How to Automate Pre-Sale Product Questions

Pre-sale questions are conversion events, not support tickets

Pre-purchase product questions are different from post-purchase support contacts in a critical way: answering them well increases revenue. A customer on a product page who asks 'will this fit a 34-inch waist?' and gets an instant, accurate answer is more likely to buy than one who waits for a response and loses the purchase momentum. Research consistently shows that live chat with immediate response during the purchase consideration phase lifts conversion rates by 10–20%.

There's a second economic benefit: pre-sale questions answered accurately prevent wrong-item purchases. A customer who buys the wrong size because they couldn't get a quick answer generates a return, an exchange, and additional support contacts — all more expensive than the original question. Accurate pre-sale AI answers are return prevention.

The most common pre-sale question types in ecommerce

Pre-sale questions cluster into predictable categories. Knowing which types apply to your store tells you which data to prioritize for your AI agent.

What data your AI agent needs for pre-sale questions

An AI agent can only answer pre-sale questions as well as the product data it has access to. Most ecommerce stores have decent product names and prices in their catalog but thin attribute data — which means the agent can confirm an item exists but can't answer 'is this machine washable?' Without good product data, pre-sale AI support is limited to availability and delivery questions.

The product data investment pays for itself in reduced returns and higher conversion. Here's what to prioritize:

1. 1Complete the attribute fields for your top-50 products by revenue — dimensions, materials, weight, care instructions, compatibility, what's included. Most stores have 80% of their traffic on 20% of their SKUs; enrich the high-traffic SKUs first.
2. 2Write a size guide for every product category — not just a generic size chart, but product-specific fit notes where relevant. 'This runs small — we recommend sizing up' is more useful than a standard size table.
3. 3Maintain a real-time inventory feed — availability questions are only answerable if the inventory data is live. A product listed as available that's actually out of stock is worse than no answer at all.
4. 4Add restock timelines for popular out-of-stock items — 'The blue option is expected back in stock in approximately 3–4 weeks' is an answer that keeps the customer in the consideration phase rather than sending them to a competitor.
5. 5Document compatibility for technical products — if you sell electronics, home goods, or anything with compatibility requirements, build a compatibility matrix for your top products and feed it to the agent. 'This works with [list]' and 'this does not work with [list]' are both valuable.

Sizing and fit: the highest-volume pre-sale category for apparel

For apparel, footwear, and accessories, sizing is the #1 pre-sale question and the #1 driver of wrong-item returns. Getting sizing AI right has a direct, measurable impact on both conversion and return rates.

• Create a product-specific size guide, not just a brand-wide one — a relaxed-fit hoodie and a structured blazer have different fit notes even if they use the same size labels. Note it explicitly.
• Include height/weight/measurement ranges for each size — 'Medium fits bust 36–38 inches, waist 28–30 inches, hip 38–40 inches' is actionable. 'Medium is our average size' is not.
• Note fit anomalies explicitly — 'this style runs 1 size small; we recommend sizing up for a standard fit.' This single note, if accurate, is the difference between a satisfied customer and a return.
• For international customers, include conversion charts — UK/EU/US size equivalents for every product category. A German customer asking for their size should get a German size answer.
• When a customer's measurements fall between sizes, give a direct recommendation based on their stated preference (loose vs. fitted) rather than saying 'it depends.' A preference question followed by a direct recommendation closes the sale; 'it depends' opens a new question.

Compatibility and technical questions

Compatibility questions have high stakes: a customer who buys an incompatible product returns it and feels foolish, not just inconvenienced. The emotional cost of a compatibility mismatch is higher than most other return reasons because the customer often feels they should have known.

For stores selling electronics, hardware, software, or accessory-dependent products, compatibility documentation is as important as pricing in your product catalog. Build it as a structured reference, not prose.

• For each product with compatibility requirements, list what it works with and what it doesn't — be exhaustive for the most common incompatibilities. 'Compatible with all iPhone models from iPhone 12 onward. Not compatible with Android devices' is clear and prevents misbuys.
• Document the 'works with X but needs adapter Y' cases explicitly — these are the most common compatibility failures because customers assume 'compatible' means plug-and-play when it doesn't.
• When the customer provides their specific device or model, the AI should confirm compatibility directly: 'Yes, this is compatible with your [specific model].' Not 'this is generally compatible with most [brand] devices.'
• If your compatibility data is incomplete for a specific product or model, the AI should say so and offer to connect to a human rather than guess. A wrong compatibility answer is worse than 'I'm not sure — let me get someone who can confirm.'

Availability and delivery timing questions

Availability and delivery questions are the easiest pre-sale category to automate because the data is already in your systems — inventory counts and shipping timelines. The barrier is usually connection, not data existence.

Configure your AI agent with live inventory access and a shipping calculator that factors in current carrier performance, the customer's location, and your fulfillment timeline.

• For in-stock items: give a specific delivery date, not a range. 'If you order before 2 PM today, this will arrive by Thursday June 5' is a confidence-building answer that converts. 'Typically arrives in 3–5 days' is less compelling and less actionable.
• For out-of-stock items: give a restock timeline if available, offer to notify the customer when it's back (if your platform supports this), and recommend an alternative if one exists. Don't just say 'out of stock' and leave the customer with nowhere to go.
• For international customers: give a delivery estimate in business days that accounts for customs processing time. 'Typically 8–14 business days to Germany; customs processing adds approximately 2–4 days' is honest and prevents WISMO contacts from international customers who expected domestic-speed delivery.
• For time-sensitive purchases (gifts, events, holidays): the AI should detect urgency language ('need this by Friday,' 'birthday on June 10') and confirm whether delivery is achievable before the customer buys. A pre-sale 'yes, it will arrive in time' that turns out to be wrong is worse than telling the customer to look for an alternative.