Install the Shopify app

Research

Discover the latest in AI-powered commerce, optimization strategies, and the future of product discovery.

AltText is not just a11y; its AI eye!

AltText is not just a11y; its AI eye!

announcements 20 min read

Gen Z Finds Brands in ChatGPT Now — and Trusts the Answer Like a Friend With Receipts

consumer insights 12 min read

The 'Purr-chase' model

conversion 3 min read

AI Luminosity - matters the most

research 8 min read
Research Paper

The Hidden Power of Alt Text: From Screen Readers to AI Answer Engines

Your images are invisible to AI. Here's what that's costing you.

╔══════════════════════════════════════════════════════════╗ ║ ┌───────────────┐ ║ ║ │ [IMAGE] │ → Without alt text: ║ ║ │ 🖼️ ??? │ • Screen readers: "Image" ║ ║ │ │ • ChatGPT: [nothing] ║ ║ └───────────────┘ • Google: [cannot index] ║ ║ ║ ║ ┌───────────────┐ ║ ║ │ [IMAGE] │ → With descriptive alt text: ║ ║ │ 🖼️ ✓ │ • Screen readers: Full description ║ ║ │ alt="..." │ • ChatGPT: Understands & cites ║ ║ └───────────────┘ • Google: Indexes in image search ║ ╚══════════════════════════════════════════════════════════╝

The New Reality: AI Crawlers Are the New SEO

In 2024, U.S. retail websites saw a 1,300% increase in traffic from generative AI searches compared to the previous year (Adobe, 2024). These visitors stay 8% longer, browse 12% more pages, and bounce 23% less than traditional search referrals.

But here's the catch: AI crawlers like GPTBot and ClaudeBot cannot see images—they read text. When ChatGPT, Perplexity, or Claude answers product questions, your images are completely invisible unless they have descriptive alt text.

"LLMs surface brands they can parse and trust. Images without strong alt text become invisible to retrieval and summarization systems."

— AltText.ai Research, 2025

The emerging llms.txt standard (proposed by Answer.AI founder Jeremy Howard) aims to help AI crawlers navigate sites—but it cannot fix missing alt text. Without proper image descriptions, your products simply don't exist in AI-generated answers.

The Accessibility Mandate: Laws Are Tightening

2025-2026 is a watershed moment for digital accessibility:

Regulation Deadline Scope Standard
ADA Title II (US) April 24, 2026 Government entities, public universities WCAG 2.1 AA
European Accessibility Act June 28, 2025 E-commerce, banking, transport across EU WCAG 2.1 AA
Section 508 (US) Active Federal agencies, federal contractors WCAG 2.1 AA

WCAG explicitly requires meaningful alt text for all informational images (WCAG 2.2 Success Criterion 1.1.1). In 2024, over 4,000 federal ADA website lawsuits were filed in the U.S. alone, with California's Unruh Act allowing $4,000 minimum statutory damages per violation (Accessibility.works, 2025).

The WebAIM Million Report: A Sobering Reality

The 2025 WebAIM Million study analyzed the top 1,000,000 home pages and found:

  • 18.5% of all images had missing alternative text (down from 21.6% in 2024)
  • ~1 in 3 images had missing, questionable, or repetitive alt text
  • 94.8% of pages had detectable WCAG failures
  • Missing alt text remains one of the six most common accessibility errors year after year

(WebAIM Million 2025 Report)

E-Commerce: Alt Text as a Conversion Engine

For online stores, alt text isn't just compliance—it's competitive advantage.

The SEO-to-AI Pipeline

Traditional Search AI-Powered Search ┌─────────────────┐ ┌─────────────────┐ │ Google Images │ │ ChatGPT/Claude │ │ ──────────────► │ │ ──────────────► │ │ Indexes alt text│ │ Reads alt text │ │ to rank images │ │ for context │ └─────────────────┘ └─────────────────┘ │ │ ▼ ▼ 25% of first-page AI answers now capture results contain images 10%+ of search queries (Marketing by Emma) (Mintlify, 2025)

Key Research Findings:

  1. 40% of users respond more to visual content in search results
  2. Products with visual content experience 12% faster traffic growth
  3. E-commerce stores using visual search see up to 30% higher conversions (Debutify, 2024)
  4. AI-generated recommendations increase click-through rates by 12.5% and conversion rates by 8.3% (ResearchGate, 2024)

What AI Needs From Your Alt Text

For product images, follow this formula:

