By industry
Customer stories
See how our clients are using Zipchat and the awesome benefits it's bringing to their businesses!
Use case
A shopper who feels heard spends more. Zipchat listens before it recommends.
No credit card required · Free forever plan available
Stores using Zipchat see an average 21% lift in AOV for chat-assisted purchases
Source: Zipchat analysis of active client deployments
At a glance
Zipchat's AI recommends complementary products and upgrades inside live chat conversations, at the moment a customer is actively engaged with a purchase decision. Unlike static product carousels, the recommendations are conversational, context-aware, and triggered by what the customer is already asking about. Stores using Zipchat average a 16.4% chat-to-sale ratio, according to Zipchat analysis, with upsell prompts contributing to higher-value orders from engaged chat sessions.
Static product carousels ("Customers also bought") fire on every product page, for every visitor, at all times. They are not timed to intent. A visitor browsing casually at the top of the funnel gets the same "complete the look" suggestions as a visitor who has already decided to buy and is reading the product description carefully. The result is noise. Most carousels are ignored.
Email upsells fire post-purchase. At that point, the transaction is closed. You might get a second purchase, but you missed the opportunity to increase the value of the order that just happened.
The gap is the live buying session: the 5-minute window when a customer has decided to buy but has not yet checked out. That is when an upsell has the highest chance of success. The customer is in buying mode. They have already decided to spend money with you. The barrier to adding one more item is at its lowest.
Static tools do not know when that window is open. They fire continuously and miss the moment. A conversation that is already happening knows exactly when that window is open.
Zipchat's upsell engine uses three inputs to decide what to suggest: your product rules, purchase pattern data, and what the customer just said.
When a visitor is in an active chat session, the AI monitors the context of the conversation. When certain signals appear, the AI introduces a relevant recommendation naturally, as part of the conversation.
Upsell signals the AI recognizes:
When a signal appears, the AI recommends a specific complementary product or upgrade using the language of the conversation rather than a generic "you might also like" insert.
Example: A customer asks whether a jacket is warm enough for winter hiking. The AI answers based on your product content, then notes that many customers who buy this jacket also pair it with your midlayer fleece for sub-freezing temperatures, and asks if that is relevant for their trip. The customer says yes. AOV increases by one product.
That is not a scripted flow. The AI derived the recommendation from the conversation context and your product catalog. No rules to write. No segments to build.
Three AOV levers:
Products that work with what the customer is already buying. Triggered by add-to-cart events and high-intent product questions.
Higher-margin variants of the product the customer is considering. Triggered when customers ask about specific features that a premium variant has.
When a customer's stated use case matches a product grouping in your catalog, the AI suggests the full set. This captures customers who would have made multiple separate purchases and accelerates the second purchase into the first order.
Zipchat reads your product pages during initial setup. The AI builds its own understanding of your catalog: product categories, features, price points, use cases, and relationships between items.
In the first two weeks, test the AI by asking it about product combinations and use cases you know well. Correct any misunderstandings. Accurate product knowledge is the foundation of relevant recommendations.
Look at your existing order data: which products are most commonly bought together? Which upgrades have the best margin? Share this context in your custom instructions so the AI prioritizes these combinations in relevant conversations.
Set up a trigger for customers who have added to cart but have not initiated checkout. The opening message of the proactive chat can be framed around ensuring they have everything they need, which naturally opens the door for complementary recommendations.
Compare the average order value for sessions that included a Zipchat conversation against sessions without. This isolates the AOV impact of AI recommendations. Most stores see this delta within the first 30 days of data.
The primary metric for this use case is AOV delta: the difference in average order value between sessions with AI chat interaction and sessions without.
Supporting metrics from Zipchat analysis across client stores:
For stores where AOV improvement is the primary goal, the most important configuration variable is the quality of product relationship data the AI has access to. Stores with well-described products and clear use case content in their product pages see more accurate complementary recommendations.
Estimate your revenue uplift: Zipchat ROI Calculator
| Upsell Scenario | Standard Approach | With Zipchat Recommended |
|---|---|---|
| Customer adds one item to cart | Static 'Customers also bought' carousel, ignored by most visitors | AI monitors chat context and introduces a relevant complementary item when the customer signals high intent |
| Customer asks about a product feature | Product page content, no escalation to recommendation | AI answers the question and introduces an upgrade that addresses the feature more directly |
| Customer mentions a specific use case | No mechanism to capture intent and match it to catalog | AI recognizes the use case, matches it to the right products, and suggests them in conversation |
| Post-purchase email upsell | 5-15% open rate; upsell missed the buying window | In-session recommendation captures the upsell during the active purchase decision |
| Bundle opportunities | Static bundle pages that customers must find | AI surfaces bundle suggestions when the customer's stated use case matches a product grouping |
AOV improvement through AI recommendations requires a product catalog with natural upsell and cross-sell relationships. A store that sells a single-category, low-price product with no natural complements (disposable goods, single-use items) will see limited AOV lift from recommendations.
The AI recommends based on what it finds in your product content. If your product pages do not describe use cases, compatible items, or product applications in detail, the AI has insufficient context to make relevant recommendations. Thin product content limits recommendation quality.
This use case also requires meaningful chat session volume. Stores with under 300 monthly visitors will generate too few chat interactions to produce a statistically measurable AOV impact in the first month.
Average order value is total revenue divided by number of orders. A store with $100,000 in revenue from 1,000 orders has an AOV of $100. AOV matters because increasing it requires no additional customer acquisition cost. You already have the customer in a buying session. Selling them one additional item costs far less than acquiring a new customer, and the margin contribution goes directly to profitability.
Product recommendation widgets are static and context-blind. They show the same suggestions to every visitor on a page, regardless of what that visitor has expressed interest in. AI chat is contextual: it knows what the customer has asked about, what they have added to cart, and what use case they described. That context produces more relevant recommendations. A relevant recommendation converts at a higher rate than a generic one, which means more recommendations accepted per session.
No. The AI reads your product catalog and learns relationships between products from your content. You do not build rules like 'if product A is in cart, show product B.' You can add custom instructions to prioritize specific high-margin combinations, but the AI derives most recommendations from product content and conversation context without manual configuration.
The highest-conversion upsell moment is immediately after the customer has demonstrated clear purchase intent, after an add-to-cart event or after expressing that a product meets their needs in conversation. This is when the customer is in buying mode and the psychological cost of adding one more item is lowest. Static carousels fire continuously and miss this window. In-chat recommendations fire at this exact moment.
Yes. AOV improvement is one outcome from AI chat. The same AI also converts visitors who would not have purchased at all, reduces cart abandonment, and handles post-purchase support. AOV is the incremental revenue layer on top of base conversion improvement.
