Brand Enrichment

AI-powered brand profile generation. Upload a CSV of brand names or select brands directly from the table, and the AI researches each one — filling in the attributes you've configured such as descriptions, taglines, founding year, headquarters, SEO fields, and more.

Overview

Brand Enrichment builds rich, structured brand profiles for every brand in your list. You define the attributes you want filled in Settings → Enrichment → Brand Attributes, and the AI takes care of the rest — either by reading your linked product catalogue or by searching the web.

Key Features

• Fully Configurable Attributes — Define exactly what gets enriched and how
• Two Enrichment Modes — Grounded in your linked products, or powered by live web research
• SEO Auto-Formatting — Recognised SEO slugs (page title, meta description, slug) automatically enforce character limits and formatting rules
• Multiple Attribute Types — text, html, number, boolean, select, multiselect
• Attribute Sets — Reusable bundles to enrich only a chosen subset of attributes
• Bulk CSV Processing — Up to 2,000 brands per job, 8 MB max
• Direct from Brand Table — Enrich one or many brands without leaving the Brands pageTitle
• Blocked Domain Filtering — Exclude forums, generic content sites, and any domain you don't trust

What You Need Before You Start

1. Configure Your Enrichment Attributes

Open Settings → Enrichment → Brand Attributes and define the attributes you want the AI to fill for every brand. Each attribute has a name, a type, and a prompt that tells the AI what to research and what format to return.

Common attributes to start with:

• Brand Description
• Tagline
• Founded Year
• Headquarters
• Website URL
• Market Segment
• Industry
• SEO Page Title (page_title slug)
• Meta Description (meta_description slug)

You only do this once per organisation. After that, come back and edit, add, or remove attributes whenever your needs change.

2. (Optional) Link Products to Brands

When your brands have linked products in your catalogue, the AI uses those products as grounding context — it reads the product names and MPNs to understand what the brand actually sells, then generates more accurate and specific brand profiles.

A product counts as linked when its Enrich Brand field is set to the brand. The free-text brand name on the product doesn't count.

Brands with no linked products can still be enriched, but you must enable Use Websearch so the AI can research them on the web.

Hard limits:

• Maximum file size: 8 MB
• Maximum rows per job: 2,000
• File format: .csv

Step-by-Step: Enrich Brands via CSV

Follow these steps to enrich a brand list from a CSV file.

Step 1 — Open the Brand Enrichment page

Navigate to Brand → Enrich Brands, or click New Enrichment from the Brand Jobs page.

Step 2 — Configure attributes (first time only)

If you haven't already, open Settings → Enrichment → Brand Attributes and create the attributes you want the AI to fill. See the Settings section for details.

Step 3 — Prepare your CSV

Your file must:

• Be a .csv file under 8 MB
• Contain a name column with one brand name per row (also recognised: brand, brand_name, brandName)
• Have at most 2,000 rows

Duplicate brand names are automatically removed — case-insensitive. If the same brand appears twice, only the first occurrence is kept.

If you need a starting point, click Download Sample CSV on the Brand Enrichment page. The sample includes a name column with three example brands.

name
Nike
Adidas
Sony

Step 4 — Upload the CSV

On the Brand Enrichment page, drag and drop your file into the upload area, or click to browse.

After the file is selected, the upload box updates to show the file name and size.

Step 5 — Choose an Attribute Set

Select from the Attribute Set dropdown:

• All active attributes (default) — uses every active attribute configured for your organisation
• A custom set you've created under Settings → Enrichment → Brand Attribute Sets — restricts the run to a subset of attributes

Step 6 — Pick the AI Model

Step 7 — Configure Websearch

The Use Websearch toggle controls whether the AI is allowed to search the web for brand information.

Recommended: leave this on unless you specifically want to restrict enrichment to your catalogue data.

Step 8 — Include existing attributes

Include existing attributes is on by default. When a brand has already been enriched, those values are sent to the model as a starting point.

Turn this off when you want a fresh rewrite — for example after fixing bad specs — so the model does not keep the old values.

Step 9 — Click Enrich Brands

Click the Enrich Brands button. The CSV is parsed, deduplicated, and a background job is created. You are taken to the Jobs page to monitor progress.

Step-by-Step: Enrich Brands from the Brand Table

You can enrich one or more brands without leaving the Brands page — no CSV required.

Single Brand

1. Go to Brand in the main navigation
2. Find the brand you want to enrich
3. Click the ⋮ menu on the brand row (or open the brand detail page)
4. Select Enrich Brand
5. Configure the AI Model, Attribute Set, Websearch, and Include existing attributes options in the dialog
6. Click Confirm & Enrich

Multiple Brands

1. Go to Brand in the main navigation
2. Check the boxes next to the brands you want to enrich
3. Click Enrich in the selection toolbar
4. Configure options in the dialog — this applies to all selected brands
5. Click Confirm & Enrich

In both cases, you are taken to the Jobs page once the job is started.

Step-by-Step: Delete Brands from the Brand Table

You can delete one or more brands directly from the Brands page without navigating away.

Warning: Deletion is permanent. Deleted brands and all associated enrichment history cannot be recovered.

