Skip to content
Standard Deviation

StandardGraph

StandardGraph is the brand-intelligence module — your brand’s living substrate inside StandardOS. Voice, visual identity, audience, AI-search visibility, social sentiment, and competitive context, in one continuously-refreshed graph. Every other module (Planning, Creative, GPT) reads from and writes back to it.

What lives in the graph

Five sub-tools:

Brand Profile

The structured representation of who the brand is:

  • Voice — tone words, do/don’t language patterns, exemplar sentences
  • Visual identity — colors, typography, logo treatments
  • Business context — what the company does, who it serves, what makes it different
  • Audience segments — demographic + lifestyle + psychographic descriptions
  • Source materials — the documents (brand bibles, prior campaigns, research) the AI has read

Auto-extraction pulls a starting profile from your website + Wikipedia/About-page sources in under 30 seconds. From there, you refine. Most brands need 30 minutes of human review to land on a profile they trust.

Search Visibility

How AI search engines describe your brand. The tool runs prompts like “tell me about [your brand]” / “compare [your brand] to [competitor]” against Claude, Gemini, and GPT-4, records the responses, and tracks change over time. Categories you define (e.g. “luxury yacht charter, Caribbean”) get a share-of-voice score per AI model.

This matters because AI search is increasingly upstream of regular search — when someone asks ChatGPT to recommend a yacht charter in the Caribbean, the model’s training data + retrieval determines whether you’re in the answer. Tracking that surface gives you a leading indicator before it shows up in conversion data.

Social Listening

Real-time brand sentiment from Reddit + news (the two surfaces where unprompted opinions live in volume). Track mentions of your brand, your competitors, and any topic relevant to the category. Sentiment scoring + trend lines + search any topic on demand.

Competitive Intelligence

Define competitor categories, compare AI share of voice (from Search Visibility), and surface positioning gaps. The output is a positioning map that updates as the AI-search landscape shifts.

StandardGPT history

Chat threads with StandardGPT (the conversational layer) are stored here, scoped to the brand. Useful when you want to retrieve a previous answer or hand a thread to a teammate.

How a typical setup runs

  1. Brand Profile — Auto-extract → review → save. ~30 minutes the first time.
  2. Define competitive set — Add 3-5 direct competitors. ~5 minutes.
  3. Define categories for Search Visibility — The 1-3 categories you want to rank in across AI search. ~5 minutes.
  4. Let Social Listening + Search Visibility run — Both refresh automatically (Search Visibility weekly, Social Listening continuously). First useful trend lines appear after 7-14 days.

Total active setup: ~40 minutes. Then the graph becomes self-maintaining and you check in on it as a dashboard.

How other modules consume the graph

  • StandardPlanning reads the brand profile + audience segments + competitive set when drafting strategy and audience modules.
  • StandardCreative reads the voice profile + visual identity for every generation. Refuses to generate if the profile is incomplete (see Brand-grounded AI).
  • StandardGPT reads everything in the graph as retrievable context. “Why did we move away from Audience B last quarter?” gets answered from the actual brand history, not made up.

What the graph is not

  • Not a CRM. Customer data lives in your CRM (HubSpot, Salesforce, etc.). The graph models the brand, not the customer base.
  • Not a CMS. Marketing content (drafts, briefs, decks) lives in the modules that produced it (StandardCreative, StandardPlanning). The graph is the why, not the what.
  • Not a data warehouse. It’s a structured representation of the brand for AI grounding, not an analytics layer.

Where to go next