WooCommerce Variable Product Chatbot: How AeroChat Handles Size, Color, and Attributes in Real Time

Variable products break most chatbot setups. Not because the technology is incapable — because most chatbots are connected to WooCommerce at the product level, not the variant level. They know you sell a hoodie. They do not know whether the forest green one in size XL is in stock, whether the heavyweight cotton version runs large, or whether "slate grey" looks more blue or more grey on the actual product.

Those are the questions customers ask. When the chatbot cannot answer them, the customer leaves.

This article covers how variant-level chatbot integration actually works on WooCommerce, what the conversations look like when it works, and how to set it up — including how AeroChat handles it specifically, and where it fits relative to other platforms.

The variable product problem nobody talks about

A standard WooCommerce product — one SKU, fixed price, no variants — is simple for any chatbot to handle. Ask "do you sell this?" and the answer is yes or no.

Variable products are different. A single WooCommerce hoodie listing might contain 48 distinct SKUs: 6 colours × 4 sizes × 2 material weights. Each combination has its own stock status, sometimes its own price, and often its own set of questions attached to it.

Here is what your customers are actually asking, and what the chatbot needs to answer each one:

None of these questions have a single-word answer. All of them require the chatbot to understand your product at the attribute level, not just the product level. And all of them, left unanswered, turn into abandoned carts.

A chatbot that only knows you sell a hoodie is useless for every one of these conversations.

How variant-level chatbot integration works on WooCommerce

The mechanism is the same across any WooCommerce chatbot platform that supports it: the WooCommerce REST API exposes variation data, and the chatbot reads that data instead of just the parent product.

Specifically, the API endpoint /wp-json/wc/v3/products/<id>/variations returns each variation as a distinct object with its own attributes (size, colour, material), stock status, price, and any variation-specific descriptions or images. A chatbot connected at this level can answer "do you have the forest green hoodie in XL?" by querying the variation data for the attribute combination colour: forest green + size: XL and returning the live stock status for that SKU.

This is SKU-level awareness, not product-level awareness. The difference is the entire gap between a chatbot that is useful for variable products and one that is not.

Platforms that support variant-level reads on WooCommerce:

The setup steps differ by platform but the underlying mechanism is the same. Where they actually differ is in what they do with the variation data — how the AI interprets a question like "the dusk blue, is it more navy or more blue?" depends on the chatbot's LLM quality and how well it surfaces the variation description text you have written.

The rest of this article covers what good variant-level integration looks like in practice, with AeroChat as the example. The technique applies to other platforms with comparable capability.

What the catalog sync covers

When AeroChat connects to WooCommerce, the sync includes:

• Parent product name, description, and images

• All defined attributes (size, colour, material, fit, scent — whatever you use)

• All active variations with their specific attribute combinations

• Stock status per variation (in stock, out of stock, on backorder)

• Variation-specific pricing where applicable

• Variation-specific images where uploaded

• Custom variation descriptions if added in WooCommerce

The catalog syncs at regular intervals and on webhook trigger when a stock change occurs in WooCommerce. A product that sells out mid-conversation updates within minutes, not hours. This matters because the worst chatbot experience is being told a variant is available, clicking through, and finding it sold out at checkout.

A note on large catalogs: Stores with 5,000+ variations can hit WooCommerce REST API rate limits during the initial sync (default WooCommerce API limit is 25 requests per second, though hosting providers may impose lower ceilings). For catalogs at this scale, schedule the initial sync during low-traffic hours and confirm your hosting plan's rate limit before starting. AeroChat handles this with batch syncing; smaller platforms may need manual configuration.

Five variable product conversations the chatbot handles automatically

These are real conversation structures, not hypotheticals. Each represents a query type that would otherwise require a human agent.

Conversation 1: Combination availability check

Customer: Hi, do you have the linen trousers in sand, size 32?

Chatbot: Yes — the sand linen trousers in size 32 are in stock. Here's the direct link to add them to your cart: [link]. Want me to check anything else about this style?

The chatbot checked the specific colour + size combination against live WooCommerce inventory and returned a direct variation link, not the general product page. The customer does not need to navigate dropdowns — they land on the exact variant.

Conversation 2: Sizing guidance

Customer: I'm usually a medium in Zara but I've heard your sizing runs small. What should I get?

Chatbot: Our sizing does run slightly smaller than Zara's standard cut. If you're a true medium in Zara, we'd recommend going up to a large with us. The large fits a 38–40 inch chest and 30–32 inch waist. Does that match your measurements?

This conversation works because the AI's knowledge base is trained on your brand's size guide and any fit notes you have added — "runs small," "relaxed fit," "true to size." The AI interprets the customer's reference point against your brand's size guidance and gives a specific recommendation. Not "check our size guide." A recommendation.

Conversation 3: Attribute comparison

Customer: What's the difference between the fleece and the heavyweight cotton versions? Which is warmer?

Chatbot: The fleece version (380gsm) is softer and warmer — better for colder weather or indoor wear. The heavyweight cotton (320gsm) is more structured, holds its shape better over time, and works well as a mid-layer. If warmth is the priority, go fleece. If you want something that looks sharper and lasts longer, the cotton is the better build.

