Lyro report says AI shopping assistants are becoming a real commerce interface

A new Tidio report examines how AI shopping assistants affect sales at Amazon, Walmart, Magalu and Salesforce, arguing that retailers should focus on incremental value rather than attributed revenue. The report also lays out a test framework retailers can use to measure whether assistants truly lift orders, conversion and margin.

Why it matters: - AI shopping assistants are moving from customer service tools to measurable commerce interfaces. - The report argues retailers should judge these systems on incremental contribution, not engagement or attributed sales alone. - The distinction matters because assistant-attributed revenue can overstate true lift if shoppers would have converted anyway.

What happened: - Tidio published a new Lyro report, AI Shopping Assistants for E-Commerce, by Bart Turczynski, head of growth at Tidio. - The report examines evidence from Amazon, Walmart, Magalu and Salesforce. - It draws on 39 published sources, including retailer disclosures, market data, industry research and randomized experiments. - The report asks whether AI assistants create incremental value that would not have existed without them. - The full report is available here. - Media information is available here.

The details: - Amazon said more than 300 million customers used its shopping assistant in 2025. - Amazon said the assistant helped deliver nearly $12 billion in incremental annualized sales in 2025. - Amazon separately said shoppers using the assistant were 60% more likely to complete a purchase. - The report notes those figures measure different things: one is described as incremental, while the other is an observed association. - Amazon's public earnings release did not describe the experimental design behind the $12 billion estimate. - Walmart said shoppers who use Sparky have an average order value about 35% higher than shoppers who do not use it. - Magalu reported conversion three times higher in its WhatsApp shopping experience than in its app. - Salesforce estimated that AI and agents influenced $262 billion in 2025 global online holiday sales. - Salesforce's figure is an attribution estimate drawn from data on more than 1.5 billion shoppers. - Etsy's A/B test showed recommendation clicks rose 28.3% after a change to recommendations. - Etsy said conversion increased by a statistically insignificant 0.22%. - Etsy said GMV fell by a statistically insignificant 0.25%. - Etsy also found organic search clicks fell significantly by 1.4%. - Etsy's causal mediation analysis found the recommendation change increased conversion by 0.50%. - Etsy said reduced search activity cut conversion by 0.28%. - The report says roughly half of the genuine conversion gain was absorbed by displaced search activity. - Tidio says smaller retailers may first see value in support automation rather than sales lift. - MattressNextDay reports its assistant resolves 73% of customer conversations without a human. - MattressNextDay says that saves more than 400 hours per month for its support team. - Renovate Direct reports its resolution rate improved to 69% from 42% after routing tickets by intent. - The report labels both retailer examples as operator-reported results rather than controlled experiments. - Tidio says proactive AI assistance can be tested with holdout groups because retailers control whether prompts are shown. - The report recommends measuring eligible sessions, exposure rate, incremental order-rate change and average order value. - The report says retailers should then account for margin, returns and operating costs. - The report identifies incremental contribution per eligible session as a useful summary metric once a holdout result is available.

Between the lines: - The report is pushing the industry away from vanity metrics and toward causal measurement. - That matters because a system can appear successful if it gets credit for orders that would have happened through search or another channel. - The Etsy example shows how engagement gains can coexist with little or no net business lift. - Tidio is also signaling that the biggest near-term payoff for many retailers may be operational efficiency, not just conversion. - Tidio founder and CEO Tytus Gołas said the retailers in the report have the engineering budgets to deploy these systems at scale, while most retailers do not.

What's next: - The report says retailers should run their own experiments before forecasting business impact. - Tidio expects proactive prompts and holdout tests to make incrementality easier to measure. - The company says the key question for retailers is where AI shopping assistants create measurable incremental value.

The bottom line: - AI shopping assistants are already influencing commerce, but the report says only retailer-specific experiments can show how much of that activity is truly new.**