Skip to content
P2Lab AI

Your first prompt

This page walks through the simplest possible workflow: generating SEO metadata for a single product. Once you’ve done it once, every other feature of P2Lab AI follows the same pattern — open a Shopware form, click an AI assistant button, review the suggestion, apply.

What you need before starting

Section titled “What you need before starting”
  • The plugin is installed and activated (Installation)
  • You have a connection with at least one chat model and the model is set as default for generate_seo.product (Connecting AI)
  • At least one product in your catalogue with a description

Step 1 — Open a product

Section titled “Step 1 — Open a product”
  1. Go to Catalogue → Products and open any product.
  2. Switch to the SEO tab.

In the SEO card header you’ll see a new AI assistant button.

Step 2 — Open the AI modal

Section titled “Step 2 — Open the AI modal”

Click AI assistant. The modal Generate SEO opens with three fields:

  • Profile — saved inline profile, optional (Inline profiles)
  • AI connection — pre-filled with the default for generate_seo.product; change if you want
  • Additional instructions (optional) — free-text, e.g. “Mention the brand at the end of the title”

Step 3 — Run the generation

Section titled “Step 3 — Run the generation”

Click Generate. The plugin builds the prompt from the active generate_seo.product template, sends it to OpenAI and shows the response in the same modal:

  • Meta title
  • Meta description
  • Keywords

Each field is editable. You’ll also see a Reasoning block explaining why the model chose what it chose.

Step 4 — Apply

Section titled “Step 4 — Apply”

Click Apply to form to copy the values into the SEO form. Save the product as you normally would in Shopware.

That’s it — your first AI-assisted edit is live.

What just happened under the hood

Section titled “What just happened under the hood”
  1. The plugin loaded the generate_seo.product template — either the default one shipped with the plugin or a fork that matches this product (see Forks).
  2. It built a system prompt and a user prompt by substituting placeholders (the product name, description, custom instructions).
  3. It called OpenAI through your AI connection, recorded token usage and estimated cost.
  4. It returned the structured response straight into the modal — no suggestion row was persisted because you ran a one-shot inline operation. Bulk jobs work differently (see Batch jobs).

Where to go next

Section titled “Where to go next”
  • Recommended next step: Tailor prompts to your shop — refine the default prompts on a real product before you run anything in bulk. The few minutes spent here pay back on every subsequent generation.
  • Explore the feature pages to see every operation P2Lab AI supports.
  • Learn how prompts are structured in Prompts overview.
  • Read about costs and statistics before running a bulk job.