Products Overview

# Mastering the Folki Platform: Your Command Center for AI-Ready Commerce

Last Updated: January 26, 2026

Reading Time: 8 Minutes

The Folki Platform is your central hub for managing the transition from "Traditional SEO" to "Answer Engine Optimization" (AEO). It is not just a viewing portal; it is a mission control center designed to give you a "God's Eye View" of your product catalog's data integrity, enrichment status, and AI readiness.

This comprehensive guide will walk you through every widget, graph, and tool available in our platform, ensuring you can leverage the full power of Folki's agentic workforce.

1. The Overview: Pulse of Your Catalog

Upon logging in, the first screen you encounter is the Overview. This view is designed to answer one critical question in milliseconds: *"Is my catalog ready for the AI era?"*

The Enrichment Score (Global Health)

At the top left, you will see a large, color-coded metric: the Enrichment Score.

What is it?

The Enrichment Score (0-100) is a weighted aggregate of three factors across your entire SKU base:

1. Completeness: What percentage of standard attributes (Weight, Dimensions, Material) are filled?

2. Depth: Do products have niche, category-specific attributes (e.g., "Sensor Type" for cameras, "Thread Count" for bedding)?

3. Verification: What percentage of these data points are backed by an authoritative citation (Manufacturer Datasheet)?

Interpreting the Score:

• 🟢 80-100 (Excellent): Your catalog is highly structured. AI agents like ChatGPT and Google Gemini will prioritize your products as "High Confidence" answers.
• 🟡 50-79 (Moderate): Your basic data is there, but you lack the technical depth required for specific recommendation queries.
• 🔴 0-49 (Critical): Your products are "Invisible" to AI. They lack sufficient structured data to be compared or recommended.

Real-Time Activity Feed

To the right of the score is the Agent Activity Stream. This is not a static log; it is a real-time window into the minds of your autonomous agents.

What you will see:

• Discovery Events: *"Agent-04 found a matching PDF manual for SKU-1920 at sony.com"*
• Extraction Events: *"Extracted 14 technical attributes for 'Ergonomic Office Chair'"*
• Conflict Flags: *"Detected conflicting voltage data for SKU-990. Requested human review."*

Pro Tip: Watch this feed during a bulk import. It gives you immediate feedback on the "Match Rate" of your products against external databases.

2. Product Manager: The Data Workbench

Clicking Products in the sidebar opens the Product Manager. This is where you will spend 80% of your time.

Advanced Filtering Strategy

When you have 10,000 SKUs, you cannot browse page by page. Use the Smart Filters to identify "Low Hanging Fruit."

• Filter: "High Revenue, Low Health"
• *Strategy:* Sort your products by sales volume (imported from Shopify) and filter by Enrichment Score < 50.
• *Action:* Prioritize these products for "Deep Research" mode immediately. These are your best-sellers that are at risk of being dropped by AI search engines.
• Filter: "Missing Critical Specs"
• *Strategy:* Isolate products that are missing the #1 ranking factor for their category (e.g., missing "Energy Efficiency Class" for appliances).
• *Action:* Bulk-select these and run a targeted enrichment focusing *only* on that missing field.

The "Confidence" Column

In the list view, every product has a Confidence badge.

• Verified (Blue Check): Data matches the official manufacturer's domain.
• Likely (Green Dot): Data matches 3+ major retailer listings (Amazon, BestBuy, Walmart).
• Uncertain (Yellow Dot): Data found on a general blog or forum.
• Unknown (Grey): No data found.

3. The Single Product View (detail mode)

Clicking on any SKU takes you to the Atomic Data View.

Left Panel: The Input

This shows the raw data coming from your store (Shopify/WooCommerce). Usually, this is sparse—a title, a price, and a blob of HTML description.

Middle Panel: The "Agent Thought Process"

This unique view allows you to audit the AI's logic.

• Search Path: See exactly which Google queries the agent used.
• *Query:* `"Sony WH-1000XM5 official specs pdf site:sony.com"`
• Source Evaluation: See which pages were rejected and why.
• *Rejected:* `techradar.com` (Reason: Third-party review, not primary source).
• Extraction Logic: See the specific text snippet used to derive the value.

Right Panel: The Knowledge Graph

This is the final output—the structured JSON-LD object.

• Attribute: The standardized name (e.g., `noiseCancellationType`).
• Value: The normalized value (e.g., `Active (ANC)`).
• Unit: The SI unit (e.g., `dB`).

4. Analytics & ROI Tracking

The Analytics tab is where you prove the value of AEO to your stakeholders.

Citation Velocity Graph

This chart tracks the number of times your products' structured data has been accessed by known AI crawlers (GPTBot, Google-Extended, Applebot).

• Trend: A rising line indicates that your products are being indexed into the Large Language Models' training sets or RAG (Retrieval-Augmented Generation) databases.

Attribute Density vs. Competitors

Folki compares your catalog against a database of "Category Averages."

• *Example:* "Your 'Laptops' category has an average of 42 attributes per product. The industry average is 12. You are in the top 1% for Data Density."

5. Team & Permissions

Folki is a collaborative tool.

Roles

• Admin: Full access to billing, API keys, and deletion.
• Editor: Can run enrichments and approve data. Cannot change billing.
• Viewer: Read-only access to analytics and dashboards. (Perfect for CMOs or Clients).

Audit Logs

Every change to a product is logged. If an intern accidentally approves bad data, you can roll back to the previous version with one click in the History tab.

Frequently Asked Questions (FAQ)

What do I do if the "Enrichment Score" is stuck at 0?

This usually means the agents are unable to match your product titles to real-world items.

1. Check your titles. Are they generic? (e.g., "Blue Shirt").

2. Add a Brand Name or MPN (Manufacturer Part Number) to the title.

3. Re-run the "match" process.

Can I export the data for use in Excel?

Yes. In the Products view, select the items and click Export > CSV. You can choose to export the "Raw Data," "Enriched Data," or a "Diff Report."

How often does the platform update?

• Agent Activity: Real-time (Socket connected).
• Analytics: Updated every 24 hours.
• Store Sync Status: Immediate.

Glossary of Platform Terms

• RAG (Retrieval-Augmented Generation): The process AI uses to fetch your specific data to answer a user query.
• Hallucination Risk: A metric indicating the likelihood of an AI inventing false details about your product due to missing data.
• Knowledge Graph: A network of entities and relationships that describes your product in a way machines understand.

By mastering the Folki Platform, you are not just "fixing data"—you are building the infrastructure for the next decade of commerce.