> This page is for Developer.

> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.bigcommerce.com/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.bigcommerce.com/_mcp/server.

# How your inputs are used

> How purpose, brand voice, keywords, and additional context become per-field instructions, and why vague input weakens every field at once.

The create screen asks for four inputs. They are not passed to the generator as written — they are turned into specific instructions for each of the seven fields.

These instructions are applied to the product data Catalog sends for each product: its first five images, name, description, and SEO keywords. See [What Catalog sends to the enrichment service](/developer/docs/ai-tools/catalog-enrichment/before-you-enrich/source-data#what-catalog-sends-to-the-enrichment-service).

![Diagram: purpose, brand voice, search keywords, and additional context are applied to every field, producing per-field instructions — for example a description written warm and benefit-led, an SEO meta description that is keyword-aware and within length, and conversational Q and A giving direct citable answers.](/_fern-files/bigcommerce.docs.buildwithfern.com/217e2a880d5f14e99242275cda03f4bbc650e23c4af26650803a738a4d825e4a/docs/assets/catalog-enrichment/inputs-to-rules.svg)

## Why this matters for how you write them

A description and an SEO meta description have different jobs, lengths, and readers. Your inputs have to serve both. "Friendly" gives every field almost nothing; "warm and plain-spoken, never uses exclamation points, says sofa not couch" gives each field something concrete to apply.

The practical consequence: vague input does not produce slightly worse copy in one place. It weakens all seven field instructions at once, which is why a first batch that reads flat is usually an input problem rather than a product problem.

## The four inputs

| Input                  | What it controls                           | Covered in                                                                                                    |
| ---------------------- | ------------------------------------------ | ------------------------------------------------------------------------------------------------------------- |
| **Purpose**            | Which outcome each field is written toward | [Purpose, keywords, and context](/developer/docs/ai-tools/catalog-enrichment/inputs/purpose-keywords-context) |
| **Brand voice**        | How the content sounds                     | [Writing brand guidelines](/developer/docs/ai-tools/catalog-enrichment/inputs/brand-guidelines)               |
| **Search keywords**    | Terms worked into copy where they fit      | [Purpose, keywords, and context](/developer/docs/ai-tools/catalog-enrichment/inputs/purpose-keywords-context) |
| **Additional context** | Anything specific to this batch            | [Purpose, keywords, and context](/developer/docs/ai-tools/catalog-enrichment/inputs/purpose-keywords-context) |

## Reuse and iteration

You can repeat an enrichment on the same products with new brand voice and context. That makes inputs cheap to iterate on a small batch and expensive to get wrong on a large one — tune on 25 products before committing 250.