Prompt Lab

Prompt Lab is EYE's real-time prompt testing and research environment. Instead of waiting for a full analysis run, you can test any query instantly across all AI platforms, build and organize prompt collections, run A/B variant comparisons, simulate buyer personas, and explore topic ownership with Research Mode.

The Four Tabs

Test

The core of Prompt Lab. Type any query, select your platforms, and get real AI responses in seconds.

Results show per platform:

• Whether the brand was mentioned
• Position in the response (1st, 2nd, 3rd…)
• Sentiment (positive / neutral / negative)
• Competitors mentioned
• Full AI response text
• Response time

Save to Library is always visible on the right panel. Before saving, you can tag the prompt with:

• Intent — Informational, Transactional, Comparison, Navigational
• Journey stage — Awareness, Consideration, Decision, Retention
• Collection — Assign to a named prompt collection
• A/B group ID — Group this prompt with variants for comparison

Library

All your saved prompts in one place, with filtering, sorting, and visual performance indicators.

Filters:

• By collection
• By intent or journey stage
• By active/inactive status
• Full-text search

Sort by: Most recent · Highest score · Visibility rate

Each prompt card shows:

• Per-platform mention indicators (green = mentioned, red = not found)
• Visibility percentage across tested platforms
• Priority score (0–100) based on mention rate, position, and sentiment
• Collection, intent, stage, and persona badges
• Active/inactive toggle

A/B Variant Groups

Prompts that share the same A/B group ID are automatically grouped together in the Library with a side-by-side comparison view. Each variant is labeled A, B, C… and the highest-scoring variant is marked ★ Best.

This lets you test different phrasings of the same query and immediately see which version performs better across AI platforms.

Industry Template Library

Click Templates to import pre-built prompt sets for your industry. Available industries:

• Healthcare & Pharma
• SaaS & Technology
• Retail & E-commerce
• Finance & Insurance
• Travel & Hospitality

Each template imports 5 prompts covering awareness, consideration, and decision stages — ready to test immediately.

Collections

Collections are named groups of prompts — the equivalent of folders. Use them to organize prompts by client, campaign, product line, or topic.

Creating a collection:

• Name (required)
• Industry (optional, for context)
• Color (for visual identification)

Batch Testing: Run an entire collection against all selected platforms in one click. After the run completes, you'll see a summary:

• Prompts tested
• Prompts where the brand was visible
• Average position across all results

Research Mode

Research Mode generates a full-funnel prompt set for any topic and tests it across platforms automatically — without you having to write a single prompt.

How it works:

1. Enter a topic (e.g. "immune health supplements", "CRM for startups", "travel insurance")
2. Optionally enter a brand to track visibility for
3. Click Run Research

EYE generates 6 prompts covering the full buyer funnel (2 awareness, 2 consideration, 2 decision), runs them across your selected platforms, and returns:

• Results grouped by funnel stage
• Per-platform mention indicators for each prompt
• Overall brand visibility percentage (if a brand was specified)
• Key Insights — auto-generated observations about brand presence, top competitors, most-cited sources, and decision-stage coverage

Use Research Mode to:

• Explore a new industry or topic before setting up a full project
• Quickly assess brand visibility for a prospective client
• Identify which funnel stages have the weakest brand presence
• Discover which competitors dominate a topic in AI responses

Any prompt from Research Mode can be sent directly to the Test tab with one click.

Personas

Personas simulate different buyer types, adding context to how a query is interpreted by AI platforms. This is especially useful when the same query might yield different brand recommendations depending on who is asking.

Built-in personas:

You can also write a custom persona — any free-text description of a buyer type relevant to your industry.

The active persona is passed to the AI platforms as context with every test, and is saved alongside the prompt when you add it to your library.

Scoring

Every saved prompt that has been tested receives a priority score from 0 to 100:

The score combines mention rate (40%), position score (30%), and sentiment (30%). Use it to prioritize which prompts need content or PR work.

Typical Workflows

Before a client pitch: Use Research Mode to run a topic analysis for the prospect's industry. In 2 minutes you have brand visibility data across 4 AI platforms to show in the meeting.

A/B testing prompt phrasing: Save two versions of the same query with the same A/B group ID. Test both. The Library shows them side-by-side with scores — pick the winner.

Building a prompt library for a new industry: Import an industry template, test all prompts with Batch Test, review which ones surface your brand, then promote the best-performing ones to your main analysis configuration.

Persona-aware testing: Test the same query with "Price-Sensitive Buyer" vs "Senior Executive" personas to see if AI platforms recommend different brands depending on who is asking.