Shopify merchants can now add Constructor’s Product Insights Agent (PIA) directly to product detail pages using the Shopify theme editor.

This makes it easier for Shopify merchants to bring product-aware Q&A into the shopping experience without requiring a custom frontend implementation. With PIA available directly in the Shopify theme editor, teams can quickly add AI-generated answers to PDPs and help shoppers get the information they need while they’re evaluating a product.

For merchandisers and customer product teams, this creates a more accessible path to launching agentic shopping experiences — reducing setup effort while giving shoppers more confidence at a high-intent moment in the journey.

What’s enabled

Easy setup: Merchants can add a PIA Q&A widget to Shopify product detail pages through Constructor’s Shopify theme app extension and configure it directly in the Shopify theme editor — without building a custom frontend integration.

Improved product discovery: Once enabled, PIA uses product page context to answer shopper questions directly on the PDP, helping shoppers get product-specific support while they’re evaluating an item and make more confident purchase decisions.

Where to find it

Once enabled, merchants can add the PIA block to a product detail page template and configure it directly in Shopify.

Documentation

Shopify UI documentation

If you have any questions or feedback, please connect with your Customer Success Manager.

S3 support for catalog data ingestion

Constructor can now fetch catalog data directly from Amazon S3, giving customers another flexible format for recurring catalog data ingestion.

This is especially useful for technical and commerce teams that want to connect existing data storage workflows to Constructor and keep product data flowing on a reliable cadence.

What this flexibility unlocks

This helps you:

• Use existing workflows: Share catalog files from an existing S3 bucket without changing established data export workflows.
• Set up faster: Connect to an existing bucket and reduce extra integration work.
• Support more ingestion patterns: Add S3 as another flexible format for catalog data ingestion.
• Keep catalogs updated reliably: Retrieve product data on the cadence that works best for your team.

Documentation

Fetching files via S3

If you have any questions or feedback, please connect with your Customer Success Manager.

Planning seasonal promotions, product launches, and other time-sensitive merchandising campaigns shouldn’t require scheduling every rule individually.

With campaign scheduling, you can now set a start time, an end time, or both for an entire campaign. The campaign schedule acts as the overall window for every rule inside it, helping you prepare launches in advance and keep campaign timing consistent.

With this update

• Schedule campaigns from one place: Set a start time, an end time, or both directly at the campaign level.
• Plan launches ahead: Prepare seasonal promotions, sales, product drops, and other time-sensitive campaigns in advance.
• Keep rule schedules aligned: Rules can still have their own schedules, as long as they stay within the campaign’s overall window.
• Prevent scheduling conflicts: If a rule falls outside the campaign window, Constructor highlights the conflict so you can adjust the affected schedule.
• See scheduling status at a glance: Quickly identify scheduled campaigns directly from the campaign list.

How to get started

1. Go to Searchandizing → Campaigns.
2. Create a new campaign or open an existing one.
3. Set a start date, an end date, or both for the campaign.
4. If needed, add individual schedules to rules within the campaign. Rule schedules must remain within the campaign’s overall window.
5. Review the campaign list to see its scheduling status at a glance.

If you have any questions or feedback, please connect with your Customer Success Manager or contact us at support@constructor.io.

Turning a merchandising idea into a rule no longer has to mean configuring every condition by hand.

With the Merchant Intelligence Agent (MIA), you can now describe what you want to achieve in plain language — such as “put this week’s new arrivals at the top” or “push items with broken size availability down.” MIA translates your intent into a suggested single-page Searchandizing rule for you to review, adjust, and publish.

With this update

• Create rules using natural language: Describe your merchandising goal instead of manually selecting every rule type, attribute, and condition.
• Start from a rule preview: Open MIA from any rule preview page to create a new rule for that page or review the rules already affecting it.
• Start from anywhere in MIA: Ask MIA to create a single-page Search or Browse rule, and it will take you to the relevant preview and prepare a draft.
• Get a suggested configuration: MIA translates your intent into an appropriate rule setup, including the rule type, attribute conditions, targeting, and timeframe.
• Make better use of your merchandising toolkit: MIA can suggest capabilities that fit your goal, including tools you may not use regularly.
• Stay in control: MIA prepares the configuration, but it does not publish rules automatically. You can preview, adjust, and approve every change before it goes live.
• Only rule creation: MIA currently supports creating rules inside the rule preview but does not support full campaign creation.

