Truace
Subscribe
70/100

Moderate evidence

Cite
More
Export CSVExport JSONEmailLinkedInX

Shoppers responded best to mid-sized AI product explanations, a study of retail experiments finds

Study title: Too little, too much, or just right? How AI recommendation information volume shapes trust and consumer behavior

Evidence type
Randomized controlled trial
Sample
n = 51,248
Sources
1 sourcePeer-reviewed
Published
2 min read
Truvace record
TRV-2026-2485cite or check it

Study title: Too little, too much, or just right? How AI recommendation information volume shapes trust and consumer behavior

Researchers tested how much explanation an AI shopping assistant should give when recommending a product. They compared five ordered explanation packages, from sparse to dense. A randomized Chinese mobile-retail experiment found an inverted-U response, and moderate packages performed best within the tested sets.

Read the study abstract

Artificial intelligence (AI) shopping assistants explain why a product is selected, yet the appropriate defaults remain unclear. We define recommendation information volume as the amount of nonredundant, product-relevant information and compare five ordered explanation packages. Study 1 identifies an inverted-U response in a randomized Chinese mobile-retail experiment. Study S1 compares identical messages attributed to an AI assistant or a human salesperson, with separate source-neutral measurement validation. S…

Too little, too much, or just right? How AI recommendation information volume shapes trust and consumer behavior
"Looking down the front end - Self checkout at Fred Meyer" by Northwest Retail, CC BY-SA 2.0.

The gain

In a series of studies including a randomized Chinese mobile-retail experiment and data on 51,248 recommendation impressions, moderate-sized AI explanation packages performed best among the tested sets.

In brief

A behavioral study covered 51,248 recommendation impressions and service outcomes for 5000 distinct users, adding real-use evidence to the experiments. A further study compared identical messages attributed to an AI assistant or a human salesperson.

The work links benefits to perceived usefulness, trust in AI and decision comfort. Costs run through information overload and what the authors call algorithmic over-justification. The authors say the results support testing default explanations category by category.

Main points

  1. Five ordered explanation packages were compared, defined by the amount of nonredundant, product-relevant information.
  2. Study 1, a randomized Chinese mobile-retail experiment, found an inverted-U response to information volume.
  3. Study 3 analysed 51,248 recommendation impressions and service outcomes for 5000 distinct users.
  4. Study S1 compared identical messages attributed to an AI assistant or a human salesperson.
  5. Positive indirect associations ran through perceived usefulness, trust in AI and decision comfort.
  6. Negative indirect associations ran through information overload and algorithmic over-justification.

The full story

The research defines recommendation information volume as the amount of nonredundant, product-relevant information in an AI shopping assistant's explanation of why a product was chosen. Five ordered explanation packages were compared across several studies.

Study 1 was a randomized experiment in Chinese mobile retail and found an inverted-U pattern. Study 2 reported conditional positive indirect associations through perceived usefulness, trust in AI and decision comfort, and negative ones through information overload and algorithmic over-justification.

Study 3 supplied behavioral evidence from 51,248 recommendation impressions and service outcomes for 5000 distinct users. Moderate packages performed best within the tested sets. The authors explain this with bounded information processing and persuasion knowledge, across AI and human sources.

Sources

1 on file
  1. Journal of Retailing and Consumer Servicesdoi:10.1016/j.jretconser.2026.105205
    Peer-reviewed2026-11-07

Talk it over

Four questions to start onWho gets this benefit, and who was left out of the study or story?

Start with a question

Shoppers responded best to mid-sized AI product explanations, a study of retail experiments findshttps://truvace.com/item/too-little-too-much-or-just-right-how-ai-recommendation-information-volume-shape

  1. Who gets this benefit, and who was left out of the study or story?

    The gain In a series of studies including a randomized Chinese mobile-retail experiment and data on 51,248 recommendation impressions, moderate-sized AI explanation packages performed best among the tested sets.

    Answer in the thread

  2. How do we know? What can this kind of evidence show, and what can’t it?

    The evidence This rests on a randomized controlled trial with a sample of 51,248, from Journal of Retailing and Consumer Services.

    Answer in the thread

  3. What does it leave open? What else would you want to know before acting on it?

    The record’s own caveat Same-session estimates do not establish causal mediation, and because amount, diagnostic priority and content composition vary jointly, the experiments identify assigned-package effects, not pure quantity effects.

    Answer in the thread

  4. Have you seen AI used this way where you study, work or live? Did it go the same way?

    Answer in the thread

The questions are the same on every record. The details under them are taken from this record as written.

What readers say

More from Business