Agent Step Prompts

Last Updated: May 19, 2026

Vasion Automate Workflow Agent Step prompts are natural language instructions that help users analyze data and generate or rewrite information from upstream sources.

Agent steps are restricted to generating text only output. Keep this restriction in mind when formulating prompts to optimize output and ensure workflow continuation.

Operation Prompts

The Agent step interprets natural language to analyze and generate output text. The following operations and examples demonstrate different approaches based on your specific needs.

Analysis Prompts

Used to evaluate, assess, or review data from upstream workflow steps.

Simple Analysis

"Review [data source] and identify [what to look for]."

  • Review this invoice and flag any line items that exceed $5,000.
  • Check this document for missing required fields and list what is absent.
  • Analyze this claim and list any common denial triggers.
Scored Assessment

"Evaluate [data source] against [criteria] and provide a [score / rating] with reasoning"

  • Evaluate this contract against our standard terms checklist and provide a completeness score from 1-10 with a list of missing clauses
  • Assess this support request for urgency (high / medium / low) and explain your reasoning
  • Review this application for completeness and rate each section as complete, partial, or missing

Generation Prompts

Used to produce new content based on workflow data.

Text Generation

"Using [data from previous steps], generate [output]"

  • Using the patient data and diagnosis codes from the previous steps, generate a letter of medical necessity for prior authorization.
  • Using this form submission, draft a personalized acknowledgment email with the submitter's name and expected timeline.
  • Using these extracted invoice fields, generate a summary suitable for the approver.
Rewriting and Formatting

"Rewrite [input] to meet [standard / format]"

  • Rewrite these free-text notes into standardized language that meets our compliance formatting requirements.
  • Reformat this extracted data into a clean summary with labeled fields: Name, Date, Amount, Status.

Prompt Approaches

Choose your prompting approach based on how well-defined the task is and what output you need.

Direct Instruction

Use clear, specific prompts when you know exactly what output you need. This approach works best when the task is well-defined.

  • Tell the agent exactly what to look for and how to format the output.

  • Specify the structure: "Respond with: [field 1], [field 2], [field 3]".

  • Include the criteria or rules the agent should apply.

Contextual Analysis

Use open-ended prompts when you want the agent to interpret and reason about the data. This approach works well for review and assessment tasks.

  • Let the agent identify issues, patterns, or anomalies.

  • Ask for reasoning alongside conclusions.

  • Refine with follow-up instructions in subsequent agent steps.

Recommendations

Following these recommendations helps customers get the best results from the Agent Step:

Define your output structure in the prompt: Since the agent outputs text, tell it how to format the response. This approach makes the output predictable and useful.

Reference previous step data explicitly: Mention what data is coming from the prior step so the agent knows what inputs to expect.

Keep instructions focused on one task: One agent step should do one cognitive job. Chain multiple agent steps for multi-stage analysis rather than cramming everything into one prompt.

Test with real workflow data: Prompt quality depends on the data flowing through. Test with actual documents and form submissions.

Remember output goes to Email: To deliver agent output to a user, add an Email step after the agent step and include the agent output variable in the email body. For configuring an email step, see Email Step.

Next Steps