How to get started

1. Open a rule preview page in Searchandizing.
2. Open the MIA chat and describe what you want to achieve.
3. Review the suggested rule configuration in the preview.
4. Adjust the configuration if needed, then save and publish it.

You can also start from the MIA chat elsewhere in the dashboard. Ask MIA to create a single-page Search or Browse rule, and it will guide you to the appropriate preview.

This capability is available to customers with MIA enabled.

If you have any questions or feedback, please connect with your Customer Success Manager or contact us at support@constructor.io

We always strive to give digital merchandisers the full transparency into what shopping experience they provide to their shoppers with Constructor. That’s why we are super excited to announce a major update to facet controls in Interact.

You can now validate what shoppers see (visible facets and their order), preview alternate ordering when hidden facets are included, and — if you participate in the Merchant Intelligence Agent (MIA) beta program — get an in-context explanation of why a facet is ranked where it is.

With this update

• See facets exactly as shoppers see them: Interact now shows visible facets only, in the same order they appear on the storefront.
• Find hidden facets in a separate section below to work with the results.
• Preview ordering when hidden facets are included: Toggle Include hidden facets in order to recalculate and preview facet order when hidden facets are requested via API.
• Move faster with better in-context controls:
  • Search facet groups and options (including hidden facets).
  • See selected facets in the header and remove quickly.
  • Jump directly to Facet Merchandising (and from there to global facet configuration) when you need to adjust rules.
• (If MIA is enabled) Explainability for facet order: Hover over a facet group or option, and you’ll notice the new Sparks icon on the right from it. Move you cursor over the icon to find a natural-language summary of what’s influencing facet position position (e.g., performance/rules/configurations/visibility), with options to copy or continue the conversation in chat.

How to get started

1. Go to Workspace → Interact and run a Search or Browse that shows a meaningful facets list.
2. In the Facets panel:
   • Use search to find facet groups/options (including hidden facets).
   • Scroll to review the Hidden facets section (shown separately by default).
3. Turn on Include hidden facets in order to preview interleaved ordering.
4. Click the green pencil button to open Facet Merchandising and adjust facet ordering as needed.
5. (If MIA is enabled) Hover a facet row and click on the Sparks icon to see why it’s ranked where it is.

If you have any questions, please connect with your Customer Success Manager or contact us through support@constructor.io.

What’s new

You can now use Collections in Offsite Discovery, making it easy to include curated (and optionally personalized) collection results in offsite channels like email, SMS, and push.

• Choose a Collection as a content piece when building a notification campaign
• Combine collections and recommendation pods in the same campaign (up to 4 content pieces total — e.g., 4 collections, 4 pods, or 2 of each)
• Our reasoning engine automatically picks the best products within your chosen collection for every shopper
• Preview a campaign before you send it

💡 Why it matters

This unlocks new marketing and merchandising use cases beyond recommendations — teams can promote curated assortments (e.g., seasonal picks, sale edits, best-of lists) through offsite campaigns while still benefiting from Constructor’s ordering and personalization when a user identifier is available.

📍 Where to find it

In Offsite Discovery / Notification Campaigns, choose a Collection as the content source (instead of a recommendation pod) and use the generated snippet/API to render product cards in your email service provider (ESP).

Learn more / questions

• Newly created collections may take ~10–15 minutes to show results in the preview.
• You don’t need to build a new collection PLP on your site — campaigns generate PDP links and redirect shoppers there. (If you want a “View all” experience linking to a collection-specific PLP, you’ll need to create that page on your end.)
• If you have any questions or feedback, please connect with your Customer Success Manager or contact us through support@constructor.io.

