---
title: "Three in five consumers use AI to research products but only 17% trust it enough to buy"
description: "Three out of five consumers now use generative AI to research products, yet only 17% to 30% trust it enough to delegate the actual purchase. Based on 2026 data, this summary explores how artificial intelligence has become the mainstream research assistant while final decisions remain human. Discover the latest market data, the fastest-growing shopping categories, and why brands must focus on selection optimization for recommendation algorithms."
url: https://www.thinkai.it/english/ai-shopping-2026-consumers-research-chatbots-purchase-trust-gap/
date: 2026-06-19
modified: 2026-07-07
author: "Gianluca Pezzi"
image: https://www.thinkai.it/wp-content/uploads/2026/06/consumatori-AI.jpg
categories: ["English"]
tags: ["Large Language Models"]
type: post
lang: it
---

# Three in five consumers use AI to research products but only 17% trust it enough to buy

Three out of five consumers currently use generative AI to research products before making a purchase. This is according to data collected between April and June 2026 by [Accenture](https://newsroom.accenture.com/news/2026/), [Gartner](https://www.gartner.com/en/newsroom/press-releases/2026-05-27-gartner-survey-finds-consumers-want-ai-shopping-help-but-not-ai-purchase-decisions), and [Invoca](https://www.invoca.com/blog/2026-report-consumers-embrace-ai-demand-fast-responses-brands-increases). On the other hand, the trust required to delegate the actual purchase remains stagnant between 17% and 30%. This gap highlights how the market is shifting speeds.

The data that triggered this shift dates back to January, when the [Capgemini Research Institute](https://www.capgemini.com/insights/research-library/what-matters-to-todays-consumer-2026/) revealed that **one out of four consumers** had made at least one purchase using AI shopping tools in 2025. In less than six months, that share has more than doubled.

Artificial intelligence has become a mainstream research assistant, but the final decision remains firmly in human hands.

## How Many Consumers Actually Use AI to Shop?

The figures add up and point in the same direction. In its [2026 Consumer Pulse Research](https://newsroom.accenture.com/news/2026/), which surveyed 25,590 consumers across 16 countries, **Accenture** reports that 74% would trust a personal AI agent more than their best friend to delegate a purchase. **Capital One Shopping**, in [its May 2026 summary](https://capitaloneshopping.com/research/ai-shopping-statistics/), highlights that AI traffic to US retail sites has skyrocketed by 4,700% year-over-year, with 63% of consumers wanting to shop with the help of AI.

Data from **Invoca**, collected between late May and early June 2026, shows that 60% of US consumers have used a chatbot like ChatGPT, Gemini, or Claude to research a major purchase. A year earlier, that figure stood at 41%.

The most surprising leap comes from Boomers. Only 11% used gen AI for pre-purchase research in 2025, compared to 34% today. The 60+ age group has tripled its usage, while Millennials and Gen Z remain stable at around 70%.

### Which Categories are Growing the Fastest

The breakdown by category, compiled from Prophet and Capital One data, shows that **electronics** (58%), **fashion** (41%), **home goods** (37%), and **toys and games** (28%) are the sectors where AI plays the most prominent role. Automotive is still underrepresented in these surveys, but [the April 2026 Prophet report](https://prophet.com/2026/04/the-2026-ai-powered-consumer-report/) shows that 41% of respondents discovered products through a chatbot that they didn’t even know existed.

## Why Does Trust in Delegating Purchases Remain Low?

There is a distinct gap between “I use AI” and “I let AI decide,” as measured by multiple surveys. **Gartner**, in [research published on May 27, 2026](https://www.gartner.com/en/newsroom/press-releases/2026-05-27-gartner-survey-finds-consumers-want-ai-shopping-help-but-not-ai-purchase-decisions), notes that 31% of consumers are willing to let AI narrow down choices for household groceries, and 28% for personal electronics. Kate Muhl, VP Analyst at Gartner, sums it up: consumers do not want to outsource purchasing decisions to AI; they want it to help them compare prices, find information, and narrow down their choices.

The [2026 Marketplace Shopping Behavior Report by ChannelEngine](https://www.channelengine.com/), based on 4,500 shoppers across five countries, confirms this trend: only **17%** feel comfortable completing a purchase through artificial intelligence. While 58% used it for research and 37% started their buying journey with an AI assistant, most prefer to double-check everything themselves when it is time to pay.

The [May 2026 Kantar survey](https://www.kantar.com/) reaches a similar conclusion: only 30% of consumers worldwide fully trust AI to make a purchase on their behalf.

