when a task is easy to delegate, not every decision should be handed to AI. AI can suggest implementations and options, but humans still decide what to build, which risks are acceptable, and whether the result is ready to publish.
Direction-Setting
People decide whether to add a feature, which page structure to use, and which design to adopt by considering the site’s purpose. Receiving an implementation proposal from AI does not automatically settle that direction.
External Services and Permissions
AI can propose steps for connecting to services such as Vercel and GitHub, or to an external API—a way for a program to request work from another service. A person checks which repository (where code, configuration, and their change history are kept), account, environment variable (a setting passed to a program, such as a connection destination), and permission will be used. An incorrect destination or permission can leave the publishing scope different from what was intended.
Verifying the Published Display
Even if AI changes the code or configuration correctly, humans still need to check whether the published page behaves as intended. Layout differences between desktop and mobile, wrong navigation targets, and wording that changes the intent can be missed if only the file diff is reviewed.
Security-Related Judgment
API keys, passwords, information that should not be public, and settings that send data outside the system should not be judged only from AI’s proposal. AI can investigate and list options, but final confirmation remains a human responsibility.
Tasks That Are Easier to Delegate
The easiest tasks to give AI are those with a clear target, a clear way to undo the change, and a clear verification method.
- Concrete implementation: change a button color, add a link to a paragraph, add a section to a specific page
- Organizing options: compare multiple implementation approaches or configuration options
- Investigating errors: use an error message to identify possible causes and verification steps
Even for these tasks, humans still decide what to execute, which option to adopt, and how the result will be checked.
Compare a UI Change and an Authentication Change Across the Delegation Criteria
“Change the button color on the top page from blue to white”
- Reversibility: high (can be reverted with Git)
- Scope of impact: narrow (one button on one page)
- Ease of verification: high (visible in browser) → Easy to delegate to AI
“Modify the user authentication system”
- Reversibility: low (problems after the change can be complex to address)
- Scope of impact: wide (affects login, registration, and session management)
- Ease of verification: low (security issues cannot be confirmed from the display) → Proceed carefully with a solid understanding of what is being changed
Use the Three-Axis Table to Assess Delegation
Use the table to identify which side of each axis applies.
| Axis | Easier to Delegate | Proceed with Care |
|---|---|---|
| Reversibility | Can be undone | Difficult to undo |
| Scope of Impact | Localized | Wide-reaching |
| Ease of Verification | Visible in browser | Requires specialized knowledge |
If any axis falls under “proceed with care,” decide how to address it before moving forward. Legal, security, permission, and publication decisions still require the appropriate specialist or owner even when the three-axis assessment looks manageable.
Summary: If Any of the Three Axes Requires Care, Design Human Review First
Delegation scope depends on more than how difficult a task is. If there is uncertainty about reversing a change, its reach, or verifying its result, define the necessary review, permissions, and stopping method first.
People retain the decisions about what to build, how much to delegate, and what counts as acceptance. With those decisions in place, work that is easy to reverse and verify can be delegated to AI.
This article is a general information summary and is not legal advice. Confirm practical decisions with a qualified specialist.