Documentation

    AI Content

    Generate optimized product descriptions, collection texts, alt tags and blog posts.

    1Generate product descriptions

    Pshly.ai generates unique, SEO-optimized product descriptions based on:

    • Product title, type and tags
    • Product images (via AI image analysis)
    • Variants (sizes, colors, materials)
    • Your custom prompt and writing style
    • Relevant keywords from Google Search Console
    1. Open a product in the Products tab.
    2. Click "Generate".
    3. Review the result.
    4. Click "Approve" and "Push to Shopify".
    Generate product descriptions

    Real-world example

    From "Box brown large" to "Corrugated Shipping Box 40×30×20 cm — Brown, Single Wall". At G&F Verpakkingen 500+ descriptions had been written by different people over the years: one long and technical, the next a single line.

    After the rewrite every title contains the product type, size and material, with the SEO title (≤60 chars) and meta description (≤155 chars) included automatically in one consistent style.


    2Generate collection texts

    Collection texts help search engines understand your category pages. Pshly.ai supports a dual text structure:

    • Top text — The Shopify body_html, visible at the top.
    • Bottom text — Stored as metafield, ideal for extra SEO content at the bottom.

    Choose between 'Top text only', 'Bottom text only' or 'Both' in settings.

    Generate collection texts

    Real-world example

    34 collections, each with a top text and a hidden SEO bottom text. The "Single wall shipping boxes" collection for G&F Verpakkingen got a short top text for visitors and a more extensive bottom text with dimensions and use cases for search engines.

    That split kept the page clean while the SEO text still contributed to the +34% rise in organic clicks.


    3Generate Image Alt Tags

    Alt tags describe your product images for search engines and accessibility.

    1. Go to the Alt Tags tab.
    2. See all product images with their current alt text.
    3. Click "Generate" for AI descriptions based on image analysis.
    4. Review suggestions and push to Shopify.

    Bulk generation is also available.

    Generate Image Alt Tags

    Real-world example

    512 product images, 0 still named "IMG_2041.jpg". For G&F Verpakkingen, the AI generated an alt text for every photo like "Corrugated shipping box 30×20×15 cm, brown, double wall, open top view".

    Besides accessibility, this delivered a measurable traffic boost via Google Images for search terms like "double wall shipping box".


    4Generate blog posts

    Generate complete blog posts to strengthen your SEO authority.

    1. Go to Blog and click "New article".
    2. Enter a topic, focus keyword, audience and tone.
    3. Click "Generate" — the AI writes a full article with SEO metadata.
    4. Edit and push to your Shopify blog.

    Use the inspiration tool for automatic topic suggestions.

    Generate blog posts

    Real-world example

    "Single wall or double wall: which shipping box should you choose?" For G&F Verpakkingen, the AI generated this article based on the focus keyword "corrugated shipping box", including a comparison table and weight guidance per box category.

    Within two months the article attracted new organic traffic that flowed through to the product pages via internal links.


    5AI Copywriting style & tone

    Customize style and tone via Settings → Writing Style:

    • Writing style: Professional, casual, luxury, playful, technical or informative.
    • Custom prompts: Your own instructions for product and collection texts.
    • Company info: Context about your brand and USPs.

    The AI supports 4 languages (NL, EN, DE, FR). Content language is independent of interface language.

    AI Copywriting style & tone

    Real-world example

    "Technical" writing style set for G&F Verpakkingen. With a custom prompt ('always mention the maximum load weight per box'), all 500+ descriptions were rewritten in the same businesslike, factual tone.

    This put an end to the mix of overly technical and too-short descriptions that had built up over the years.


    6A/B variants for products

    Generate A/B variants — two versions of the same text.

    • The AI uses higher variation for more diversity.
    • Compare both versions side by side.
    • Choose the version that fits best.
    A/B variants for products

    Real-world example

    Two titles for the same shipping box. G&F Verpakkingen tested "Shipping Box 40×30×20 cm — Single Wall, Brown" against "Corrugated Box 40×30×20 cm for Shipping, Single Wall" and picked the second variant based on click behaviour.

    This let the team decide with data which phrasing better matched customer search intent, instead of relying on gut feeling.


    7Bulk content generation

    Bulk generation creates content for multiple items at once.

    1. Select products/collections (or "Select all").
    2. Click "Bulk generate".
    3. Track progress in the progress bar.

    Uses credits per item. Processing continues in the background.

    Bulk content generation

    Real-world example

    18 fields in a single run. In the bulk modal every field is enabled (material, dimensions, flute type, weight, units per box, colour, application, …) plus the extra actions "Suggest Shopify category" and "Suggest internal links".

    Across 500 products that produced an average of 15 of 18 filled fields per product — so size and material filters work in the storefront right away.


    8Optimising URLs (handles)

    Bulk content generation can also optimise the product URL (handle) based on the new AI title.

    • Toggle Optimise URL in the bulk modal; the default lives in Settings.
    • For draft products the new slug is applied directly.
    • For active products Pshly automatically creates a 301 redirect in Shopify so no SEO value is lost.
    • The progress bar badges products whose URL changed.
    Optimising URLs (handles)

    Real-world example

    From /products/product-4471 to /products/shipping-box-40x30x20-single-wall. During the bulk run for G&F Verpakkingen, 480 URLs were automatically optimised based on the new title, with 301 redirects from the old handle.

    This prevented existing backlinks and Google rankings from being lost while rewriting the catalog.


    9Title quality & filler words

    Product names should be concrete, not full of marketing words. Pshly filters automatically:

    • Filler words such as "elegant", "stylish" or "luxurious" are removed from titles and SEO titles, and forbidden in the AI prompt.
    • Repeated words (e.g. "Earrings Earrings") are deduplicated.
    • The structure stays Name – Type – Material, keeping titles unique and searchable.
    Title quality & filler words

    Real-world example

    "Beautiful stylish shipping box" became "Shipping Box 40×30×20 cm, single wall". For G&F Verpakkingen, Pshly automatically filtered filler words out of 500+ titles written by different employees.

    The result: titles that are scannable on size and material, exactly what B2B buyers search for.


    10Matching products (AI)

    The Matching products job uses AI to find complementary products — not more of the same.

    • The AI judges the product image on style, colour and material, not just the name.
    • Earrings therefore suggest a bracelet or necklace, not another pair of earrings.
    • The maximum number of suggestions per product is configurable (default 2, range 1–10).
    • The job runs server-side and ends in a review modal.
    Matching products (AI)

    Real-world example

    A 40×30×20 cm shipping box was suggested with matching filler material and tape. For G&F Verpakkingen, the AI assessed product photos on size and use case, not just the title.

    This surfaced relevant add-on products during the customer's visit, which raised the average order value.


    11Collection suggestions (AI)

    The Collection suggestions job places products into the right Shopify collections.

    • Brand detection based on collection titles.
    • With more than 60 collections, suggestions are processed in chunks for higher accuracy.
    • Post-validation checks brand, audience and product type.
    • Min. score and max. suggestions control strictness and breadth.
    Collection suggestions (AI)

    Real-world example

    34 collection pages from 512 products. With min. score 70 and max. 3 collections per product random matches drop out; type and application checks prevent a pallet cover from ending up in "Shipping boxes".

    Above 60 collections the AI splits the analysis into chunks, which noticeably improves accuracy.