A Framework for
Aligned Intelligence
"The gap between who you are and how AI perceives you is not a technical problem. It is a meaning problem."
The Phynom Protocol is QIQ's core methodology for understanding, measuring, and closing the distance between a brand's declared intent and its latent reality in AI systems. It is built on three disciplines — Measure, Position, and Steward — operating as a continuous cycle rather than a one-time intervention.
The full framework as a PDF — methodology, diagrams, TAG Schema reference, and implementation guide. Free for anyone building in this space.
Download PDF — Free →Latent Space
Optimisation
Every AI language model builds an internal representation of the world during training. This representation — called latent space — is a multi-dimensional map of concepts, entities, relationships, and associations. When an AI responds to a query, it draws on this map to construct its answer.
Your brand, your name, your expertise — they all exist somewhere in that map. The question is not whether AI has a representation of you. It does. The question is whether that representation is accurate, aligned with your intent, and positioned where you want to be.
"Latent Space Optimisation is the practice of intentionally shaping AI's internal model of your brand — not through manipulation, but through semantic coherence, signal alignment, and consistent intent."
LSO is to AI systems what SEO was to search engines — except the signals are subtler, the stakes are higher, and the window to establish early positioning is closing fast. Brands that understand this now will own their categories in AI's knowledge space. Brands that don't will be defined by default.
Traditional SEO
- Optimises for keyword ranking
- Targets search engine crawlers
- Measured in clicks and impressions
- Works on page-level signals
- Primarily technical
Latent Space Optimisation
- Optimises for semantic positioning
- Targets AI latent representations
- Measured in perception alignment
- Works on entity-level signals
- Primarily meaning-based
The core insight of LSO is that meaning precedes visibility. Before an AI can recommend you, cite you, or describe you accurately — it needs a coherent, well-defined model of who you are. That model is built from every signal you put into the world — your content, your language, your associations, your consistency.
Where Phynom
comes from
The word Phynom is a deliberate fusion of two philosophical concepts that sit at the heart of what QIQ does — phenomenon and noumenon. These two terms, drawn from Kantian philosophy, describe the fundamental tension between how things appear and what they actually are.
In the context of AI and brand perception, this tension is not abstract — it is the lived reality of every brand that has ever asked an AI system about itself and received an answer that felt both familiar and wrong.
The world as it is perceived and experienced — the appearance of things as they present themselves to observation. In AI terms: how your brand is described, clustered, and recalled by AI systems.
The thing-in-itself — reality as it exists independent of perception. In AI terms: your actual identity, values, expertise, and intent as you experience and declare them.
The Phynom Protocol exists precisely in the space between these two — the gap between the noumenal brand (what you actually are) and the phenomenal brand (what AI perceives you to be). Closing that gap is the work.
Core Principals of the Phynom Protocol
Measure. Position.
Steward.
The Phynom Protocol is not a linear process — it is a continuous cycle. Each pillar feeds the next, and the cycle repeats as AI models update, markets shift, and your brand evolves. The goal is not a fixed destination but a maintained state of alignment.
Measure
Before you can close the gap you need to see it. The Measure pillar is the systematic interrogation of AI systems to map your current latent position — how you are described, clustered, and associated right now.
This produces your Observed Latent Vector (OLV) — a precise picture of AI's current perception of your brand across multiple systems and query types.
- AI perception audit across ChatGPT, Claude, Grok
- Competitive latent mapping
- Intent gap quantification
- Misalignment flag identification
Position
Positioning in latent space is not about gaming algorithms — it is about achieving coherence. When your declared intent, your content, your language, and your associations all point in the same direction, AI models align with your meaning.
This pillar produces your TAG Schema — a structured set of semantic anchors designed to move your Observed Latent Vector toward your Declared Intent Vector.
- Declared Intent Vector (DIV) definition
- TAG Schema development
- Signal architecture design
- Content realignment roadmap
Steward
Latent space is not static. AI models are retrained. New content enters the ecosystem. Competitors reposition. Without ongoing stewardship, even a well-aligned brand drifts. The Steward pillar is the discipline of maintaining alignment over time.
This pillar operates as a continuous monitoring and adjustment system — tracking your latent position, identifying drift, and reinforcing your signal architecture.
- Monthly latent position monitoring
- Drift detection and alerts
- TAG Schema maintenance
- Quarterly alignment reporting
The Mechanics of Alignment
Semantic Anchors for Latent Positioning
A TAG (Thematic Alignment Geometry) is a carefully chosen word, phrase, or concept that consistently signals your intended latent position to AI systems. TAGs are not keywords — they are semantic anchors that cluster your brand with the concepts, values, and entities you want to be associated with.
A TAG Schema is the full set of primary, secondary, and contextual TAGs that define your latent positioning strategy. When applied consistently across your content, profile, and communications, they gradually shift your Observed Latent Vector toward your Declared Intent Vector.
Declared Intent vs. Observed Latent Vector
The Declared Intent Vector (DIV) is how you intend to be positioned — the precise description of your brand's meaning, values, and category that you would give to an AI if asked directly.
The Observed Latent Vector (OLV) is how AI systems actually describe and position you, extracted through structured Phynom queries. The distance between DIV and OLV is your alignment gap — and closing it is the entire work of the Phynom Protocol.
Apply the Framework
Where would you like to begin?
The Phynom Protocol applies at every scale — from how you personally work with AI, to how your brand is understood across AI systems.
For individuals · Discover & Explore
Apply this to how you work with AI personally
Start with who you are in AI. Build a PersonAI file that anchors every conversation to your actual identity, thinking, and intent. Then explore the Skills that let you work with AI the way you think.
Begin with
For brands · Expand
Apply this to how AI understands your brand
Find out how AI currently represents your brand — your latent position across ChatGPT, Claude, and Grok. Then close the gap between what you mean and what AI understands.
Begin with
The framework is the same. The entry point depends on where you are. Both paths lead to the same place — clarity, alignment, and intentional presence in an AI-mediated world.