How conversational AI is turning e-commerce support into a revenue channel

For decades, customer support has been classified as a cost centre. The goal was simple: resolve the issue as cheaply and quickly as possible. Conversational AI is rewriting this equation.

The support-sales convergence

Every support conversation is a moment of attention. The customer is engaged, focused, and receptive. A well-designed AI agent uses this moment not just to resolve the issue, but to recommend relevant products, suggest upgrades, or offer personalised promotions.

The Feastables example

Feastables deployed a Certainly agent that handles support queries while also recommending products. The result: 6.5% of total revenue attributed to the AI agent, a 20% increase in average order value, and an NPS of 52 versus an industry average of 42.

How it works technically

The AI agent accesses the customer's order history, browsing behaviour, and product catalogue in real time. During a support conversation about a delayed order, it might say: 'While your order is on its way, have you seen the new flavour we just launched? Customers who liked what you ordered tend to love it.'

The personalisation advantage

Unlike static product recommendations on a website, conversational recommendations are contextual. The agent knows what the customer just asked about, what they have purchased before, and what is in stock. The recommendations feel natural, not spammy.

Measuring revenue impact

Track three metrics: conversations that include a product recommendation, click-through rate on recommendations, and revenue attributed to AI-recommended products. Most e-commerce brands see a measurable revenue lift within the first month.

The brands that still see support as pure cost are leaving money on the table. Literally.