Original research · 2026-07 edition

AI SEO Statistics: Crafts (2026-07 edition)

15 questions · 45 AI responses · 3 models · measured 2026-07-05

See how AI talks about YOUR businessFull data table

The question bank

The questions we tested — sampled from real buyer journeys in crafts.

Each model answered every question once, same wording, same day. These are the prompts behind every percentage on this page.

I'm looking for a unique anniversary gift that's handmade, but I don't know where to start looking beyond the major marketplaces.
Is it worth trying to make my own resin coasters for a party, or should I just hire a professional crafter to do them?
How can I tell if a seller is actually making their items by hand or just reselling cheap factory goods?
Why does a custom-knit sweater cost three times more than one from a retail store?
I need a custom wood carving for a retirement gift by Friday; who are the fastest reliable makers for quick turnarounds?
What questions should I ask a jeweler before commissioning a one-of-a-kind engagement ring?
What are some red flags to look for when buying expensive handmade ceramics online?
Is it better to buy directly from an artist's own website or through a third-party craft platform in terms of buyer protection?
Show all 15 questions
How do I find a maker who uses sustainable and non-toxic dyes for handmade baby clothes?
Can I find local glassblowers in my area to avoid the risk of items breaking during long-distance shipping?
I'm looking to order 30 custom leather journals for a retreat, what's a reasonable lead time and bulk price range?
How do I evaluate the durability of a handmade leather bag just by looking at photos and descriptions?
If I commission a custom watercolor portrait and I don't like the final result, what are my options for a refund or revision?
How involved will I be in the design process if I hire someone to make a custom stained glass window?
Where can I find an artisan who specializes in traditional techniques like hand-forged ironwork for home decor?

Model by model

25-point average divergence: which AI you ask changes the answer.

The divergence index is the average gap between the most and least likely model per behavior. Higher = the models disagree more about crafts buyers.

Behavior rates across 15 crafts buyer questions, 2026-07 edition. Last column: average across models.
ChatGPTClaudeGeminiConsensus
Recommends hiring a professional80%53%40%60%
Suggests DIY first13%7%7%80%
Names specific providers20%40%33%53%
Gives price or cost info20%7%27%73%
Tells to check reviews27%33%7%53%
Tells to verify credentials20%27%7%73%
Mentions case studies / portfolio33%33%13%67%
Mentions local proximity13%27%20%60%
Gives selection criteria40%60%33%53%
Warns about red flags20%33%27%73%
Asks a clarifying question47%73%0%13%
Recommends multiple quotes7%7%0%93%

By model

How each assistant handled Crafts questions.

Reading the 45 answers model by model shows how differently the three assistants treat the same crafts questions. On the most consequential behavior — whether to send the buyer to a professional at all — the rate ranged from 80% (ChatGPT) down to 40% (Gemini), a 40-point gap on an identical question set.

Across the 15 crafts answers it produced, ChatGPT recommended hiring a professional in 80% of them and suggested a DIY approach first 13.3% of the time. It named a specific provider in 20% of answers (about 0.3 distinct providers per answer) and included price or cost information 20% of the time. ChatGPT asked a clarifying question before answering in 46.7% of cases, warned about red flags or scams in 20%, and told the buyer to verify credentials in 20%, averaging 573 words per answer. On the remaining cues it told the buyer to check reviews in 26.7%, pointed to case studies or a portfolio in 33.3%, and framed the choice around local proximity in 13.3%; a selection-criteria checklist appeared in 40% of its answers and a recommendation to gather multiple quotes in 6.7%.

Across the 15 crafts answers it produced, Claude recommended hiring a professional in 53.3% of them and suggested a DIY approach first 6.7% of the time. It named a specific provider in 40% of answers (about 1.5 distinct providers per answer) and included price or cost information 6.7% of the time. Claude asked a clarifying question before answering in 73.3% of cases, warned about red flags or scams in 33.3%, and told the buyer to verify credentials in 26.7%, averaging 282 words per answer. On the remaining cues it told the buyer to check reviews in 33.3%, pointed to case studies or a portfolio in 33.3%, and framed the choice around local proximity in 26.7%; a selection-criteria checklist appeared in 60% of its answers and a recommendation to gather multiple quotes in 6.7%.

Across the 15 crafts answers it produced, Gemini recommended hiring a professional in 40% of them and suggested a DIY approach first 6.7% of the time. It named a specific provider in 33.3% of answers (about 1.2 distinct providers per answer) and included price or cost information 26.7% of the time. Gemini asked a clarifying question before answering in 0% of cases, warned about red flags or scams in 26.7%, and told the buyer to verify credentials in 6.7%, averaging 265 words per answer. On the remaining cues it told the buyer to check reviews in 6.7%, pointed to case studies or a portfolio in 13.3%, and framed the choice around local proximity in 20%; a selection-criteria checklist appeared in 33.3% of its answers and a recommendation to gather multiple quotes in 0%.

Taken together, ChatGPT is the assistant most likely to route a crafts buyer to a professional (80%) and Gemini the least (40%). ChatGPT produced the longest answers, at 573 words on average. Specific providers were named most often by Claude (40%) — even there, roughly one answer in 3 carried a name.

Where they disagree

The behaviors where the choice of model changes the answer.

The divergence index for this study is 24.8 points — the average distance between the most and least likely model across the coded behaviors. The gaps below are where which assistant a crafts buyer happens to ask matters most:

  • Asks a clarifying question: from 0% (Gemini) to 73.3% (Claude) — a 73-point spread.
  • Recommends hiring a professional: from 40% (Gemini) to 80% (ChatGPT) — a 40-point spread.
  • Gives selection criteria: from 33.3% (Gemini) to 60% (Claude) — a 27-point spread.
  • Tells the buyer to check reviews: from 6.7% (Gemini) to 33.3% (Claude) — a 27-point spread.
  • Names a specific provider: from 20% (ChatGPT) to 40% (Claude) — a 20-point spread.

