The exact text we saved

This is the version exactly as it was saved, character for character. The SHA-256 hash below is a 64-character code worked out from that text. Run the same text through SHA-256 and you get the same code; change a single letter and the code comes out completely different. Each version also stores the previous version’s hash, which chains them together so no version can be swapped out quietly.

TRUVACE RECORD VERSION
record: TRV-2026-2491
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-11-12T07:42:17.403764Z
status: published
lens: g_space
sector: business
headline: Generative Artificial Intelligence and Sales Productivity: Field Experiments in Online Retail
dek: We quantify the short-term impact of generative artificial intelligence (GenAI) on sales performance through a series of large-scale randomized field experiments involving millions of users and products at a leading crossborder online retail platform. Over 2023–2024, the platform integrated GenAI into seven business workflows spanning customer service, consumer-product matching, advertising, and seller services. We find that GenAI adoption increases sales in most workflows, with effects ranging from no detectabl…
gain_title: In randomized field experiments with millions of users at a cross-border online retailer, generative AI raised sales by up to 16.3% in most of seven workflows, worth roughly $5 per consumer a year across the four that worked.
problem_title: (none)
trace_subject: (none)
gain_reading: In randomized field experiments with millions of users at a cross-border online retailer, generative AI raised sales by up to 16.3% in most of seven workflows, worth roughly $5 per consumer a year across the four that worked.
gain_evidence: the implied annual incremental value is roughly $5 per consumer
problem_reading: (none)
problem_evidence: (none)
quick_read: A series of randomized field experiments at a leading cross-border online retail platform tested generative AI across seven business workflows in 2023 and 2024, covering customer service, consumer-product matching, advertising and seller services. Sales rose in most workflows, by as much as 16.3% in the best case, while some workflows showed no detectable effect.

Across the four AI applications with positive sales effects, the authors estimate an incremental value of roughly $5 per consumer per year. They describe that as meaningful given the retailer's scale and the early stage of adoption.

The gains came mainly from higher conversion rates rather than bigger cart values. Product return rates and customer ratings did not worsen, and less experienced consumers gained more than experienced ones.
limitation: The study measures only short-term effects at one cross-border retailer, and some of the seven AI workflows showed no detectable sales impact.
tag: Evidence-backed gain
key_points: The experiments involved millions of users and products at a leading cross-border online retail platform. | The platform integrated generative AI into seven business workflows over 2023–2024. | Sales effects ranged from no detectable impact to 16.3%, depending on how much AI added over existing firm practices. | The four workflows with positive effects imply roughly $5 in incremental annual value per consumer. | Gains came through higher conversion rates rather than larger cart values. | Return rates and customer ratings did not deteriorate. | Consumers with less experience saw larger gains than experienced shoppers.
rundown: The study used randomized field experiments, which allows causal claims about the effect of generative AI on sales. It ran inside a real retail platform rather than in a lab, across customer service, consumer-product matching, advertising and seller services.

Results varied by workflow. The size of the effect depended on the marginal contribution of AI relative to what the firm already did, so some uses produced no detectable lift while the strongest reached 16.3%.

The authors link the conversion gains to lower search, information, communication and personalization frictions. Because returns and ratings held steady, the added sales do not appear to come at the cost of worse purchase outcomes.
sources:
- peer_reviewed | Management Science | https://doi.org/10.1287/mnsc.2026.00623 | 2026-11-06
prev: 0000000000000000000000000000000000000000000000000000000000000000

SHA-256 hash

73f3c21956a6871524858e2e69c526bd24d87ede90966f2e72de4c15c2439c16

Previous version’s hash

0000000000000000000000000000000000000000000000000000000000000000