Figma Scores a Perfect 100 in AI Recommendations. Here Is Who Else Made the List.

{
  "title": "The $26B Design Tool Lost to an Open-Source Underdog in AI Recommendations",
  "meta_description": "We tracked 5 design tools across 7 AI models for 4 days. Figma was untouchable. But the real story? Penpot outscored Canva — by a mile.",
  "tags": ["design", "ai", "ux", "devto", "webdev"],
  "body_markdown": "We ran 7 AI models through the same design tool questions for 4 consecutive days.\n\nI expected Figma to win. I did not expect what happened to Canva.\n\n---\n\n## The Setup\n\nWe tracked five design tools — Figma, Canva, Penpot, Framer, and Miro — across GPT-4o, Claude, Gemini, Perplexity, and three others, asking variations of \"what design tool should I use\" and related prompts. We used our monitoring platform, [AIAttention.ai](https://aiattention.ai), to compute an AI Attention Score (AAS) for each brand: 0 means invisible, 100 means the AI thinks of you first, always.\n\nFour days of data. Seven models. Here's what we found.\n\n---\n\n## Figma Is Untouchable\n\nFigma scored **AAS 100.00**.\n\nNot 98. Not 99.5. One hundred.\n\nEvery model. Every prompt variation. First position, every time. That's not dominance — that's a monopoly on AI memory. When an AI model thinks \"design tool,\" Figma isn't one answer among many. It's the answer.\n\nThis is what category ownership looks like in the age of AI recommendations.\n\n---\n\n## The Upset Nobody Saw Coming\n\nHere's where it gets interesting.\n\nPenpot — an open-source design tool built by a small Spanish startup — scored **AAS 47.14**, appearing in **5 of 7 models**.\n\nCanva scored **AAS 42.45**.\n\n**Penpot outscored Canva.**\n\nLet that sit for a second. Canva is valued at **$26 billion**. It has 185 million users. It spends aggressively on marketing. Penpot has a fraction of that budget, a fraction of those users, and — until this experiment — I assumed it was a footnote.\n\nIn AI recommendations, Penpot ranks higher.\n\n---\n\n## Why This Happened\n\nThe gap isn't huge — 47.14 vs 42.45, about **5 points** — but the pattern held across four days. This wasn't noise.\n\nMy theory: AI models are trained heavily on developer and design community content. Penpot gets disproportionate coverage on GitHub, Hacker News, Reddit's design communities, and tech blogs. It's the tool people *write about* — open-source launches, self-hosting guides, comparisons to Figma.\n\nCanva gets coverage too, but it skews toward mainstream media and marketing blogs. The audiences AI models appear to weight most heavily — technical writers, design bloggers, developer communities — talk about Penpot more than their traffic numbers would suggest.\n\nCanva's massive brand awareness hasn't fully translated into AI recommendation visibility. Yet.\n\n---\n\n## The Miro Problem\n\nMiro only appeared in **3 of 7 models**. AAS: **31.96**.\n\nThat's not just a low score — it's a categorization problem. When AI models respond to \"design tool\" questions, they're pulling from a mental model of what \"design\" means. Miro lives in the collaboration/whiteboard space. It's not wrong — Miro is great — but AI doesn't consistently file it under \"design tools.\"\n\nIf you're at Miro and optimizing for AI visibility, you have a framing problem, not a quality problem.\n\n---\n\n## Final Scores, Day 4\n\n- **Figma** — AAS 100.00 (7/7 models)\n- **Penpot** — AAS 47.14 (5/7 models)\n- **Canva** — AAS 42.45 (5/7 models)\n- **Framer** — AAS 40.14 (5/7 models)\n- **Miro** — AAS 31.96 (3/7 models)\n\nFigma is in a class of its own. The middle tier (Penpot, Canva, Framer) is surprisingly tight — less than 7 points separating all three. Miro is playing a different game.\n\n---\n\n## The Takeaway\n\nAI recommendation visibility doesn't perfectly mirror market share, brand spend, or user counts. It mirrors *how communities talk about you* — the depth and specificity of coverage in the places AI models pay attention to.\n\nPenpot's open-source positioning, its developer-adjacent community, and the richness of technical content written about it gave it an edge over a $26B competitor.\n\nThat's either encouraging or alarming depending on which side of the table you sit on.\n\n---\n\n**What patterns are you seeing with AI recommendations in your category? Are the market leaders always winning — or are there upsets like this one?**"
}