THE SHORT ANSWER

An agent can fit variable, multi-step work that requires choosing among tools as evidence changes. Avoid an agent when rules are stable, exact predictability is required, reliable tools or data are missing, latency and cost must be minimal, or a mistake cannot be contained. Start with the least autonomous design that meets the outcome.

Look for justified variability

Signals for and against an agent
QuestionAgent may helpSimpler design may be better
PathNext steps depend on new observationsSteps and rules are known in advance
JudgmentInputs are messy and require interpretationOutputs must follow exact calculations
ToolsReliable read or reversible tools existCritical systems lack safe interfaces
VerificationSuccess can be checked from environment stateCorrectness is hard to observe
ConsequenceErrors can be contained or approvedOne wrong action creates unacceptable harm
EconomicsVariability justifies added calls and latencyHigh volume rewards a fixed efficient path

Evidence & context: Anthropic

Test simpler alternatives first

  • A deterministic rule for exact policy or calculation.
  • One model call for classification, extraction or drafting.
  • Retrieval plus an answer when the need is information, not action.
  • A fixed AI-enabled workflow when stages are known.
  • A human-owned process when the decision requires accountable judgment.

This sequence is not anti-agent. It keeps complexity proportional to the unresolved variability. Anthropic's published distinction similarly recommends workflows for predefined paths and agents where model-directed choice is genuinely needed.

Evidence & context: Anthropic

Assess action, not only answer quality

A research agent that proposes sources can be corrected before publication. An agent that sends money, changes access or communicates externally creates a different risk even if its language quality is identical. Consider reversibility, blast radius, sensitive data and who can detect an error in time.

Require human approval when consequences exceed the system's tested ability to decide. Confidence scores can help route cases, but they are not proof of correctness or permission.

Evidence & context: NIST

Use a staged decision

  1. Define the goal and accepted final state.
  2. Map fixed and variable decisions.
  3. Confirm tools, permissions and observable outcomes.
  4. Estimate full cost, including review and failure.
  5. Run representative cases in a contained environment.
  6. Compare against the simplest credible alternative.
  7. Proceed only if added autonomy improves the measured task.

Sources & further reading

  1. Building effective agents

    Anthropic. A provider's engineering taxonomy of agents and workflows, not a universal industry definition. We use the conceptual distinction, not its changing product recommendations.

  2. Generative Artificial Intelligence Profile (NIST AI 600-1)

    NIST. Risk-management guidance, including confabulation. It does not establish a universal error rate.

  3. Evaluate a model router

    Microsoft Learn. Official guidance for evaluating routing across representative workloads using quality, cost, latency and policy criteria. It is not evidence that routing always improves results.

Examples and exercises are illustrative unless attributed to a source. No independent expert review is claimed.

A correction, a counterexample or an experience worth sharing?

Join the conversation ↗