Bad:  alt="product"
Bad:  alt="IMG_4521.jpg"
Bad:  alt="blue jacket winter jacket coat jacket sale discount buy now"

Good: alt="Navy blue insulated parka with faux-fur hood trim, front zip closure"

AI answer engines need:

  • Color and material (for visual search matching)
  • Product type and category (for contextual understanding)
  • Key differentiating features (for comparison queries)
  • No keyword stuffing (flagged as spam by modern AI)

The CHI 2024 Research: What Screen Reader Users Actually Want

The 2024 CHI Conference paper "From Provenance to Aberrations" (Das et al., ACM CHI 2024) studied alt text preferences across screen reader users and found:

  • Vision-to-language AI captions are often too generic ("a man standing") or verbose
  • Context matters: the same image needs different alt text depending on its purpose
  • Human-authored alt text (when done well) outperforms AI-generated descriptions
  • Combining AI suggestions with human editing produces the best results

The American Foundation for the Blind notes: "AI-generated descriptions appear to offer an elegant, scalable remedy... [but] AI models don't know what's relevant, emotionally salient, or functionally important." (AFB, 2025)

The Bottom Line: One Attribute, Three Audiences

╔════════════════════════════════════════════════════════════╗ ║ ALT TEXT ║ ║ │ ║ ║ ┌───────────────┼────────────────┐ ║ ║ ▼ ▼ ▼ ║ ║ ┌──────────┐ ┌──────────┐ ┌──────────┐ ║ ║ │ USERS │ │ SEARCH │ │ AI │ ║ ║ │ (a11y) │ │ ENGINES │ │ ANSWERS │ ║ ║ ├──────────┤ ├──────────┤ ├──────────┤ ║ ║ │ Legal │ │ Image │ │ Product │ ║ ║ │ mandate │ │ indexing │ │ citations│ ║ ║ │ │ │ │ │ │ ║ ║ │ 1B users │ │ 25% of │ │ 1,300% │ ║ ║ │ globally │ │ SERPs │ │ growth │ ║ ║ └──────────┘ └──────────┘ └──────────┘ ║ ╚════════════════════════════════════════════════════════════╝

Alt text is the rare optimization that serves accessibility compliance, SEO, and the emerging AI answer economy simultaneously. As AI search continues its exponential growth—ChatGPT went from 300M to 400M weekly users in just two months—the brands that invest in descriptive, meaningful image descriptions today will be the ones that AI systems cite tomorrow.

Quick Action Checklist

  • Audit your catalog: Run WAVE or a similar tool to find missing alt text
  • Prioritize product images: These have the highest commercial impact
  • Use the formula: [Color/material] + [Product type] + [Key features]
  • Avoid keyword stuffing: AI penalizes spam-like descriptions
  • Consider AI-assist tools: Use them for drafts, but human-edit for quality
  • Set up monitoring: Track AI referral traffic in your analytics

References

Published by Pagerank.ai — Optimizing e-commerce for the AI era.
Consumer Insights

Gen Z Finds Brands in ChatGPT Now — and Trusts the Answer Like a Friend With Receipts

Search used to be a place (a box). For Gen Z, discovery is becoming a conversation (a back-and-forth).

Jan 24, 2026

And the new mall directory isn't a list of blue links—it's an AI that summarizes, compares, and recommends in one breath.

That shift changes how trust is formed. When someone discovers your brand inside ChatGPT (or Gemini / Perplexity / Claude), they're not "clicking to learn." They're outsourcing uncertainty—and judging you by whether you show up as the right answer.

Below is a research-oriented, one-page synthesis of what the latest studies and industry data say about Gen Z + AI-driven brand discovery, why it increases trust (sometimes), and why AI Visibility (GEO / AIO) is quickly becoming a growth imperative.

What's changing: from "search results" to "answers"

Multiple surveys now show younger consumers using AI tools as a starting point for exploration and shopping research.

  • In an Adobe survey write-up, 24% of U.S. respondents reported going to ChatGPT first for online information, with Gen Z highest at 28%—and 47% of Gen Z said they discovered a new product or brand through ChatGPT.
  • A 2025 consumer survey report commissioned by Wildfire found 61% of U.S. consumers had tried gen-AI tools (ChatGPT, Gemini, Claude, Perplexity), and among AI users, Gen Z reported the highest "often/almost always" use of AI in online shopping (44%).

