A new era of AI shopping: How brands are chosen by algorithms

Capgemini survey data shows 25% of consumers used generative-AI shopping tools in 2025 and a further 31% plan to adopt them, with 63% seeking hyper-personalised content but 71% worried about how AI uses their data. Retailers are advised to supply machine-readable, continuously refreshed product data and balance digital convenience with human support—trends that could shift e-commerce discovery and advertising dynamics (notably Amazon's move to limit OpenAI access) but are unlikely to trigger immediate market-wide moves.

Analysis

Market structure: AI-driven selection favors marketplaces and merchants that publish machine-readable, attribute-rich catalogs; winners include ETSY and SHOPIFY merchants that can be surfaced by LLMs, while gatekeepers that restrict access (notably AMZN) trade short-term control for potential loss of model-preferred inventory. Expect a reallocation of ~5–10% of digital marketing spend toward feed/data engineering and review aggregation over 12–24 months, boosting margins for SaaS feed/ratings providers and compressing weaker retailers' CAC economics.

Risk assessment: Tail risks include privacy/consent regulation (GDPR-style restrictions or US federal rules) that could remove LLM training/use cases, and model hallucinations creating large-scale consumer fraud/recall exposure; both can hit revenues in weeks and capex budgets over quarters. Hidden dependencies: third-party review platforms, inventory sync latency, and crawler access; catalysts that will accelerate adoption include OpenAI/Amazon integrations or major retailers deploying purchase-capable assistants in the next 3–9 months.

Trade implications: Tactical longs: exposure to SHOP (merchant monetization) and ETSY (curated inventory) with 3–12 month horizons; defensive long WMT for omni-channel capture of AI-assisted spend. Use options to skew risk: 3–6 month call spreads on SHOP to capture upside, and collars/purchase-protective puts on AMZN if shorting due to uncertain regulatory/tech battles. Rebalance ad-tech and pure-search exposure into retail-platforms and feed/SaaS names over 1–6 months.

Contrarian angles: Market consensus underrates how quickly clean, enriched product data can dethrone closed ecosystems—if AI adoption rises from 25% to 40% within 12 months, expect a 10–20% re-rating uplift for feed-integrated marketplaces. Conversely, shorting AMZN is binary and risky: Amazon’s own AI-ad control could protect margins long-term, making small, option-defined bearish positions preferable to large naked shorts.

Key Decisions for Investors

• Establish a 2–3% long position in SHOP (SHOP) over 3–12 months; complement with a 6-month 20% OTM call spread (buy 20% OTM, sell 40% OTM) sizing 0.5–1% notional to capture merchant monetization upside while limiting premium spend.
• Establish a 1–2% long position in ETSY (ETSY) for 3–9 months to capture curated inventory preference by LLMs; add a 3-month protective put if ETSY breaches -15% from entry to limit downside from a consumer-sentiment shock.
• Add a 1% long position in WMT (WMT) as defensive omni-channel exposure for 12 months, anticipating retention of customers who demand human/assistant hybrid experiences; trim 50% if same-store sales decelerate two consecutive quarters below consensus by >150 bps.
• Avoid large outright short on AMZN; instead size a 0.5–1% tactical bearish position using 3–9 month puts or put spreads (depending on cost) and only increase if regulatory action (FTC/DOJ filing or EU draft ruling) occurs within 90 days or if AMZN blocks major LLM integrations and revenue guidance misses by >3%.
• Rotate 3–5% of ad-tech/digital-search exposure into feed/SaaS vendors and review-aggregation plays over next 6–12 months, targeting companies with recurring revenue and API-first product catalogs; prioritize names that can demonstrate <24-hour feed refresh and 99% uptime SLAs.