The chatbot is drawing on attribute-level product data and any additional description you have written for each variation. If your WooCommerce variation descriptions are thin, the AI answers with thin information. This is the most important reason to write meaningful variation descriptions — not for SEO, but for the conversations your customers are already trying to have.

Conversation 4: Colour description for online shoppers

Customer: The 'dusk blue' — is it more of a light blue or closer to navy? The photo looks different on my phone vs laptop.

Chatbot: Dusk blue sits between a medium blue and a muted slate — closer to a faded denim than a bright blue, and noticeably lighter than navy. Think washed denim shade.

Honest framing: the AI is retrieving and presenting a description you wrote in the variation description field. It is not interpreting the colour from the photo. If your variation description for "dusk blue" just says "dusk blue," the chatbot has nothing to answer with. The setup section below covers how to structure descriptions so this conversation is possible.

Conversation 5: Out-of-stock restock intent

Customer: The burgundy one in size S is showing out of stock. Any idea when it's back?

Chatbot: We're expecting a restock on the burgundy size S within 10–14 days. Want me to send you a WhatsApp message the moment it's back in? I just need to confirm you're happy to receive that notification.

When the customer confirms, the chatbot adds them to a restock notification list. When WooCommerce marks the variant back in stock (via webhook or sync), the notification fires automatically. Customers who opted in for restock notifications convert at meaningfully higher rates than cold traffic — they have already decided they want the product.

Guided selling: when the customer does not know what they want

The five conversations above are reactive — the customer asked a specific question. Guided selling handles a different scenario: the customer who says "I want to buy a hoodie as a gift but I'm not sure which one."

The chatbot handles this through a structured attribute narrowing flow:

1. Budget or price range — "What's the budget you're working with? I can narrow it down from there."

2. Use case or occasion — "Is this for everyday wear, outdoor use, or more of a smart-casual situation?"

3. Size or fit preference — "Do you know their size, or would you prefer something with a more forgiving fit?"

4. Colour preference — "Any colour preferences? I can show you what's in stock in the colours that match what you're describing."

At the end of this flow, the chatbot presents two or three specific variations — not the full catalogue — with direct add-to-cart links. Decision paralysis is the reason customers abandon configurable products. Narrowing 48 variants to 2 or 3 relevant options is what the guided flow does.

This works for any product category with multiple attributes: furniture (material, size, finish), supplements (flavour, format, strength), clothing (size, fit style, fabric), electronics (storage, colour, model tier).

Writing WooCommerce variation descriptions that make the chatbot useful

The AI answers from the data in your WooCommerce catalog. The quality of the answers is directly determined by the quality of what you have written.

Thin variation description (chatbot cannot help):

Forest Green, Size XL

Useful variation description (chatbot can have a real conversation):

Forest Green: a deep, muted green — closer to pine than bright olive. Photographed in natural light; indoor lighting renders it slightly darker. Heavyweight 320gsm cotton. This colourway runs true to size. Available in S through XL.

The second version gives the chatbot material to work with for colour description questions, sizing questions, and material questions — all without a human agent involved.

Practical rule: For each variation that differs meaningfully from others (colour appearance, fit, material weight, use case), write 2–3 sentences of plain-language description inside the WooCommerce variation description field. This investment in product data pays back in every customer conversation that gets resolved automatically — regardless of which chatbot platform you use.

Multilingual stores

If you run WooCommerce with WPML or Polylang, variation descriptions are translatable per language. Whichever chatbot platform you use needs to read the language-specific variation data, not just the default language. AeroChat handles this when WPML's REST API integration is enabled; verify the same with any other platform before assuming multilingual coverage works.

Setting up AeroChat for WooCommerce variable products

The setup steps below are AeroChat-specific. The structure (REST API connection, variation-level sync verification, content prep, knowledge base, automation flows) applies to any chatbot platform that supports variant-level reads — the field names and screens differ.

Step 1: Connect AeroChat to WooCommerce. In your WooCommerce dashboard, generate a REST API key pair (Consumer Key and Consumer Secret) under WooCommerce → Settings → Advanced → REST API. Set the permissions to Read. Enter these credentials in AeroChat's WooCommerce integration settings. The initial full catalog sync runs automatically — for catalogs above 1,000 variations, expect 2–10 minutes; above 5,000, schedule for low-traffic hours.

Step 2: Verify variation-level sync. After the initial sync, search for a specific product variation inside the catalog view. Confirm that individual colour/size combinations appear as distinct entries with their own stock status. If only parent products appear, the API connection is at the product level — check the API permissions and re-sync.

Step 3: Add variation descriptions in WooCommerce. For your top 20 selling variable products, open each product, navigate to the Variations tab, expand each variation, and add a plain-language description to the Description field. Focus on: colour appearance in plain language, fit notes (runs large/small/true), material feel and weight, and use-case guidance. This is the highest-leverage content work you will do for chatbot performance.