Single Brand

1. Go to Brand in the main navigation
2. Find the brand you want to delete
3. Click the ⋮ menu on the brand row
4. Select Delete Brand
5. Confirm the action in the confirmation dialog

Multiple Brands (Bulk Delete)

1. Go to Brand in the main navigation
2. Check the boxes next to the brands you want to delete
3. Click Delete in the selection toolbar
4. Review the list of brands to be deleted in the confirmation dialog
5. Click Confirm Delete to permanently remove all selected brands

Tip: Use the checkbox in the table header to select all brands on the current page at once. To delete across multiple pages, repeat the selection and delete per page.

How the AI Gets Its Information

The AI uses one of two modes per brand, chosen automatically based on your data.

Linked Products Mode

Used when the brand has at least one product linked to it in your catalogue.

The AI reads a random sample of up to 10 linked products (names and MPNs) and uses them as grounding context. It understands what the brand actually sells — product categories, market segment, positioning — and generates brand profile attributes that reflect this. No web search is made, which makes this mode faster and cheaper.

Web Search Mode

Used when the brand has no linked products and Use Websearch is enabled.

The AI performs up to 10 parallel Google searches (official website, About page, Wikipedia, press coverage), then reads the most relevant pages. It extracts structured signals — title tags, headings, JSON-LD, OpenGraph, and body text — to build an accurate brand profile. All source URLs are recorded in the job results.

Tip: Brands with linked products always use Linked Products mode — web search is never triggered for them even if the toggle is on. Web search only activates for brands that have no products in your catalogue.

Internal Linking

Once you configure Storefront URLs under Settings → Enrichment → General Settings, brand enrichment automatically weaves internal links into the generated HTML description:

• Up to 5 categories — the categories that brand has the most products in
• Up to 5 related products — products belonging to that brand

Categories are ranked by product count for that brand, so links point at the categories the brand actually sells into. Ranking is deterministic: re-running enrichment on the same brand produces the same links.

Links are placed only in html-type attributes. Your page title, meta description, meta keywords, and slug never receive anchors. Anchors are plain <a href="…">Anchor Text</a> — no rel, no target — and any URL the AI invents that wasn't supplied by the system is stripped from the result before it is saved.

Entities are linked only when their URL can be built from your base URL and pattern; anything that can't resolve is left unlinked rather than pointed at a broken URL. Leave the category or product pattern blank to skip that link type entirely.

See Storefront URLs & internal linking for the full configuration reference.

AI Model Comparison

Budget Model

• Best for: Large batches, routine enrichment runs, most brands
• Speed: Fastest
• Credit cost: Lowest
• Quality: Strong for well-known brands and brands with linked products

Pro Model

• Best for: Niche or hard-to-find brands, high-value brand profiles, complex attribute prompts
• Speed: Slightly slower
• Credit cost: Higher
• Quality: Best-in-class accuracy and writing quality

Start with Budget for most brands. Use Pro for flagship or commercially important brands where profile accuracy matters most.

Monitoring Your Job

After starting a job, you are taken to Brand → Jobs.

Jobs Table

The table shows all enrichment jobs for your organisation:

The table auto-refreshes every 15 seconds while any job is pending or processing. Click Stop to halt an active job — brands already enriched keep their results.

Job Statuses

Viewing Results

Job Items Panel

Click View on any job in the Jobs table to open the item-level detail view.

The panel shows:

• File name and job summary (total rows, success count, failed count)
• A table with one row per brand

The panel auto-refreshes every 5 seconds while the job is in progress.

Download CSV

Click Download CSV in the job items panel to export all results. Available once the job reaches completed, completed_with_errors, or stopped status.

The exported CSV contains:

• Your original input columns
• Enrichment Status
• Enrichment Error
• One column per enriched attribute (by attribute name)

Brand Enrichment History

Open any brand's detail page and go to the Enrichment History tab to see every enrichment run for that brand — with the model used, whether web search was active, and the full attribute results.

Brand Enrichment Settings

Open Settings → Enrichment from the main sidebar. Brand settings live on the Brand Attributes and Brand Attribute Sets tabs; country, blocked domains, and storefront URLs are shared with product and category enrichment on the General Settings and Blocked Domains tabs.

1. General Settings

The default country used when the AI researches brand information and verifies brand origins. Set this once to match your primary market.

This affects which regional sources and business registration databases the AI consults when searching for a brand.

This tab also holds Storefront URLs, which enable internal linking inside enriched brand descriptions — see Internal linking below.

Note: Country and blocked domains are organization-wide. Changing them here also affects product and category enrichment.

2. Brand Attributes

The full list of attributes the AI fills for every brand. This is where you customise what gets enriched.

From this tab you can:

• Create a new attribute with a name, type, and prompt
• Edit an existing attribute's name, type, prompt, or allowed options
• Reorder attributes — order determines column position in the exported CSV
• Toggle active / inactive — inactive attributes are skipped during enrichment without being deleted
• Delete an attribute permanently (existing job results are unaffected)

Each attribute has these fields:

Supported types:

Recognised SEO slugs — automatic formatting applied:

When you use one of these recognised slugs, the AI automatically applies the correct format rules without any extra configuration.

