Screenshot Business Analyzer Agent

Extracts functional requirements, data entities, business rules, and domain logic from UI screenshots for product analysis.

When to Use This Agent

Choose this agent when you need to:

• Identify business functions, data entities, and domain-specific workflows visible in application screenshots
• Extract functional requirements and value propositions from UI designs without access to source code or documentation
• Produce structured JSON analysis of product capabilities, business rules, and data relationships from visual evidence

Consider alternatives when:

• You need to analyze visual design quality, typography, and color theory rather than business function extraction
• Your goal is interaction pattern analysis or user flow mapping, which the interaction analyzer agent handles specifically

Quick Start

Configuration

Copy
name: screenshot-business-analyzer-agent
type: agent
category: ui-analysis

Example Invocation

Copy
claude agent:invoke screenshot-business-analyzer-agent "Analyze the business functions in this e-commerce dashboard screenshot"

Example Output

Business Analysis: E-Commerce Dashboard

Product Domain: E-commerce management platform

Functional Modules:
  [Core] Order Management
    - Order listing with status filters
    - Bulk status update operations
    - Revenue summary calculations
    Priority: core

  [Core] Inventory Control
    - Stock level monitoring
    - Low-stock threshold alerts
    - Supplier reorder triggers
    Priority: core

  [Supporting] Customer Analytics
    - Cohort retention charts
    - Lifetime value calculations
    Priority: supporting

Data Entities: 5 identified
  - Order (id, status, total, customer_id, created_at)
  - Product (sku, name, stock_level, category)
  - Customer (name, email, order_count, ltv)
  ...

Business Rules: 8 identified
  - Orders transition: pending -> processing -> shipped -> delivered
  - Low stock alert triggers at threshold <= 10 units
  ...

Core Concepts

Business Analysis Dimensions

Aspect	Details
Functional Modules	Core features driving primary value, supporting features, and administrative functions identified from UI elements
Data Entities	Objects displayed in the UI with visible attributes, CRUD operation indicators, and inter-entity relationships
Business Rules	Validation constraints, permission indicators, workflow state machines, and conditional logic implied by UI state
Domain Concepts	Industry-specific terminology, categorization taxonomies, status workflows, and business process stages
Value Analysis	Core value proposition, differentiating features, monetization signals, and user engagement mechanisms

Analysis Pipeline Architecture

┌────────────────┐    ┌────────────────┐    ┌────────────────┐
│  Screenshot     │───>│  Element       │───>│  Module        │
│  Input          │    │  Identification│    │  Grouping      │
│                 │    │                │    │                │
│ Raw UI capture  │    │ Buttons, tables│    │ Core features  │
│                 │    │ Forms, labels  │    │ Support funcs  │
│                 │    │ Status badges  │    │ Admin tools    │
└────────────────┘    └────────────────┘    └───────┬────────┘
                                                     │
┌────────────────┐    ┌────────────────┐            │
│  Structured    │<───│  Domain        │<───────────┘
│  JSON Output   │    │  Extraction    │
│                │    │                │
│ Modules, rules │    │ Entities       │
│ Entities, value│    │ Rules, flows   │
│ Workflows      │    │ Terminology    │
└────────────────┘    └────────────────┘

Configuration

Parameter	Type	Default	Description
analysisFocus	string	"comprehensive"	Analysis scope: comprehensive, modules-only, entities-only, or rules-only
outputFormat	string	"json"	Structured output format: json, markdown-table, or narrative prose
domainHint	string	""	Optional industry domain hint to improve terminology recognition (e.g., "healthcare", "fintech")
priorityClassification	boolean	true	Classify modules as core, supporting, or administrative priority levels
includeValueAnalysis	boolean	true	Include monetization signals, value propositions, and engagement feature analysis

Best Practices

1. Focus on What the System Does, Not How It Is Built - Business analysis extracts functional requirements, not implementation details. Identify that a screen shows "order status tracking" rather than speculating about database schemas or API designs. The output should be meaningful to product managers and business analysts, not just developers.

2. Distinguish Data Entities from Display Artifacts - A table column header represents a data attribute, but a sorting icon is a UI interaction element. Carefully separate genuine data entities and their attributes from presentational elements. A "Last Modified" column reveals a timestamp attribute on the entity, while a pagination control reveals collection size, not entity structure.

3. Infer Business Rules from Visual State Indicators - Greyed-out buttons imply permission restrictions or prerequisite conditions. Color-coded status badges reveal workflow state machines. Disabled form fields suggest derived or system-managed values. Each visual signal encodes business logic that should be captured as an explicit rule in the analysis output.

4. Map Entity Relationships Through Co-occurrence - When a customer name appears in an order table, it reveals a relationship between Customer and Order entities. Column headers like "Assigned Team" within a ticket list expose assignment relationships. Trace these connections to build a domain relationship map from purely visual evidence.

5. Classify Feature Priority by Visual Prominence - UI designers allocate screen real estate proportionally to feature importance. Primary navigation items, hero sections, and prominent CTAs indicate core features. Secondary sidebar items and footer links suggest supporting functions. Settings and configuration screens typically represent administrative capabilities.

Common Issues

1. Confusing UI Components with Business Features - A dropdown menu is a UI component, not a business feature. The business feature is "category filtering" or "status selection." Analysis should abstract from specific widget types to the underlying business capability they enable, producing findings that remain valid even if the UI implementation changes.

2. Missing Off-Screen Business Logic - A single screenshot captures one state of one screen. Workflows spanning multiple screens, conditional features visible only to certain roles, and error states not currently displayed all represent business logic that cannot be fully captured from a single view. Flag incomplete analysis areas explicitly rather than assuming the visible screen represents the complete system.

3. Over-Specifying Data Entity Attributes - Visible table columns reveal some attributes, but entities typically contain many more attributes not displayed in the current view. Document only the attributes directly observable in the screenshot and note that the entity likely contains additional attributes not visible in this particular screen state.