What goes in a company identity database?

Include your mission statement, vision, core values with behavioral examples, brand positioning statement, unique differentiators, target audience definitions, brand voice attributes with writing examples, messaging pillars with proof points, competitive positioning, boilerplate descriptions at various lengths, and approved terminology. Structure everything as queryable fields, not long-form documents.

How does a company identity database improve AI content?

When AI tools generate content, they can query the identity database for relevant brand context, voice guidelines, and messaging frameworks. This grounds every output in your actual brand identity rather than generic patterns. The result is content that sounds like your company across all channels, models, and team members who prompt the AI.

How is this different from a brand guidelines PDF?

A brand guidelines PDF is a static document that humans read. A company identity database is a structured, versioned, queryable system that both humans and machines consume. It can be injected into AI prompts, queried by content tools, and updated without redistributing a file. It is living infrastructure, not a document.

How do you structure a company identity database?

Organize into sections: foundation (mission, vision, values), positioning (differentiators, competitive stance, category), audience (ICPs, personas, pain points), voice (attributes, tone modifiers, examples), and messaging (pillars, proof points, objection handling). Use key-value pairs and tagged content blocks for easy retrieval.

How often should you update your company identity database?

Review quarterly and update when strategic direction shifts. Version every change so you can track how positioning evolves. Minor updates like new proof points or customer quotes happen continuously. Major changes to mission, positioning, or voice require cross-functional review. Treat it like a living product with its own changelog.

Intelligent OperationsDeep Dives

The 23-Column Company Identity Framework

A structured approach to codifying who your company is.

The Prompt Engineering Project February 13, 2025 13 min read

Quick Answer

A company identity database is a structured, queryable repository that stores your mission, vision, values, positioning, voice guidelines, differentiators, and messaging frameworks in a format that both humans and AI systems can consume. It replaces scattered brand documents with a single source of truth that feeds content generation, sales enablement, and product copy. Every AI-generated output becomes more consistent when grounded in a well-structured identity database.

Every company has a brand. Most companies cannot describe that brand in a way that a machine can use. They have a logo file, a color palette in a PDF somewhere, maybe a tone-of-voice document that was written three years ago and has not been updated since. When a human writer needs to produce on-brand content, they absorb these materials through osmosis and intuition. When an AI agent needs to produce on-brand content, it needs structured data.

The Company Identity prompt library solves this by organizing everything an AI agent needs to know about a company into 23 typed columns. Each column captures a specific dimension of identity. Each row is a complete, machine-readable identity document for one company or product line. The result is not a brand guidelines PDF. It is a queryable database that any AI agent can reference to produce consistent, on-brand output across every touchpoint.

This article walks through all 23 columns, organized into six groups. For each column, we cover what it captures, how to fill it well, and how AI agents use the data in practice.

Group 1: Core Identity

The first four columns define who the company is at the most fundamental level. These are the columns that every other column builds on.

Column 1: Company Name. The canonical name, including any specific capitalization, spacing, or stylistic conventions. If the company is "IntelligentOperations.ai" and not "Intelligent Operations," this column makes that explicit. AI agents reference this for every mention of the company name, ensuring consistency across thousands of generated documents.

Column 2: Tagline. The single sentence that captures the company's promise. This is not a mission statement. It is the phrase that appears below the logo, in the hero section, on the business card. A good tagline is specific enough to differentiate and concise enough to remember. AI agents use it for headlines, introductions, and any context where the company needs to be described in one line.

Column 3: Mission. What the company exists to accomplish. The mission is present-tense and action-oriented: "We build infrastructure for prompt engineering at scale." AI agents use the mission when generating about-us content, investor materials, job descriptions, and any long-form content that needs to anchor back to purpose.

Column 4: Vision. Where the company is going. The vision is future-tense and aspirational but specific: "A world where every business decision is informed by structured AI intelligence." AI agents use the vision for strategic content, thought leadership, and forward-looking narratives.

Column 5: Values. The principles that govern how the company operates. Values should be stated as behaviors, not abstract nouns. "We ship documentation before features" is a value. "Innovation" is not. AI agents use values to calibrate tone and make judgment calls in content generation -- when a value like "clarity over cleverness" is specified, the agent knows to prefer direct language over creative metaphors.

State values as behaviors, not nouns. "We document everything" is actionable. "Transparency" is ambiguous. AI agents need specificity to make consistent decisions.

Group 2: Market Position

The market position columns tell the AI agent where the company sits in its competitive landscape. Without this context, AI-generated content is generic. With it, the content is differentiated and strategically positioned.

Column 6: Industry. The specific market category the company operates in. Not just "technology" but "AI infrastructure for marketing and operations teams." The more specific the industry definition, the more precisely the AI agent can target its language to the audience that exists within that industry.

Column 7: Target Market. Who the company sells to, described with enough specificity that an AI agent can adjust its register. "Series A to Series C B2B SaaS companies with 50-500 employees" is a target market. "Businesses" is not. AI agents use this column to calibrate vocabulary, assumed knowledge level, and the types of examples and analogies they use.