We’ve added badge support to Recommendations, so shoppers can see the same high-signal labels directly on recommended items — creating a more connected experience across Search, Browse, Product details pages and Recommendations.

Why this is important

Badges are small visual labels that highlight key information about an item  and can encourage purchasing by making it easier for shoppers to decide what to click next.

What it unlocks for you

• A more consistent shopper journey: the same merchandising cues can show up across onsite touchpoints, so shoppers don’t lose decision signals when they move from Search/Browse into recommendation modules.
• Higher engagement with recommendations modules: badges help shoppers quickly identify standout items, which can translate into stronger interaction with recommendations.
• A scalable merchandising lever: once badges are configured, they update automatically over time.

If you’re interested in using badges in Recommendations, please connect with your Customer Success Manager — they can guide rollout and best practices.

We’ve released Image Search v2, which adds optional automatic image segmentation to Image Search. This lets shoppers upload a scene photo (for example, an outfit or a room) and get results for each detected item — instead of treating the whole image as a single query.

What’s enabled

• A new endpoint: POST >
• Optional segmentation mode (use_segmentation=true) that detects multiple objects in an image and searches each one
• Optional per-segment metadata in the response (return_segments=true) so clients can render selectable regions (bounding boxes / masks)
• Segment crop options: search_crop_type=bbox (default) or search_crop_type=segment (masked crop)

What it unlocks for you

• Multi-item discovery: Let shoppers go from “this whole photo” to “show me results for the jeans / shirt / shoes”.
• Better relevance on complex images: Reduced noise when the upload contains multiple products or background clutter.
• Stronger demo / evaluation flows: Make “shop from photo” experiences more interactive by letting users select what they mean within the image.

API docs

• Image Search v2 reference: https://docs.constructor.com/reference/v2-image-search-retrieve-search-results

If you have any questions or feedback, please connect with your Customer Success Manager.

We’ve enabled Product Details API endpoints that let customers fetch enriched information for items and variants, so PDPs can be powered by the same enrichment layer as the rest of the Constructor experience.

This is especially important for PDPs as it’s where shoppers make high-intent decisions. Having enriched signals available (instead of only on listing pages) makes the experience more consistent and easier to trust.

What’s enabled

• Retrieve enriched data by item
• Retrieve enriched data by variation
• Return enriched signals that customers can use to improve PDP UX (including badges)

What it unlocks for you

• Consistency: Badges and other enriched signals don’t disappear after a shopper clicks into a product.
• Cleaner integrations: A dedicated PDP-friendly retrieval path, instead of repurposing PLP responses.
• Better PDP storytelling: We’re especially excited about serving badges on PDP, so key signals can show up exactly where purchase decisions happen.

API docs

• Retrieve item product details: https://docs.constructor.com/reference/v1-product-details-get-items
• Retrieve variation product details: https://docs.constructor.com/reference/v1-product-details-get-variations

If you have any questions or feedback, please connect with your Customer Success Manager.

You can now paginate through Items and Variations with cursor-based pagination for more reliable, scalable listing - especially for large catalogs.

What’s new

• GET /v2/items and GET /v2/variations now support cursor-based pagination via next_page_cursor.
• Each response may return next_page_cursor; pass it in the next request to fetch the next page of results.

Why it matters

• More reliable pagination for large result sets (reduces offset-related gaps/duplication).
• Easier to build long-running or resumable export/audit workflows.

Common customer use cases

• Catalog audits / backfills: Iterate through the full catalog safely, without relying on offset math.
• Resumable scripts: Store the last cursor to resume processing after interruptions.
• Large-catalog integrations: Reduce risk of missed/duplicated entities when the underlying catalog is changing.

How to use it (high level)

• Make a request and read next_page_cursor from the response.
• Repeat requests with next_page_cursor until it’s empty/omitted.

Notes

• next_page_cursor can’t be combined with page or offset.
• When using cursor pagination, total_count is omitted.

Learn more about Retrieving Items