Prompt System & Methodology

How the Brand module works, why it relies on open-ended prompts, and how it differs from Search Analysis.

Understanding how AI models perceive your brand is a strategic necessity. In the era of generative AI, your Share Of Model — the presence and perception of your brand inside LLMs — is a key driver of future growth, and that strategic pursuit begins with the Brand module.

The Brand module reveals the semantic associations LLMs consistently make with your brand, across models and over time. It sets the foundation for all tactical optimisation that follows in Search, Assets and Activation.

What the Brand module measures

The Brand module provides a macro view of how LLMs conceptually understand your brand. It looks beyond mere appearances in specific user questions and systematically maps the recurring concepts, attributes and category cues that models naturally associate with your brand and its competitors.

It answers the foundational question:

What does the model think my brand stands for?

This strategic layer acts as the north star for marketers — establish a foundational understanding before diving into tactical layers.

Why we use open-ended prompts

To map a model’s conceptual landscape, we use a deliberate approach: simple, open-ended prompts. The system is engineered for stability, comparability and depth. Prompts run daily across all major LLMs to consistently surface:

• Concepts common to the category — universal associations models make.
• Associations unique to each brand — what sets your brand apart.
• Differences in model interpretation — how different LLMs perceive the category.

Rather than generating thousands of one-off queries, we focus on consistency over time — essential as new LLMs are released roughly every six weeks.

Example. Across one month, a Brand analysis uses a set of different core prompts, applied repeatedly across models. An average analysis generates over 700 iterations of core strategic questions every month.

Examples of prompts

The analysis relies on simple, open-ended patterns that reveal the conceptual understanding of the category and the brand:

• “What brands come to mind when you think of non-alcoholic spirits?”
• “I’m interested in non-alcoholic spirits — what brands would you recommend?”
• “Which is better in the non-alcoholic spirits category, Brand A or Brand B?”

All client inputs (country, brand names, personas, attributes) act as multipliers: they automatically generate many structured variations of the same prompt pattern, enriching the semantic landscape without losing consistency.

Brand vs. Search analysis

Brand and Search analyses answer different questions and operate at different levels of the marketing journey.

	Brand Analysis	Search Analysis
Purpose	Understand semantic associations and conceptual brand perception (what the model thinks of your brand).	Measure visibility for user queries (where your brand appears for specific questions).
Insight type	Macro, strategic.	Granular, tactical.
Questions used	Simple, open-ended prompts.	Real, high-intent search queries.
Prompt management	Fixed set, for stability and comparability.	Editable by clients, focused on real-time user intent.
Role in the journey	First step (foundation) — get macro insight and context.	Second step (execution) — tactical optimisation and measurement.
Core value	Clarifies what AI models think your brand represents (meaning).	Shows where you appear when consumers ask specific questions (presence).

In short:

• Brand = Meaning (strategic).
• Search = Presence (tactical).

Both layers complement each other, but Brand comes first — it provides the conceptual context needed to properly understand and optimise tactical Search results.

What’s next

Brand Analysis Starter Kit

Run your first Brand analysis end to end.

Search Visibility Overview

The tactical layer that complements Brand.