Original framework

Choose the operating pattern before choosing the AI technology.

The Workflow Suitability Framework helps teams decide whether a workflow calls for deterministic automation, enterprise search, retrieval-augmented generation, an agentic system, predictive machine learning, or no AI at all.

Executive summary

Start with the job, uncertainty, source material, and cost of failure.

Job shape

Is the work a fixed transformation, information lookup, grounded synthesis, multi-step execution, or prediction?

Input stability

Are the rules and inputs structured and predictable, or variable and dependent on interpretation?

Failure exposure

Can an error be reversed cheaply, or could it create financial, safety, compliance, or customer harm?

Evidence need

Must every answer be traceable to approved sources, records, or model features?

Action authority

Should the system only recommend, or may it write data, trigger tools, and advance a process?

Evaluation path

Can quality be measured before launch and monitored under real operating conditions?

Decision map

Match the workflow to the least complex pattern that can do the job reliably.

PatternBest fitPrimary controlWarning sign
Deterministic automationStable rules, structured inputs, repeatable transformationsTests, validation rules, exception queuesThe task depends on open-ended interpretation
Enterprise searchUsers need to find exact approved documents or recordsPermissions, indexing, ranking, source freshnessThe user actually needs a synthesized answer
RAGAnswers must synthesize changing, approved source materialCitations, retrieval evaluation, abstentionThe corpus is incomplete, stale, or poorly owned
Agentic workflowBounded multi-step work across tools with recoverable actionsTool permissions, checkpoints, logs, rollbackErrors are irreversible or cannot be detected
Predictive MLHistorical patterns can improve forecasts, scores, or anomaly detectionBacktesting, drift monitoring, decision thresholdsThere is no stable outcome label or usable history
No AI yetThe workflow, ownership, value, or source quality is unresolvedProcess redesign and data remediationA tool is being used to avoid fixing the process
Selection sequence

Use a six-step routing test.

  1. Define the exact input, task, output, user, and business consequence.
  2. Remove steps that can be simplified or eliminated without AI.
  3. Identify whether the core need is retrieval, synthesis, execution, or prediction.
  4. Set the maximum acceptable action authority and failure exposure.
  5. Choose the simplest viable pattern and define a measurable baseline.
  6. Pilot against real exceptions before expanding scope or autonomy.
Complexity rule

Do not use an agent where a rules engine works, do not use generation where search works, and do not use AI where a process fix removes the problem.

Hybrid systems

Real workflows often combine patterns, but each boundary should remain explicit.

Search + RAG

Search exposes exact records while RAG synthesizes an answer with citations. Evaluate retrieval and answer quality separately.

RAG + agent

Retrieval grounds a recommendation before an agent takes a bounded action. Require approval for consequential steps.

Predictive ML + rules

A score informs a deterministic threshold or human decision. Monitor both model drift and downstream outcomes.

Automation + human review

Rules complete the routine path while exceptions route to a person with enough context to decide.

Stop conditions

“No AI yet” is a valid and often valuable result.

  • No accountable workflow owner or adoption path
  • No baseline against which improvement can be measured
  • Unusable source data or unclear access rights
  • Failure consequences exceed the available controls
  • Low process volume or value cannot justify operating cost
Next step

Resolve the process, ownership, data, or control gap first. Then repeat the suitability review with a narrower workflow target.

Apply the framework

Turn technology selection into an accountable implementation decision.

Use this framework after the AI Readiness Framework establishes that a workflow is worth pursuing. Then connect the selected pattern to the appropriate implementation and evaluation plan.