Improve Brand Architecture, AI Onboarding, Projects & Knowledge/RAG

The current Brand setup needs a clearer hierarchy and onboarding flow.

Right now, users are asked to choose a writing style before AIValance has enough information about the brand. The following Brand Guidance screen then repeats similar information across multiple fields.

I suggest restructuring Brand around an AI-first analysis and inheritance model.

CORE HIERARCHY

Workspace → Brand → Project → Whiteboard → Output

Each level should inherit relevant context from the level above while allowing controlled overrides.

For example, a Project inherits its Brand voice, audience, knowledge and assets, but can override the tone or add project-specific instructions without changing the master Brand.

  1. START BRAND CREATION WITH SOURCE ANALYSIS

Instead of starting with “Pick the closest version of you”, first ask the user to provide sources AIValance can analyse:

• Website URL
• LinkedIn profile or company page
• Other public URLs
• Google Drive
• Uploaded documents
• Pasted text

AIValance should crawl and analyse the available sources and automatically generate a proposed Brand Profile.

The user then reviews and edits the results instead of defining everything manually before AI has any context.

  1. GENERATE A STRUCTURED BRAND PROFILE

AI should extract and propose separate fields such as:

• Brand Summary
• Business
• Positioning
• Offers
• Target Audiences
• Audience problems / Jobs-to-be-Done
• Brand personality
• Tone of voice
• Writing style
• Speaking style
• Content pillars
• Messaging themes
• Preferred vocabulary
• Language to avoid
• Signature phrases

Where possible, AI-generated conclusions should show their source and confidence level so users can understand why something was inferred.

  1. MOVE VOICE CALIBRATION AFTER BRAND ANALYSIS

Direct, Conversational, Educational, Storyteller, High Energy, etc. should not be mutually exclusive choices.

AI should first infer the communication style and then allow users to adjust dimensions such as:

• Directness
• Energy
• Authority
• Warmth
• Playfulness
• Formality
• Educational depth
• Storytelling

The existing radar visualisation could be used for this.

  1. ADD PROJECTS UNDER BRANDS

A Brand should support multiple Projects.

Example:

Workspace
→ AIValance Brand
→ LinkedIn Campaign
→ Product Launch
→ Commercial Film

Each Project should contain:

• Brief and goals
• Project instructions
• Selected audiences
• Selected offers
• Project-specific knowledge
• Assets
• Whiteboards
• Outputs
• Versions and history

Projects inherit Brand defaults but allow specific overrides.

  1. ADD WHITEBOARDS INSIDE PROJECTS

Whiteboards should be project-level thinking environments for research, strategy, brainstorming, customer journeys, campaigns, storyboards and content planning.

Users should be able to:

• Create multiple Whiteboards
• Link existing Whiteboards
• Add documents, images, URLs and AI outputs
• Ask AI about the entire board or a selected frame
• Choose whether a board is available as AI context
• Promote validated information into Project or Brand Knowledge

Whiteboard brainstorming should not automatically become trusted RAG knowledge.

  1. ADD DEDICATED KNOWLEDGE / RAG

Knowledge should be separate from Brand Guidance.

Support:

• Google Drive connection
• Google Drive folder and file selection
• Document uploads
• Websites and URLs
• Pasted text
• Existing Project knowledge
• Approved Whiteboard content

Knowledge should have a clear scope:

Workspace Knowledge
→ Brand Knowledge
→ Project Knowledge
→ Whiteboard Context

Higher-level knowledge becomes available downstream while Project-specific information remains isolated.

Sources should also support authority/status levels such as:

• Canonical
• Approved
• Reference
• Research
• Draft
• Archived

This should help prevent conflicting or outdated information from contaminating RAG results.

  1. ADD BRAND ASSETS

Knowledge and Assets should be separate concepts.

Brand Assets should contain:

• Logos
• Colours
• Fonts
• Product images
• Photography
• Videos
• Icons
• Screenshots
• Audio
• Approved creatives

Projects can inherit Brand Assets while also adding Project-specific assets.

  1. MAKE AUDIENCES AND OFFERS REUSABLE OBJECTS

Target Audience and Offer should not exist only as text fields.

A Brand should support multiple reusable Audiences and Offers that Projects can select.

Example:

Brand
→ Audiences
→ SaaS Founders
→ Agencies
→ Enterprise Teams

Brand
→ Offers
→ Starter
→ Pro
→ Agency
→ Enterprise

This provides stronger and more consistent grounding for AI content generation.

  1. ADD OUTPUTS AS THE FINAL HIERARCHY LEVEL

The workflow should not end at Whiteboards.

Workspace → Brand → Project → Whiteboard → Output

Outputs could include:

• Social posts
• Ads
• Articles
• Emails
• Landing pages
• Images
• Videos
• Commercial scripts
• Presentations
• Campaign material

Outputs should retain their lineage, including:

• Brand
• Project
• Audience
• Offer
• Whiteboard
• Knowledge sources
• Relevant generation settings

  1. ADD INHERITANCE, VERSIONING AND GOVERNANCE

Users should always be able to see whether information is inherited or overridden.

For example:

Brand default
Project override
Reset to Brand default

AI should never silently overwrite established Brand information after re-analysing a website or connected source.

Instead, AIValance could show:

“We detected 3 potential changes to your Brand Profile.”

The user can then:

• Review
• Accept
• Reject

Brand Profile fields and important knowledge should also support states such as:

• Draft
• Reviewed
• Approved
• Locked

CORE INFORMATION MODEL

Brand = Who we are, what we offer, who we serve and how we position ourselves.

Voice = How we communicate.

Knowledge = What AI knows and can retrieve.

Assets = What AI can use when creating content.

Project = What we are currently trying to accomplish.

Whiteboard = Where we research, think, organise and develop ideas.

Output = What we ultimately produce.

The proposed hierarchy is therefore:

Workspace

Brand

Project

Whiteboard

Output

Knowledge, Assets, Audiences, Offers, Instructions and Integrations should operate as reusable resources around this hierarchy rather than being mixed together inside Brand

Guidance.

This would make Brand substantially more useful while creating a scalable foundation for AIValance Projects, RAG, content generation, collaboration and future workflows.

Please authenticate to join the conversation.

Upvoters
Status

In Review

Board
💡

Feature Request

Date

7 days ago

Author

Marcus

Subscribe to request

Get notified by email when there are changes.