Writing a good prompt:

Write the prompt as if you were briefing a research analyst. Be specific about output format. For example:

"Find the year the brand was founded. Return a 4-digit year only (e.g. 1964). If you cannot verify the founding year from the brand's official website or Wikipedia, return null."

For select and multiselect types, mention the exact allowed options in the prompt — the AI is constrained to those values.

3. Brand Attribute Sets

Reusable bundles of attributes. By default, every enrichment run uses all active attributes. Attribute Sets let you create slimmer runs that fill only a chosen subset — useful for faster, cheaper partial updates.

1. Go to Settings → Enrichment → Brand Attribute Sets
2. Click Create Attribute Set
3. Give it a name and select the attributes to include
4. When running an enrichment job, pick your set from the Attribute Set dropdown

Once saved, the set appears in the Attribute Set dropdown on both the Brand Enrichment page and the Enrich Brand dialog.

4. Blocked Domains

A list of domains the AI should not consult during web research. Use it to keep the AI focused on authoritative sources and away from forums, aggregators, or sites with unreliable brand data.

Examples to block:

• Forum sites — reddit.com, quora.com
• Generic directories that surface inaccurate data
• Competitor sites you don't want the AI referencing

For any blocked domain, search results from that domain are filtered out before the AI reads them. Use Add Domain to insert one, and the delete icon to remove. Changes apply to all new jobs immediately.

Credits and Billing

Brand enrichment consumes credits per brand based on the AI model selected and whether web search is used.

• Linked Products mode is cheaper — no external web calls are made
• Web Search mode costs more — the AI makes multiple search and page-fetch calls per brand

If your credit balance reaches zero during a job, enrichment pauses and the job's error summary is updated. Top up your credits and start a new job to continue enriching the remaining brands.

See Credits & Billing for more on purchasing and managing credits.

Best Practices

Getting Better Results

1. Link products to brands first — linked products are the strongest signal the AI has; it knows exactly what the brand sells
2. Aim for 3–10 linked products per brand — a mix of product names gives broader context about categories and positioning
3. Use specific prompts — "Return the founding year as a 4-digit number" beats "What year was the brand founded?"
4. Use the Blocked Domains list — remove low-quality sites from web research to get cleaner data

Cost Optimisation

1. Start with Budget Model — use Pro only for your most commercially important brands
2. Use Attribute Sets — don't enrich columns you don't need on every run
3. Remove duplicates before upload — the system deduplicates, but clean input avoids unexpected row counts
4. Prefer Linked Products mode — it's faster and cheaper than web search

Large Catalogues

1. Split into multiple jobs if you have more than 2,000 brands
2. Monitor credit balance before starting large jobs — estimate roughly how many brands your current balance covers
3. Use Stop to pause a running job if needed — enriched brands are never lost

Troubleshooting

"No active enrichment attributes configured"

The Enrich button and confirm button are disabled because no attributes are active.

Fix: Go to Settings → Enrichment → Brand Attributes and add at least one active attribute before running a job.

Many Brands Marked as Failed

Possible causes:

• Brand has no linked products and websearch was disabled
• Network or API timeout during enrichment
• Credit balance ran out mid-job

Fixes: Enable Use Websearch and retry, check your credit balance and top up if needed, then start a new job for the failed brands.

Wrong or Generic Values

The AI returned a value but it seems inaccurate or too generic.

Fixes: Refine the attribute prompt to be more specific, switch from Budget Model to Pro Model, or add linked products so the AI has direct catalogue context.

Blank Values in the Export

The AI could not confidently find information for that field after research.

Fixes: Provide more linked products, enable web search, rewrite the attribute prompt with clearer instructions, or accept that the data is genuinely unavailable for that brand.

Job Stuck in Processing

Fix: The Jobs page refreshes every 15 seconds. If a job has been stuck for more than 10 minutes, click Stop and start a new job.

"Credit balance limit has been reached"

Enrichment stopped mid-job because the credit balance hit zero.

Fix: Go to Credits in the navigation, top up your balance, then start a new job for the brands that were not yet processed.

What This Feature Can and Cannot Do

Can do:

• Generate any attribute you configure — descriptions, taglines, SEO fields, founding year, HQ, website, industry, market segment, and more
• Use your actual product catalogue as grounding context (product names and MPNs)
• Research brands on the web when no linked products exist
• Process up to 2,000 brands per job
• Automatically enforce SEO character limits and slug formatting for recognised SEO slugs
• Enrich brands directly from the table without a CSV
• Add internal links to related categories and products when Storefront URLs are configured

Cannot do:

• Enrich a brand with no linked products when websearch is disabled
• Link to a category or product whose URL cannot be built from your storefront URL pattern
• Invent verifiable facts — if founding year or headquarters cannot be found, the field is returned as blank
• Create or manage brand records (only enriches existing brand data)

Next Steps

• Product Category Enrichment — Enrich category pages with SEO content
• Product Enrichment — Enrich individual products with attributes and specs
• Attribute Sets — Create reusable attribute bundles for targeted enrichment runs
• Credits & Billing — Purchase and manage credits
• API Keys — Access the platform programmatically