Step 4: Build the size guide knowledge base entry. Create an entry titled "Size Guide" and paste your full size guide content including measurement guidance (chest, waist, hip), brand-specific fit notes per product line, and comparison notes for common reference brands if your customers frequently ask about sizing relative to Zara, H&M, or ASOS. The AI draws on this for every sizing conversation.

Step 5: Configure the restock notification flow. In the automation settings, create a flow triggered by intent phrases like "out of stock" or "when will it be back." The flow captures the customer's confirmation to receive a restock notification, stores the product variation and customer contact, and triggers the notification when WooCommerce marks the variant back in stock. Test end-to-end with a test product before going live.

Step 6: Set up guided selling entry points. Configure a proactive chat trigger on product pages with more than 8 variations. After 45 seconds on the page with no add-to-cart action, the chatbot opens: "Lots of options here — want me to help narrow it down?" This catches decision paralysis before it becomes a bounce.

When variable product chatbot integration is the wrong investment

Most articles on this topic skip this section. It's the most useful one for actually deciding.

Skip variant-level chatbot setup if:

• Your catalog has fewer than 50 active variations. The setup work and content investment outweigh the conversion lift. Stick with a basic chatbot or no chatbot.

• Your variations don't differ meaningfully. If the only difference between SKUs is size, customers don't need a chatbot to interpret them. Variant-level integration shines when colour, material, or fit varies.

• Your variation descriptions are thin and you don't plan to invest in writing them. The chatbot can only surface what you've written. Without good variation copy, you're paying for a feature that has nothing to retrieve.

• You sell mostly through a single channel where customers don't ask pre-purchase questions. B2B wholesale, subscription replenishment, and similar models often don't have the conversational pre-purchase pattern that justifies this setup.

Fix the underlying content and catalog hygiene first. The chatbot amplifies what's already there.

What this does to conversion (with honest framing)

Variant-level chatbot integration tends to lift conversion in three specific ways. The exact size of the lift varies meaningfully by category, AOV, and traffic source, so treat any single number as directional rather than predictive.

• Variant-specific conversion uplift. Customers who get an accurate, direct answer to "do you have this in [colour] [size]?" tend to convert at significantly higher rates than customers who self-navigate variant dropdowns, because the friction of finding the right SKU is removed. The exact gap varies, but consistently shows up as the largest single lift in test data we've seen across AeroChat-connected stores.

• Higher AOV on guided selling conversations. The narrowing flow surfaces complementary products and higher-value variants naturally because the AI is asking about use case and preference, not just confirming availability.

• Restock notification conversion. Opted-in restock notifications convert at meaningfully higher rates than cold traffic, because the customer has already decided they want the product. This is revenue most stores write off as permanently lost.

If you want hard numbers for your specific store, the only reliable way to get them is to A/B test with chatbot on a percentage of traffic for 30 days. Industry benchmarks are useful for direction; your own data is what you should plan around.

FAQ

Can a WooCommerce chatbot read individual variation stock status in real time?

Yes, if the chatbot connects via the WooCommerce REST API at the variation level. The /products/<id>/variations endpoint exposes stock status per SKU. AeroChat, Tidio, Gorgias, and several other platforms support this. When a customer asks about a specific colour and size combination, the chatbot checks that exact SKU's stock status and responds accordingly.

What happens if a customer asks about a variation that does not exist?

A well-configured chatbot identifies that the requested combination (for example, "white in size XXS") does not exist as a variation and offers the closest available options — the same product in the nearest size, or white in the available sizes. The chatbot should not invent availability or redirect the customer to the product page without an answer.

Do I need to write descriptions for every WooCommerce variation?

Not every variation, but the ones where visual or fit differences matter. Colour variations that look noticeably different from each other, material variations with different weight or feel, and fit variations (slim vs relaxed) benefit most from written descriptions. Generic variations (just a size difference with no other attribute change) need less.

Does AeroChat handle variable products across WhatsApp and Instagram DMs as well as website chat?

Yes. The same catalog connection and conversation logic applies across WhatsApp, Instagram DMs, and website chat. A customer who DMs on Instagram asking about a specific variant gets the same SKU-level answer as a customer using the website chat widget.

What WooCommerce product types does AeroChat support beyond variable products?

Simple products, grouped products, and external/affiliate products are all readable via the WooCommerce REST API. Variable products require the most configuration to get full value from (variation descriptions, size guide, restock flows) — but all product types benefit from the catalog connection.

Can the chatbot send restock notifications automatically when WooCommerce stock updates?

Yes, via webhook. When WooCommerce marks a variation back in stock, a webhook fires to the chatbot platform and the restock notification flow sends the message to any customer who opted in for that specific variation. This works on AeroChat and on any platform that supports inbound WooCommerce webhooks tied to stock change events.

Does this work with WPML or Polylang for multilingual WooCommerce stores?

It depends on the chatbot platform. The chatbot needs to read language-specific variation data, not just the default language. AeroChat reads WPML translations when WPML's REST API integration is enabled. Verify this explicitly with any other platform before assuming multilingual coverage works.