### What Holds Back Full Delegation

Three recurring factors stand out across all surveys. The first is privacy: 71% of Capgemini respondents are concerned about how gen AI tools use their data. The second is transparency: 76% demand clear rules regarding when AI acts on their behalf. The third is the need for control: the final decision, especially for purchases with an emotional component or a high price tag, remains a personal choice.

## How is AI Changing the Buyer’s Journey?

The most transformative data relates to ChatGPT and other chatbots serving as alternative search engines. **Capital One Shopping** reports that 58% of buyers use gen AI instead of traditional search to find product recommendations. Nearly six out of ten consumers have already used AI for shopping, and 92% have encountered AI at some stage of their online journey.

According to **Invoca**, the top three use cases are: getting a quick summary of options (32%), comparing companies and brands (25%), and understanding complex topics like insurance and financing (22%). This last data point is particularly interesting for businesses selling products that require technical choices—ranging from cars and medical devices to financial products and insurance policies.

Research by [Optimizely](https://contentstamp.com/), cited by eMarketer, shows that 29% of consumers have already bought a product based on a single AI-generated response. Among those who did, 87% reported being satisfied with their purchase.

### The Best-Performing Use Cases

Surveys converge on four functions where AI is replacing traditional tools: summarizing reviews, comparing models, finding discounts and deals, and filtering options based on specific needs. Conversely, consumers remain skeptical when AI has to manage emotionally driven or complex situations: 43% recognize that AI struggles with context, 42% point to issues with complex problems, and 36% note a lack of empathy. For purchases that require empathy or context, consumers still prefer a human being.

## What Are the Implications for Brands?

The most profound shift affects sellers. [Dreen Yang, Global Industry Leader for Consumer Products & Retail at Capgemini](https://www.capgemini.com/insights/research-library/what-matters-to-todays-consumer-2026/), summarized the transformation: brands must move beyond search optimization and focus on selection optimization. Success increasingly depends on being selected by algorithms, not just being found by consumers.

This shift mirrors the transition from classic SEO to AEO (Answer Engine Optimization). Except here, it doesn’t just concern articles or web pages, but how products are presented to the algorithms that recommend them to buyers.

[Accenture](https://newsroom.accenture.com/news/2026/) arrives at a similar conclusion: 87% of surveyed consumers believe that AI will change the role of brick-and-mortar stores. Meanwhile, 31% think physical retail locations will become more important for creating moments of enjoyment and connection, as the functional aspect of purchasing shifts increasingly online and into AI interfaces.

### The Risks: DIY AI for Businesses

In its [2026 B2C marketing predictions](https://www.businesswire.com/news/home/20251028197329/it), **Forrester** sounds an alarm: one-third of companies will damage their customer experience with do-it-yourself gen AI chatbots. Cost pressures are driving brands to use virtual assistants in contexts where they are ineffective, resulting in a loss of trust. According to this analysis, the consumer adoption curve is outrunning the corporate maturity curve.

## What is Changing in Italy?

Specific data regarding Italy remains fragmented. The [2026 idealo Italian E-commerce Report](https://www.bitmat.it/internet/ecommerce/e-commerce-italiano-2026-italiani-sempre-piu-smart-nello-shopping-online/) shows that Italian shoppers have become more strategic and informed online, though a survey comparable to US studies on explicit AI tool usage is still missing. Capgemini included Italy in its 12-country sample, but the press release only provides aggregate data without a country-by-country breakdown.

An indirect indicator comes from ChatGPT’s penetration in Italy. According to late 2025 **OpenAI** data, monthly active users in the country surpassed 4 million. This user base roughly translates to one out of ten adult Italians, a figure that continues to grow month over month.

## What to Expect by the End of 2026?

Trend lines are converging. [Prophet](https://prophet.com/2026/04/the-2026-ai-powered-consumer-report/) estimates that by the end of 2026, over 80% of consumers in five key markets (US, UK, Germany, Australia, India) will use at least one gen AI function during their buying journey. **Accenture** predicts that autonomous AI agents, which currently see only 17-30% trust levels, will reach 50% trust within 18 months, driven by the rollout of agentic checkout features across Google, Amazon, and ChatGPT.

The turning point will be trust. It is not enough for AI to perform better; it must be perceived as transparent and controllable. Once consumers understand exactly how AI makes decisions on their behalf and can step in at any moment, delegation will rise.

For now, the landscape remains clear: **AI has already become the primary assistant for search**, but the moment of purchase remains a decision people prefer to make on their own.