AI Product Builder

The AI Product Builder turns the files you already have — spreadsheets, PDF catalogs, ZIP archives, and product photos — into a fully structured TimberCloud catalog. Instead of hand-building every attribute table, option list, and category, you upload your existing files and let AI propose an editable mapping that you review and create in a few clicks.

Tip: This is the fastest way to onboard. Use it to migrate an entire existing catalog, or to stand up a single complex product (all its attributes, options, photos, and a starting price) in minutes.

The AI Product Builder is part of TimberCloud's native AI suite — it is included from the Commerce plan up, not a separately metered add-on. For the full picture of where AI shows up across the platform, see the AI overview. For the pricing side of AI, see the AI Pricing Assistant.

What it does

You upload one or more files, optionally add a sentence of guidance, and AI:

• Reads your spreadsheets, PDF catalogs, and images.
• Detects product attributes, selectable options, parts, and (in bulk mode) whole products and a category hierarchy.
• Proposes an editable mapping — a preview of everything it plans to create.
• Can match product photos pulled from a PDF catalog to the right options.
• Can suggest a starting pricing formula based on pricing it detects in your files.

Nothing is created until you review the mapping and confirm. You stay in control of every attribute, option, and product before it lands in your catalog.

The two modes

The AI Product Builder works in two modes. The mode controls what it builds.

Attributes mode (one product)

Use this when you are setting up a single product and want AI to build its attribute groups and options for you — for example, a cabinet door with Wood Species, Finish, and Glass Style attributes, each with their own option tables. Open the product, launch AI Product Setup, and upload a spreadsheet or PDF that describes that product's options. The mode toggle is hidden here because you are working inside one product.

Products mode / Catalog Import (many products)

Use this when you want to bulk-create your whole catalog. From the Products tab in the Catalog, click Import Catalog to open the builder with the mode toggle visible, then choose Products. AI will propose multiple products organized into categories, each with its own attributes and options, all in one reviewable mapping.

Supported file types

You can upload any mix of the following:

ZIP archives are unpacked in your browser, so only the relevant spreadsheets, PDFs, and images inside are uploaded — you can drop in a large export without sending gigabytes to the server.

Best practices for clean results

A little structure in your source files goes a long way:

• Spreadsheets: use one attribute per sheet/tab, with rows as options and columns as fields (option name, price, SKU, etc.). This maps most cleanly onto TimberCloud's attribute-and-option model.
• PDF catalogs are the best choice when you want AI to match photos to options — it can pull images from catalog pages.
• Keep one concept per file where you can (one product, one attribute, one category), and give files descriptive names.
• Use the instructions field (see below) to clear up anything ambiguous — your terminology, which column is price, how to group things.

Walkthrough: upload to create

The builder is a guided, multi-step flow. From start to finish:

1. Open the builder.
   • Bulk import: Catalog → Products tab → Import Catalog, then pick Products mode.
   • Single product: open the product → AI Product Setup (already in Attributes mode).
2. Upload your files. Drag and drop, or browse to select. You can add several files at once and remove any before continuing. Unsupported file types are rejected, and supported files inside any .zip are pulled out for you.
3. Add optional instructions. Type a sentence or two of guidance — for example, "Each sheet is a wood species; the 'Up-charge' column is the option price" or "Group these into Interior Doors and Exterior Doors." This is optional but improves accuracy.
4. Run the analysis. AI reads your files and proposes a mapping. This can take a little while for large catalogs; you can cancel a long-running analysis if needed.
5. Review and edit the proposed mapping. This is the important step. You see exactly what AI intends to create — attributes, options, parts, photos, categories, and (in bulk mode) products. Toggle items on or off, and adjust details so the mapping matches your catalog.
6. Create. Confirm, and TimberCloud builds the entities. A completion summary shows what was created so you can jump straight into the new product(s).

Important: Nothing is written to your catalog until you click Create on the review step. The proposed mapping is a preview you can fully edit or trim first.

Reviewing the mapping before you create

The review step is where you turn an AI proposal into a clean catalog:

• Toggle items on/off to skip anything you don't want created.
• Check option tables — option names, prices, and any detected SKUs or fields.
• Confirm categories and product groupings in bulk mode so products land where you expect.
• Verify photo matches (see below) before accepting them.

Spend the most time here — a quick review now prevents cleanup later.

Matching photos from a PDF catalog

When you upload a PDF catalog, the builder can extract product photos from the pages and match them to the right attribute options — the Catalog Photos matcher. Review each proposed photo match and apply the ones that look right, so your option tables come in with images already attached instead of you uploading them one by one.

Optional AI-proposed pricing formula

If your uploaded files contain pricing context, the builder can also suggest a starting pricing formula for the product. It arrives as a proposal you can accept as-is or refine — and like every formula in TimberCloud, you can keep tuning it afterward with the AI Pricing Assistant or the Formula Builder. Treat the proposed price as a starting point and confirm the numbers against your real pricing before publishing.

Permissions and availability

• Plan: the AI Product Builder is included from the Commerce plan and up. It is not a separately billed AI add-on, and it works during the 14-day trial.
• Role permission: your user role needs the Create / Edit Products permission to run product setup and create catalog entries. Roles are configured in Settings → Company → Permissions.

If you don't see the Import Catalog or AI Product Setup buttons, check your plan and your role permission.

What to double-check after an import

AI gets you most of the way, but always give a fresh import a once-over:

• Attribute structure and input types — confirm each attribute is the right kind (Dropdown, Text Input, or Description). See Attributes.
• Option names, prices, and SKUs — spot-check values against your source files.
• Categories and product placement (bulk imports) — make sure products are grouped correctly.
• Photos — confirm matched images landed on the right options.
• Pricing — review any AI-proposed formula and test a sample configuration before you publish.
• Visibility and behavior — set conditional visibility, "Match across items" locks, and per-document visibility as needed.

When everything checks out, finish the product the usual way in the product builder.

Steering the AI company-wide

Beyond the per-import instructions field, you can set a company-wide Global AI Instructions prompt at Settings → Catalog → AI Settings. This teaches every AI feature — including the Product Builder — your terminology, brand voice, and defaults, so you don't have to repeat the same guidance on each import.

Related

• Create a Product — the standard ways to add products, including the AI paths.
• Attributes — the attribute, option, and input-type model the builder fills in.
• AI Pricing Assistant — refine pricing in plain English after import.
• AI overview — every AI capability and which plan unlocks it.