Conversational commerce is the practice of selling through real-time conversation interfaces - SMS, chat, voice, messaging apps, and AI chat - rather than through traditional browse-and-purchase e-commerce flows. A shopper asks a question, gets a relevant answer, and completes the purchase inside the conversation without being sent to a product page or checkout form. The term was coined in 2015 but has become substantially more relevant in the last 18 months as AI chat interfaces and the maturation of SMS commerce have made the format practical at scale.
The shift in shopping behavior is real and accelerating. A growing share of product discovery happens inside AI chat - ChatGPT, Perplexity, Claude - rather than traditional search. Consumer engagement with SMS marketing has risen dramatically, with open rates in the 90%+ range versus email's 20-30%. And messaging apps (WhatsApp, Messenger, Instagram DM) have become de facto customer service and sales channels in many markets. The combined effect is that conversational surfaces are no longer a niche channel; they're a primary way shoppers engage with brands, especially outside the US.
SMS and MMS via Klaviyo, Attentive, and Postscript. Most mature. Drives 10-20% of total revenue for many Shopify brands running well-built SMS programs alongside email.
AI chat on the storefront - Shopify Inbox, Gorgias, and custom MCP-integrated assistants - that answer product questions, handle returns, and close sales in-context.
AI assistants and shopping agents - ChatGPT Shopping, Perplexity Shopping, Amazon Rufus - that recommend and transact on behalf of consumers. See agentic commerce for the broader implications.
Messaging apps - WhatsApp Business, Messenger, Instagram DM - especially important in Latin America, Southeast Asia, and much of Europe where shoppers expect brand conversations through these channels.
Voice - Alexa, Siri, Google Assistant - still a small share of commerce but growing with improved AI voice interfaces.
The signature of a well-built program is that conversations feel useful rather than promotional. Messages answer specific questions the shopper actually asked. Response latency is low (seconds on AI surfaces, minutes on human-staffed channels). Product recommendations are anchored to what the shopper said they want, not to whatever the brand most wants to sell. Recovery flows (abandoned cart SMS, re-engagement sequences) are tuned for conversion rather than sending the maximum allowed volume.
A conversational channel that generates low open or response rates typically signals one of three things: the audience didn't genuinely opt in (often a sign that the signup was buried in a broader form or coerced by a discount); the content is promotional rather than useful (treating SMS like email is a common pattern that erodes list quality fast); or response latency is too high (a 2-hour wait for a chat response converts at a fraction of a 2-minute wait). Diagnosis requires looking at response rate, unsubscribe rate, and conversion rate separately - each signals a different kind of problem.
The durable patterns:
Build the list the hard way. Opt-ins from exit-intent popups and aggressive incentives produce short-term list growth and long-term deliverability problems. The highest-quality SMS and chat lists come from genuine value exchanges (restock alerts, VIP early access, order updates) rather than "subscribe for 10% off" shortcuts.
Segment by actual shopper behavior. High-value customers, recent purchasers, cart abandoners, and browse abandoners each deserve different message frequency and content. Treating the list as a single audience under-performs segmented sending by large margins.
Invest in response quality, not volume. One well-crafted recovery message converts better than three generic ones. The channels that burn out fastest are the ones brands send the most on.
Treat AI chat as a product, not a feature. AI chat that actually helps shoppers - finding products, answering specific questions, resolving issues - drives meaningful revenue. AI chat that deflects to FAQ pages and hands everything off to email support erodes trust.
Measure contribution, not engagement. Open rate and click rate are diagnostic; revenue per recipient and net contribution margin are the numbers that matter.
.png)
.png)