Column 8: Competitive Landscape. The three to five closest competitors, with a one-sentence description of how each positions itself. This is not a competitive analysis. It is a positioning map that tells the AI agent what claims have already been staked so it can avoid generic positioning and emphasize genuine differentiation.

Column 9: Unique Value Proposition. The single capability or approach that no competitor matches. This column answers the question: "Why should someone choose us instead of the alternative?" AI agents use the UVP as the primary differentiator in sales content, landing pages, and comparison materials.

Without competitive context, AI-generated content is generic. With it, every piece of content is strategically positioned against the alternatives your customers are actually considering.

Group 3: Brand Personality

Brand personality is where most brand guidelines fail AI agents. Describing a voice as "professional yet approachable" gives a model almost no useful information. These four columns force specificity.

Column 10: Voice. The consistent character of the brand's communication. Voice is defined by what the brand always sounds like, regardless of context. A voice description should include at least three positive attributes and three negative ones: "Direct, technical, confident. Never casual, never condescending, never hedging." AI agents use the voice column as a persistent style constraint across all content types.

Column 11: Tone Range. How the voice flexes across different contexts. While voice is constant, tone varies. This column specifies the range: "In documentation, tone is instructional and precise. In blog posts, tone is conversational and opinionated. In error messages, tone is calm and solution-oriented." AI agents use this to adjust register based on the content type they are generating.

Column 12: Communication Style. The structural preferences for how the brand communicates. Does it prefer short sentences or complex ones? Active voice or occasional passive? Lists or prose? Concrete examples or abstract principles? This column governs the micro-level writing decisions that accumulate into a recognizable style.

Column 13: Vocabulary. The specific words the brand uses and avoids. This includes preferred terms ("prompt library" not "prompt template"), industry jargon that is acceptable ("LLM," "token window"), and words that are banned ("utilize" instead of "use," "leverage" as a verb, "synergy" in any context). AI agents reference this column for word-level decisions that would otherwise default to the model's training distribution.

brand-personality-example.json

{
  "voice": {
    "always": ["direct", "technical", "confident", "precise"],
    "never": ["casual", "condescending", "hedging", "hyperbolic"]
  },
  "tone_range": {
    "documentation": "instructional, precise, no filler",
    "blog": "opinionated, conversational, evidence-based",
    "marketing": "confident, benefit-focused, specific",
    "error_messages": "calm, solution-oriented, no blame"
  },
  "communication_style": {
    "sentence_length": "short to medium, rarely compound",
    "voice": "active, first-person plural for company",
    "structure": "lead with the point, evidence follows",
    "examples": "concrete and specific, never hypothetical"
  },
  "vocabulary": {
    "preferred": {
      "prompt library": "not 'prompt template'",
      "system prompt": "not 'AI instructions'",
      "structured output": "not 'formatted response'"
    },
    "banned": ["utilize", "leverage (verb)", "synergy",
               "revolutionize", "game-changing", "cutting-edge"]
  }
}

Group 4: Visual Identity

Visual identity columns inform AI agents that generate visual content, select imagery, or produce content that references the brand's visual system.

Column 14: Color Palette. The complete color system, specified in hex values with semantic labels. Not just "primary: #7C3AED" but "primary-400: #7C3AED (used for interactive elements, links, and emphasis in dark mode)." AI agents that generate HTML, CSS, or design specifications need this level of detail to produce on-brand visual output.

Column 15: Typography. The font stack, with specific weights and sizes for each level of hierarchy. "Headings: Inter 700, 32-56px. Body: Inter 400, 16-18px. Code: JetBrains Mono 400, 14px." AI agents use this when generating documentation, presentations, or any content that specifies typographic treatment.

Column 16: Imagery Style. The visual treatment that characterizes the brand's imagery. "Abstract geometric patterns on dark backgrounds. No stock photography. No photographs of people. Dot grids and gradient accents. Monochromatic with purple (#7C3AED) as the sole accent color." This column guides AI agents that select or generate imagery, and it prevents the default drift toward generic stock photography.

Specify colors with semantic labels, not just hex values. An AI agent that knows a color is "used for interactive elements in dark mode" can apply it correctly in new contexts. An agent that only knows a hex value cannot.

Group 5: Operational Context

Operational context columns give AI agents the background information they need to calibrate the scale, maturity, and constraints of the company. This prevents an AI agent from writing enterprise-scale case studies for a seed-stage startup or suggesting bootstrapped strategies for a venture-backed company burning through its Series B.

Column 17: Founding Story. A two-to-three paragraph narrative of why the company was started, by whom, and what problem motivated its creation. AI agents use this for about-us pages, founder bios, pitch decks, and any content that needs to establish credibility through origin narrative.

Column 18: Team Size. The current headcount and team structure. "12 people: 4 engineers, 2 designers, 3 prompt engineers, 2 operators, 1 founder." AI agents use this to calibrate recommendations and content to the company's actual capacity. A 12-person team does not need an enterprise content calendar. It needs a focused, sustainable publishing cadence.

Column 19: Growth Stage. The company's current stage: pre-revenue, seed, Series A, growth, mature, public. AI agents use this to calibrate the ambition and scope of generated content. A seed-stage company talks about potential and vision. A growth-stage company talks about traction and scale.

