Pre-purchase questions in live chat come in many forms, but they usually mean the same thing: “I don't know which product is right for me.” Someone compares two coffee machines or several versions of one vacuum cleaner and cannot tell how they differ for their needs.
This guide gives your agents a repeatable advice script: three questions, a recommendation of one to three products, a close without pressure and a note that feeds your product pages. The shops in the examples are fictional.
The advice conversation in five steps
A pre-purchase conversation has one job: narrow the choice to a product the shopper understands and can buy with confidence. The order below works in chat because each step is one short message.
- Ask about the use.What for, where and how often the product will be used. This question settles the most.
- Ask about the budget.A range, not a single figure. Without it you will suggest something the shopper is not considering.
- Find the constraints.Size, compatibility, delivery date, color, anything that cannot change.
- Recommend one to three products.Each with one reason and a link to its page, never a list of ten options.
- Close without pressure.Leave the decision with the shopper and say what else they can check.
Where to show chat in a store at all, and which questions content handles better, is covered in the guide to live chat for an online store.
Three questions: use, budget, constraints
Ask one question at a time, or two per message at most. Start with the use, because it narrows the range the most. Put the budget second and offer ranges; they are easier to answer than “how much do you want to spend?”.
“Happy to help you choose. What will you mainly use the machine for: one coffee in the morning or several milk drinks a day?” Then: “Which price range are you looking at: up to 350 euros, up to 600 euros or more?” Finally: “Is there anything that has to fit, such as the width of the space on your counter or a delivery date?” The amounts are examples; adjust them to your range.
Do not ask about what you can already see: in Meetlead, agents see the visitor's current page and the path of their visit. When the chat is about clothing, the third question becomes a size question, which we cover separately in the guide to size questions in a clothing store.
The recommendation: one to three products, each with a reason
After three answers you know enough to suggest a specific product. Offer two or three only when they differ in something only the shopper can decide, such as price or size. Justify each in one sentence based on what the shopper wrote.
Scroll the table sideways to see all columns.
| What the shopper wrote | Recommendation | One-sentence reason |
|---|---|---|
| One black coffee each morning, up to 350 euros | One filter coffee machine | It does exactly what you need and stays within budget |
| Several milk coffees a day, up to 600 euros, narrow counter | Two machines up to 25 cm wide | Both have a milk frother; they differ in water tank size |
| A gift, needed before Friday | Three models in stock | We can ship any of them tomorrow; choose by price |
“For several milk coffees and a narrow counter I would suggest model X [product link], because it has a frother and is 24 cm wide. If you make coffee for several people, also look at model Y: it has a bigger tank but costs about 70 euros more.” Replace the brackets with the product page address from your own store.
The product link is not optional, because that is where the shopper checks photos and terms. Save repeated parts, like the greeting, as team macros, but write the reason fresh each time. The guide to reply templates shows how to avoid sounding like a bot. A request for a lower price is a separate conversation: see what to do when a customer asks for a discount.
Closing without pressure
A good close hands the decision back to the shopper. Instead of “it's the last one, better hurry”, say what else they can check: delivery time, return terms, the difference between two models. If low stock is true, state it as a fact without drama. For example: “If you like, I can check whether delivery will arrive before Friday. It's your call, and I'll stay in the chat.”
Directive 2005/29/EC on unfair commercial practices lists in Annex I, point 7, as always unfair, falsely stating that a product will only be available for a very limited time in order to push an immediate decision. EUR-Lex: Directive 2005/29/EC. The rule targets false time pressure, not every mention of a promotion ending. The directive applies in EU countries through national law; other rules apply outside the EU, including in the UK. This article is not legal advice.
A chat need not end in a purchase right away. Note what you recommended and why, so the next agent does not start from scratch if the same person writes tomorrow.
When to start a chat proactively on a product page
A proactive message helps when the shopper is clearly hesitating: reading one product page for a long time or moving between several similar products. In Meetlead Automations, a rule can start on a page URL visit or on time spent on the page, and conditions limit it to the current page, device or number of visits. Ready-made behavior-based scenarios are collected in the article on live chat automation.
- The rule runs only on product pages with the most questions, not across the whole store.
- The message appears after a longer read of the page, not right after arrival.
- The wording refers to the product: “Want help choosing between versions?” rather than a generic “How can I help?”.
- No proactive message appears on cart or payment pages.
- The rule runs during hours when someone will actually reply.
Automations is an App available on selected plans. Meetlead does not connect to your product catalog or your cart, so it will not recommend a product on its own and does not know what the shopper added to the cart. A person always prepares the recommendation.
Pre-sales questions as input for product pages and the knowledge base
Every advice conversation leaves a trace: what the shopper asked and what they did not find on the page. Mark these threads with one tag, for example “advice”, and add the product and the deciding question to the note. See the guide to tagging conversations and tickets. Once a month, filter the inbox by that tag.
A question asked several times about one product points to a gap on its page: a missing version comparison, a dimension or a note on who the product suits. How to move such gaps onto the page is explained in the article on the product page. You can add the same answers to the knowledge in Meet AI (Scale plan), which Copilot uses to prepare private drafts for agents. A draft never sends itself, and Meet AI does not pick products from a catalog; the agent checks every reply.
How to check whether chats end in a purchase
You need a goal that means a purchase. In Meetlead you define a goal in Analytics, such as “purchase”, and your store calls meetlead.goal() on the order confirmation page, optionally with a value and currency. The Traffic and conversion tab then shows how many sessions opened the widget, started a conversation and reached the goal. From the result you can open session details, and a visit links to its conversation if there was one.
Compare the same period and the same product pages before and after you change the script. A purchase after a chat does not prove the chat caused it, because people who would have bought anyway may also ask for advice. So alongside the goal, watch whether the same questions about a product become less frequent. For choosing events to measure, see the article on website micro-conversions. Report coverage depends on your plan.
Common questions
Should you ask about budget if it might feel awkward? Yes, but offer ranges to choose from. A question with ready answers is easier to reply to than a request for an amount.
What if no product fits? Say so plainly and name what is missing, for example “we don't have a model narrower than 25 cm”. An honest answer saves a return and builds trust.