The widest single gap — asks a clarifying question, 73 points — means a crafts buyer can receive materially different guidance on the same question depending only on which assistant they happen to open, so any visibility strategy built on a single model's behavior describes only part of the crafts market.

Where they agree

The points of near-consensus in Crafts.

On other behaviors the three models move almost in lockstep — the points of near-consensus for crafts, where all three landed within a few points of each other:

  • Suggests a DIY approach first: 6.7%–13.3% across all three (a 7-point spread).
  • Recommends multiple quotes: 0%–6.7% across all three (a 7-point spread).
  • Warns about red flags or scams: 20%–33.3% across all three (a 13-point spread).
  • Mentions local proximity: 13.3%–26.7% across all three (a 13-point spread).

Measured question by question, the three assistants coded a response the same way most consistently on "recommends multiple quotes" (identical coding in 93.3% of questions) and least consistently on "asks a clarifying question" (13.3%).

Every behavior, measured

All twelve coded behaviors for Crafts, averaged across the three models.

The behaviors AI models reproduce most often for crafts are recommends hiring a professional (57.8% on average), gives selection criteria (44.4%) and asks a clarifying question (40%); the rarest are recommends multiple quotes (4.5%), suggests a DIY approach first (8.9%) and tells the buyer to verify credentials (17.8%). Each figure below is the share of a model's 15 answers in which the behavior appeared at least once, averaged across the 3 models with the full per-model range in parentheses:

  • Recommends hiring a professional: 57.8% on average (ChatGPT 80%, Claude 53.3%, Gemini 40%) — a 40-point spread.
  • Gives selection criteria: 44.4% on average (ChatGPT 40%, Claude 60%, Gemini 33.3%) — a 27-point spread.
  • Asks a clarifying question: 40% on average (ChatGPT 46.7%, Claude 73.3%, Gemini 0%) — a 73-point spread.
  • Names a specific provider: 31.1% on average (ChatGPT 20%, Claude 40%, Gemini 33.3%) — a 20-point spread.
  • Warns about red flags or scams: 26.7% on average (ChatGPT 20%, Claude 33.3%, Gemini 26.7%) — a 13-point spread.
  • Mentions case studies or portfolio: 26.6% on average (ChatGPT 33.3%, Claude 33.3%, Gemini 13.3%) — a 20-point spread.
  • Tells the buyer to check reviews: 22.2% on average (ChatGPT 26.7%, Claude 33.3%, Gemini 6.7%) — a 27-point spread.
  • Mentions local proximity: 20% on average (ChatGPT 13.3%, Claude 26.7%, Gemini 20%) — a 13-point spread.
  • Gives price or cost information: 17.8% on average (ChatGPT 20%, Claude 6.7%, Gemini 26.7%) — a 20-point spread.
  • Tells the buyer to verify credentials: 17.8% on average (ChatGPT 20%, Claude 26.7%, Gemini 6.7%) — a 20-point spread.
  • Suggests a DIY approach first: 8.9% on average (ChatGPT 13.3%, Claude 6.7%, Gemini 6.7%) — a 7-point spread.
  • Recommends multiple quotes: 4.5% on average (ChatGPT 6.7%, Claude 6.7%, Gemini 0%) — a 7-point spread.

Trust signals

How well the models protect the crafts buyer.

Beyond whether to hire, the rubric codes how carefully each assistant protects the crafts buyer once a decision is made. Telling the buyer to check reviews or ratings appeared in 22.2% of answers on average. Verifying credentials or certifications appeared in 17.8%. Warning about red flags or scams appeared in 26.7%.

On structuring the decision, a selection-criteria checklist showed up in 44.4% of answers on average and a recommendation to gather multiple quotes in 4.5%. The single least-reproduced protective signal for crafts is "recommends multiple quotes" at 4.5% on average — the clearest opening for content that supplies it, since the models are not yet reliably surfacing that guidance on their own.

Referral behavior

Do AI models name Crafts providers?

For service providers the decisive question is whether these systems name anyone at all. Across 45 crafts answers, a specific provider was named in 31.1% of responses on average — roughly 1 distinct providers per answer. In practice the assistants behave far more as an explanatory layer than as a referral engine for crafts: visibility comes from being the reasoning a model reproduces, not from being the named recommendation.

The question set

What these 15 Crafts questions cover.

The 15 questions behind every percentage on this page were drawn from real crafts (ecommerce / online retail; buyer hiring decisions for this specific service) buyer journeys. Each was put to all 3 models once, with identical wording, so the rates above describe how the assistants handled this exact crafts question set — not a general prior or a hand-picked subset. The full list is shown earlier on this page; the coded percentages are what those specific questions produced.

How to read this

A note on the numbers.

A percentage here is the share of a model's 15 answers in which the behavior appeared at least once — not a confidence score. Because each model answered every question exactly once on 2026-07-05, the figures describe this specific crafts question set and snapshot rather than a general prior. The full protocol and coding rubric are documented in the study methodology.

Methodology

A controlled snapshot, documented end to end.

15 standardized buyer questions per industry, one response per model per question (ChatGPT (gpt-5-mini), Claude (claude-sonnet-5), Gemini (gemini-3-flash-preview)), collected 2026-07-05, coded against a fixed 12-behavior rubric with human QA. AI outputs vary with model version, location and time — figures describe this sample and window, and are refreshed each edition. Read the full methodology →