Column 20: Revenue Model. How the company makes money. "SaaS subscription with three tiers: Starter ($49/mo), Professional ($199/mo), Enterprise (custom pricing)." AI agents use this for pricing pages, sales content, ROI calculators, and any content that touches on the economics of the product.

Operational context prevents AI from writing enterprise playbooks for startups or bootstrapped strategies for funded companies. Scale-appropriate content requires scale-aware context.

Group 6: Digital Presence

The final three columns define where the company shows up online and what it talks about. These columns are the most frequently updated because digital presence evolves faster than core identity.

Column 21: Website and Properties. The primary domain, any secondary domains, and key pages. AI agents use this for internal linking in blog content, for generating meta descriptions, and for any content that needs to reference specific pages or features.

Column 22: Social Channels. Each active social platform with its handle, posting frequency, and content focus. "LinkedIn: @intelligentops, 3x/week, long-form thought leadership and case studies. GitHub: @intelligentops, continuous, open-source tools and documentation." AI agents use this to tailor content for specific platforms, matching the length, tone, and format expectations of each channel.

Column 23: Content Pillars. The three to five thematic areas that all content maps to. "Prompt engineering craft. AI infrastructure. Design systems. Intelligent operations. Content at scale." Every piece of content the company produces should map to one of these pillars. AI agents use the pillars as a relevance filter: if a content idea does not map to a pillar, it does not get produced.

A Complete Example Row

Below is a condensed example of a complete Company Identity row. In practice, several columns contain multi-paragraph entries. This example shows the structure and the level of specificity each column requires.

company-identity-row-example.json

{
  "company_name": "IntelligentOperations.ai",
  "tagline": "Prompt engineering infrastructure for teams that ship.",
  "mission": "Build the tools and frameworks that make prompt engineering a repeatable, measurable discipline.",
  "vision": "Every business operation informed by structured AI intelligence.",
  "values": [
    "Ship documentation before features",
    "Measure everything, assume nothing",
    "Clarity over cleverness in every sentence",
    "No emoji, no decoration, no filler"
  ],
  "industry": "AI infrastructure for B2B marketing and operations",
  "target_market": "Series A-C B2B SaaS, 50-500 employees, technical leadership",
  "competitive_landscape": [
    "PromptLayer: developer-focused prompt management",
    "Humanloop: enterprise prompt optimization",
    "Langchain: framework-level orchestration"
  ],
  "unique_value_proposition": "68 production-ready prompt libraries with typed schemas, not a framework or SDK.",
  "voice_always": ["direct", "technical", "confident", "precise"],
  "voice_never": ["casual", "condescending", "hedging", "hyperbolic"],
  "tone_range": "Instructional in docs, opinionated in blog, confident in marketing.",
  "communication_style": "Short sentences. Active voice. Lead with the point.",
  "vocabulary_preferred": ["prompt library", "system prompt", "structured output"],
  "vocabulary_banned": ["utilize", "leverage", "synergy", "game-changing"],
  "color_palette": { "primary": "#7C3AED", "surface_dark": "#0B0B0F", "text_primary": "#F4F4F5" },
  "typography": "Inter for prose, JetBrains Mono for code",
  "imagery_style": "Abstract geometric on dark backgrounds. No stock photography.",
  "founding_story": "Started as a single prompt library. Grew to 68 when the pattern proved repeatable.",
  "team_size": "Small, engineering-heavy, design-led",
  "growth_stage": "Early growth, post-product",
  "revenue_model": "SaaS subscription, three tiers",
  "website": "intelligentoperations.ai",
  "social_channels": { "linkedin": "3x/week thought leadership", "github": "open-source tools" },
  "content_pillars": ["prompt engineering", "AI infrastructure", "design systems", "intelligent operations"]
}

When an AI agent receives this row as context, it has everything it needs to generate content that sounds like this company, is positioned correctly in this market, and operates within the constraints of this stage and team size. No guessing. No generic defaults. Every dimension of identity is specified and structured.

The Company Identity framework is not a branding exercise. It is an AI enablement exercise. The 23 columns exist because AI agents need structured, complete, machine-readable identity data to produce consistent output. A brand guidelines PDF is useful for humans. A 23-column database record is useful for machines. And in an era where machines are producing an increasing share of your content, the machine-readable version is the one that matters.

Key Takeaways

AI agents cannot absorb brand identity through osmosis the way human writers can. They need structured, typed, queryable identity data.

The 23 columns are organized into six groups: Core Identity, Market Position, Brand Personality, Visual Identity, Operational Context, and Digital Presence.

Brand personality columns must be specific. "Professional yet approachable" gives an AI agent no useful information. Explicit voice attributes, tone ranges, and vocabulary lists do.

Operational context prevents scale mismatches -- AI will not write enterprise playbooks for startups when it knows the team size, growth stage, and revenue model.

A complete Company Identity row gives any AI agent everything it needs to produce on-brand, strategically positioned, scale-appropriate content without guessing.

Frequently Asked Questions

Common questions about this topic

Building Customer Intelligence Databases with AI Team Structures for AI-Augmented Organizations

Intelligent Operations