Translation: the top of the funnel is being re-routed through conversational interfaces. If your brand isn't legible to these systems, you're not "ranking lower"—you're often not in the room.

Why AI can create more trust (and where the trust comes from)

Across academic work on AI in shopping contexts, a consistent pattern emerges: AI doesn't magically create trust; it earns trust through a bundle of psychological shortcuts that reduce effort and increase confidence.

Here are the big trust mechanisms—supported by recent research:

1) Cognitive relief: less work, more certainty

Gen Z is digitally fluent, but not digitally patient. Research using technology acceptance models finds that perceived usefulness and ease-of-use of AI tools significantly shape Gen Z's intention to transact in AI-assisted shopping environments.

When AI compresses 30 tabs into one coherent recommendation, it delivers the modern luxury: less thinking.

2) "Personalization" feels like being understood

Recent Gen Z-focused studies report that AI exposure, attitudes toward AI, and perceived AI accuracy increase brand trust, which then boosts purchasing decisions. Another line of work suggests that perceived personalization and relevance can meaningfully influence trust and purchase intent in AI recommendation settings.

In plain terms: when an AI says "this fits your vibe," it feels less like advertising and more like translation.

3) Brand credibility becomes the trust "multiplier"

Research on intelligent personal assistants shows that the effect of AI features on Gen Z purchase intention is moderated by brand credibility. The assistant can open the door—but your credibility decides whether the room feels safe.

4) AI-backed trust is often "trust in process," not blind faith

Even as adoption grows, consumers remain selective about what they'll trust AI to do. YouGov research suggests trust is higher for AI helping with tasks like price comparisons, but much lower for fully autonomous actions like placing orders.

This matters because AI-driven discovery is often evaluation before it becomes automation. Brands that win early trust (information quality, transparency, consistency) are more likely to be chosen when the user moves from "Tell me" to "Buy it."

So why is AI Visibility imperative (especially for Gen Z)?

Because conversational engines act like gatekeepers of attention:

  • They decide which brands enter the "shortlist."
  • They summarize reviews and reputation into a single paragraph.
  • They reward pages that are structured, explicit, and easy to quote.
  • They quietly penalize vague claims, missing specs, and "trust gaps."

A BCG analysis (drawing on consumer research) reports that shopping-related gen-AI use grew 35% from Feb 2025 to Nov 2025, and shoppers like gen-AI because it's "direct, objective, transparent, and personalized." That's a description of why the channel converts: it feels like advice, not a pitch.

If Gen Z increasingly trusts AI to do the filtering, then your marketing job changes: you're not just persuading the shopper—you're also persuading the model that your brand is safe, relevant, and cite-worthy.

A practical playbook: "Be easy for AI to trust"

Think of AI Visibility as a three-layer stack:

Layer 1: Be legible (so you can be retrieved)

  • Put product facts in clean, explicit language (materials, sizing, compatibility, care, warranties, shipping/returns).
  • Use schema.org structured data (Product, Offer, AggregateRating, FAQPage) so machines don't have to guess.
  • Avoid "mystery meat" pages: missing specs, hidden policies, empty alt text, no brand story.

Layer 2: Be quotable (so you can be summarized accurately)

  • Write short "truth sentences" that an assistant can lift without distortion:
    • "Ships in 2 business days. Free returns within 30 days."
    • "OEKO-TEX certified. 100% organic cotton."
  • Make comparisons easy: include "Compared to X, this is better for Y" with evidence (lab tests, certifications, review themes).

Layer 3: Be verifiable (so trust survives scrutiny)

  • Showcase third-party signals: reputable press, certifications, lab tests, transparent sourcing.
  • Publish an FAQ that answers the real Gen Z questions (ethics, sustainability, durability, inclusivity, repairability).
  • Make review content easy to parse (summary blocks + full reviews), because AIs increasingly summarize reviews as part of research flows.

Tiny ASCII postcard from the future

GEN Z: "I need a sneaker that's comfy, ethical, and not cringe." AI: "Here are 3. Here's why. Here's what reviewers repeat." BRAND: (either shows up… or becomes an unsaid option)

The future is not "SEO vs AI." It's attention vs invisibility—and Gen Z is handing more of the steering wheel to conversational systems.

So yes: optimize for humans. But also optimize for the new librarian that hands humans their first shortlist.

References (papers, reports, reputable articles)

Published by Pagerank.ai — Optimizing e-commerce for the AI era.