An old idea, under other names
Long before AI, three classic models landed on the same three parts. Treacy and Wiersema's value disciplines call them product leadership, operational excellence and customer intimacy, and advise leading on one while staying at the industry standard on the other two [1]. Amazon's 2008 shareholder letter names vast selection, low prices and fast, convenient delivery, and its report for 2025 still does [2]. In PwC's 2018 survey of 15,000 consumers, experience ranked third, after price and product quality, and was worth up to a 16% higher price [3].
The right product
56% of shoppers look at the product images first, and a quarter of sites' images fall short [4]. In a 2026 shopper survey by Salsify, a vendor, 45% said they had returned an online purchase because the product information was wrong or misleading [5].
The right price
46% of US adults compare prices across stores, and only 9% never do. 24% already compare with an AI assistant like ChatGPT or Gemini [6]. And the right price is rarely the lowest: in a classic study, the best-known booksellers on a price-comparison site charged $1.72 more and still won the sale, because trust earns a premium [7].
The right experience
About 70% of carts are abandoned. Besides just browsing, shoppers leave over extra costs (40%), slow delivery (20%), not trusting the site with their card (19%) and the returns policy (13%) [8]. Five reviews instead of none raise the chance of a purchase by 270% [9], and a mobile site 0.1 seconds faster saw 8.4% more retail conversions, an observed link rather than a test [10].
AI now checks all three
AI shopping assistants ask the same three questions, from your data. OpenAI's product feed asks for the price, stock, a main image and a description, with fields for ratings, shipping and returns [11]. Google's Store Quality scores delivery, returns, photos, site speed and the store's rating, every day [12]. Shopify says AI agents need accurate product data, prices, stock and policies [13]. And a 2025 study at Columbia University found that AI shopping agents react to price, ratings and reviews, by different amounts from one model to the next [14].
One more part: being found
A store with the right product, price and experience still fails if nobody arrives. Shopify reported that orders from AI search grew elevenfold in 2025 [15], though only 14% of shoppers rely on AI picks regularly so far [5]. Being found is the part that changes fastest.
Where we fit
You choose what to sell and what to charge. Our Agent Apps make every product page complete, show where your prices stand against the market, and make your store easy to shop, from the menu to the answers shoppers get.
Sources
- Michael Treacy and Fred Wiersema, Harvard Business Review, 1993, and The Discipline of Market Leaders, 1995.
- Amazon.com, letter to shareholders, 2008, and annual report (Form 10-K) for 2025.
- PwC, Future of Customer Experience survey, 15,000 consumers in 12 countries, 2018.
- Baymard Institute, research on product image resolution and zoom, 2020.
- Salsify consumer research, 2,712 shoppers, a vendor survey, January 2026.
- Akeneo and Dynata, 1,000 US adults, via Chain Store Age, August 2026.
- Michael Smith and Erik Brynjolfsson, Consumer decision-making at an internet shopbot, Journal of Industrial Economics, 2001.
- Baymard Institute, cart abandonment rate across 50 studies, updated September 2025.
- Spiegel Research Center, Northwestern University, How online reviews influence sales, 2017.
- Deloitte and 55, for Google, Milliseconds make millions, 2020.
- OpenAI developer documentation, commerce product feed specification, read October 2026.
- Google Merchant Center Help, Store Quality, read October 2026.
- Shopify, Winter '26 Edition, agentic storefronts, December 2025.
- Allouah and colleagues, Columbia University, arXiv 2508.02630, 2025.
- TechCrunch, 4 November 2025: Shopify says orders from AI search are up 11 times since January 2025.