Before

Design tool decisions are increasingly made before a buyer ever reaches the vendor's website. Product managers, designers and engineers now ask ChatGPT, Claude, Gemini and Perplexity which AI design tool to use, and build their shortlist from the answer.

For Magic Patterns, the goal was to be positioned correctly as the #1 AI design tool for product teams, especially when buyers compare it with Figma, Claude Design and Lovable. In late July 2026, Magic Patterns appeared in just 17.3% of AI answers in its category, far behind Figma.

Insight

The Prompting Company found that AI assistants were answering design tool questions from someone else's material. In this category, 79% of AI citations still come from independent third-party sites, almost all of them blog posts, and Figma's own resource pages are among the most cited. So when buyers asked which tool to use, assistants named Figma far more often than Magic Patterns.

Many of the most-cited pages are broad roundups of AI prototyping tools, not comparisons on what product teams actually need, like designing from an existing design system, importing current designs or handing work to a coding agent. That made it harder for AI to explain where Magic Patterns fits, especially for product managers, one of its core buyers. On product manager workflow questions, it appeared in just 6.4% of answers.

Action

The Prompting Company benchmarked how ChatGPT, Claude, Gemini and Perplexity answered the questions Magic Patterns' buyers ask, then built the answers those assistants were missing.

The work focused on making Magic Patterns easy for AI to recommend for the jobs product teams actually do, from designing with an existing design system to handing screens to a coding agent.

• Tracked 100+ buyer questions across ChatGPT, Claude, Gemini and Perplexity, and agreed a priority set with Magic Patterns' team. When MagicOS launched, the question set was audited again to match it.
• Published the answer pages on magicpatterns.com, so each citation builds Magic Patterns' own authority.
• Wrote 300+ pages, each answering a specific buyer question in Magic Patterns' voice. 9 in 10 have already been cited by AI.
• Wrote head-to-head pages against Claude Design and Figma, two of the tools buyers weigh Magic Patterns against.
• Got the answer pages' sitemap listed in robots.txt so AI crawlers can find them.
• Twice a week, the team reviewed where Magic Patterns was still missing, and new content followed those gaps, including pages on MagicOS after its October launch.

The goal was for assistants to describe Magic Patterns the way its own team does: as the AI design tool for product teams.

Shift

After the work, AI assistants started treating Magic Patterns as a first-choice AI design tool for product teams. On unbranded questions, it now leads Figma 48% to 44% over the 30 days to Nov 12.

The bigger change was in positioning. On unbranded questions from product teams, the ones that don't mention Magic Patterns by name, it now comes up in more than half of answers in these categories:

Across every question tracked for Magic Patterns, those category-level gains added up to Magic Patterns overtaking Figma for the top spot.

Magic Patterns is now the top answer on Claude and Perplexity, #2 on ChatGPT and Gemini, and leads 5 of the 11 categories built on unbranded questions. 99% of cited pages were still being cited over the last four weeks.

What's next

Product teams increasingly start their search for an AI design tool inside an AI assistant. That makes it important to be described accurately when assistants compare tools and recommend one.

For Magic Patterns, the goal now is to make MagicOS and its product-team strengths just as easy for AI assistants to understand and explain, especially as buyers weigh it against Claude Design. The next gap to close is use-case questions, like which tool lets a team compare many generated screens side by side or stay in control of the canvas.