Campaign Brief Analysis

Campaign Brief Analysis lets you upload a campaign brief document -- a PDF, image, Word document, or PowerPoint -- and have AI automatically extract audience targeting recommendations. The AI reads the brief, identifies the brand, objective, target audience, channels, and KPIs, then recommends 3 to 5 audience topics that can be classified into your taxonomy.

This is ideal for media planners and audience strategists who receive campaign briefs from clients and need to quickly translate them into actionable audience segments.

Overview

The campaign brief workflow fits naturally into the classify chat interface. Instead of typing a topic name, you upload a file. The AI analyzes the document and presents structured recommendations that you can select and classify.

Upload brief → AI analysis → Review recommendations → Select topic → Classify → Add to Library

Supported File Formats

AudienceGPT accepts a range of document and image formats for brief analysis:

Format	MIME Type	Content Type	Notes
PDF	application/pdf	Document	Most common format for campaign briefs
Word (.docx)	application/vnd.openxmlformats-officedocument.wordprocessingml.document	Document	Microsoft Word documents
PowerPoint (.pptx)	application/vnd.openxmlformats-officedocument.presentationml.presentation	Document	Slide decks and presentations
PNG	image/png	Image	Screenshots or scanned briefs
JPEG	image/jpeg	Image	Photos of printed briefs
GIF	image/gif	Image	Animated or static images
WebP	image/webp	Image	Modern web image format
Plain Text	text/plain	Text	Simple text briefs

warning

Maximum file size: 10 MB. Files larger than 10 MB will be rejected. If your brief exceeds this limit, consider splitting it into smaller sections or converting high-resolution images to a lower resolution before uploading.

Upload and Analysis Flow

Step 1: Upload Your Brief

On the Classify page (/classify), click the file upload button (paperclip icon) in the chat input area. Select your campaign brief file from your computer.

Alternatively, you can drag and drop the file directly onto the chat input area.

The file is uploaded and sent to the AI for analysis. You will see a processing indicator while the AI reads and interprets the document.

Step 2: AI Analysis

The AI analyzes your brief using Claude Sonnet 4.6 and extracts the following structured information:

Field	Description
Brand	The advertiser or brand name from the brief
Objective	The campaign's primary goal (awareness, conversions, etc.)
Target Audience	The audience description from the brief
Channels	Media channels mentioned (display, social, CTV, etc.)
KPIs	Key performance indicators (CTR, ROAS, reach, etc.)
Budget	Budget information if available in the brief
Timeline	Campaign dates or flight periods

Based on this analysis, the AI recommends 3 to 5 audience topics that could be built from the brief.

Step 3: Review AI Recommendations

The AI presents its recommendations in the chat as a list of topics. Each recommendation includes:

• Topic name -- A behavioral intent topic (what people are actively researching or doing online)
• Rationale -- Why this topic is relevant to the campaign brief
• Segment type -- The suggested segment type (B2B, B2C, etc.)
• Objective -- How this audience supports the campaign objective

tip

The AI generates a mix of specificity levels: some topics are tightly focused on the brand/product mentioned in the brief, while others are broader category-level topics that capture the wider audience opportunity.

Example recommendations from a Nike running shoe campaign brief:

1. Nike Running Shoes -- Direct brand interest, people researching Nike running footwear
2. Marathon Training -- Functional interest, runners preparing for distance events
3. Athletic Footwear -- Category-level interest, general running shoe shoppers
4. Fitness Wearables -- Adjacent interest, people tracking running performance
5. Sports Nutrition -- Adjacent interest, runners researching nutrition and recovery

Step 4: Select and Classify

Pick the topic you want to classify by clicking on it or telling the AI which recommendation to proceed with (e.g., "Let's classify Nike Running Shoes"). This moves you into the standard classification flow:

1. The AI may ask follow-up questions for additional context
2. You confirm the topic details
3. The 7-layer classification engine processes the topic
4. You review results and add to your Library

You can classify multiple recommendations by returning to the brief results after each classification and selecting the next topic.

info

Each recommended topic is classified independently. You can choose different classification modes (AI-powered vs. rule-based) for different topics if you wish.

Best Practices for Brief Formatting

To get the best recommendations from the AI, structure your campaign briefs to include:

Include These Elements

• Clear brand/product identification -- State the advertiser name and specific products or services being promoted
• Campaign objective -- What the campaign aims to achieve (awareness, consideration, conversions)
• Target audience description -- Who the campaign is trying to reach, including demographics, interests, and behaviors
• Media channels -- Which platforms and channels will be used
• KPIs and success metrics -- How campaign performance will be measured
• Timeline and budget -- Flight dates and budget ranges help the AI calibrate recommendation specificity

Formatting Tips

• Use clear headings and sections -- Structured briefs with labeled sections (Objective, Audience, Channels) produce better extraction results
• Avoid excessive design elements -- For image uploads, ensure text is readable and not obscured by heavy graphic design
• Include specific audience descriptions -- "Health-conscious women ages 25-44 interested in premium skincare" is better than "target demographics"
• Mention competitor brands -- If the brief references competitors, the AI can recommend competitive conquest audience topics

What to Avoid

• Briefs with only visual content and no readable text (the AI needs text to analyze)
• Password-protected or encrypted PDFs
• Files with primarily tabular data and no narrative context
• Extremely large briefs (more than 20--30 pages) -- consider uploading the most relevant sections

File Size and Processing Limits

Parameter	Limit
Maximum file size	10 MB
Supported formats	PDF, DOCX, PPTX, PNG, JPEG, GIF, WebP, TXT
Recommendations per brief	3 to 5 topics
Processing time	Typically 5--15 seconds
Model used	Claude Sonnet 4.6

info

Campaign brief analysis consumes API credits. Each analysis uses the same model and pricing as AI-powered classification. The cost is based on the size of the uploaded document (input tokens) plus the structured response (output tokens).

Next Steps

• Matrix Generation -- Create combinatorial taxonomies from dimensional data
• CSV Import -- Bulk import topics from spreadsheet files
• Library Management -- Manage and organize your classified topics
• Classification Deep Dive -- Understand the full 7-layer classification system