Skip to content
storywi.se documentation storywi.se documentation storywi.se documentation

AI Instructions

Guides

AI Instructions - Customize AI Behavior

AI Instructions allow you to customize how the AI behaves across your organization and projects. Set organization-wide standards that can be overridden at the project level, ensuring the AI follows your company’s specific writing style, naming conventions, and project requirements.

Overview

AI Instructions provide a powerful way to guide the AI’s behavior in Storywise. You can define how user stories are written, how epics are structured, how acceptance criteria are formatted, and provide context about your company’s practices and project-specific requirements. These instructions are applied to all AI calls, including the AI Chat, ensuring consistent output that matches your organization’s standards.

Structure of AI Instructions

AI Instructions are organized into two main categories:

General Context

Context Passed to Each AI Request

General context is information that is included with every AI request to provide background and constraints:

  • Tenant Context - Organization-level context that applies to all projects in your organization

    • Company-wide standards and practices
    • Organization naming conventions
    • General business context
    • Default behavior for all projects

  • Project Context - Project-specific context that overrides or supplements tenant context

    • Project-specific requirements (e.g., “This is an offline project that must work without internet”)
    • Domain-specific terminology for this project
    • Special constraints or regulations (e.g., HIPAA compliance)
    • Project-specific business context

Rules for Entity Creation

Guidelines for Creating Specific Entities

Rules define how specific entities should be created and formatted. You can set rules for:

  • Epics - How epics should be structured, named, and described
  • Stories - Format and style for user stories
  • Tasks - How tasks should be written and organized
  • Original Sentences - Guidelines for processing original requirement sentences
  • One Sentences - Format for one-sentence summaries
  • Roles, Entities, and Glossary Entries - Naming and definition standards
  • Acceptance Criteria - Format and style requirements for acceptance criteria
  • Test Scenarios - How test scenarios should be written (for acceptance criteria)
  • Features - Structure and naming conventions for features
  • Releases - Format and organization standards for releases

Each rule type can be configured at both the tenant (organization) and project level, with project-level rules overriding tenant-level defaults.

Common Use Cases

Tenant Context Examples

Organization-wide context that applies to all projects:

  • “We use camelCase for all technical terms and feature names”
  • “All acceptance criteria must be written as numbered lists with Given-When-Then format”
  • “Epics should be written as high-level features, not as collections of stories”

Project Context Examples

Project-specific context that overrides or supplements tenant context:

  • “This is an offline project - all functionality must work without internet connectivity”
  • “This project is for a healthcare application - all features must comply with HIPAA regulations”
  • “This is a mobile-first project - prioritize mobile user experience in all suggestions”
  • “This is a financial services project - use banking terminology and ensure PCI compliance”
  • “This project uses domain-driven design principles - structure entities accordingly”

Entity-Specific Rules Examples

Rules for creating specific entity types:

Stories:

  • “User stories must include a clear value statement in the ‘so that’ clause”
  • “Each story should be independently testable and deliverable”

Epics:

  • “Epics should be named using the pattern: [Module]-[Functionality]”
  • “Epic descriptions must include business objectives”

Acceptance Criteria:

  • “Acceptance criteria must be written in Given-When-Then format”
  • “Each criterion should be testable and measurable”

Roles and Entities:

  • “User roles should always bear a persona name and get a description”

Test Scenarios:

  • “Test scenarios must be written in Gherkin format”
  • “Each scenario should cover positive and negative test cases”

Configuration

Setting Up AI Instructions

  1. Configure Tenant Context - Set organization-wide context in organization settings
  2. Set Entity Rules - Define rules for each entity type (Epics, Stories, Tasks, etc.) at the organization level
  3. Add Project Context - Override or supplement with project-specific context when needed
  4. Override Project Rules - Customize entity rules for specific projects if necessary