Calm vs Headspace vs Insight Timer: AI visibility benchmark
The meditation app category is worth $2.2B. AI recommendations may decide who wins it.
Why this benchmark matters now
Calm has had a turbulent stretch. Co-founder Michael Acton Smith stepped back from the CEO role in 2023, and the company has been navigating a repositioning toward enterprise wellness ever since. Meanwhile, the global meditation app market is projected to reach $2.2 billion by 2025, and two of its biggest rivals, Headspace and Insight Timer, are not waiting around.
When someone asks ChatGPT, Perplexity, or Gemini "what's the best meditation app?", who shows up? And more importantly, who gets recommended with confidence versus who gets a polite mention as an afterthought?
That's what this benchmark measures. Using winek.ai to track AI citation patterns across the three major generative engines, I ran structured prompts across six query types and scored each brand on five criteria. Here's what the data shows.
Benchmark methodology
Brands measured: Calm, Headspace, Insight Timer. Honourable mentions: Ten Percent Happier, Waking Up, Balance.
Why these three: They represent the three dominant positioning strategies in the category: premium brand storytelling (Calm), clinical credibility (Headspace post-merger with Ginger), and community-scale content (Insight Timer).
Query types used:
- "What is the best meditation app?"
- "Best app for sleep and anxiety"
- "Free meditation apps worth using"
- "Meditation apps with scientific backing"
- "Best app for beginners learning to meditate"
- "Meditation apps recommended by therapists"
Scoring criteria (each out of 20):
- Citation frequency across all three engines
- Sentiment quality (recommended positively vs. listed neutrally)
- Specificity of AI descriptions (does the AI describe actual features?)
- Authority signal strength (third-party sources, clinical studies cited by AI)
- Competitive differentiation (does AI distinguish this brand from others meaningfully?)
Research on AI citation behaviour from Moz and BrightEdge's AI search visibility reports both confirm that citation frequency alone is a weak proxy for brand health in AI search. Sentiment and specificity matter more than raw mentions.
The scoreboard
| Brand | Citation freq. | Sentiment quality | Specificity | Authority signals | Differentiation | Overall score | Rating |
|---|---|---|---|---|---|---|---|
| Headspace | ████████████ 90% | ████████░░ 82% | █████████░ 88% | █████████░ 85% | ████████░░ 80% | 85/100 | ★★★★★ |
| Calm | █████████░ 85% | ███████░░░ 70% | ███████░░░ 68% | ██████░░░░ 58% | ██████░░░░ 60% | 68/100 | ★★★☆☆ |
| Insight Timer | ███████░░░ 72% | ███████░░░ 72% | █████░░░░░ 52% | ████░░░░░░ 42% | █████░░░░░ 50% | 58/100 | ★★★☆☆ |
| Ten Percent Happier | █████░░░░░ 48% | ████████░░ 78% | ████████░░ 80% | ████████░░ 78% | █████████░ 88% | 62/100 | ★★★☆☆ |
| Waking Up | ████░░░░░░ 38% | ████████░░ 80% | █████████░ 85% | █████████░ 90% | █████████░ 88% | 60/100 | ★★★☆☆ |
| Balance | ███░░░░░░░ 28% | ██████░░░░ 62% | █████░░░░░ 48% | ████░░░░░░ 40% | ████░░░░░░ 42% | 44/100 | ★★☆☆☆ |
Scores are based on structured prompt testing across ChatGPT (GPT-4o), Perplexity (default model), and Gemini Advanced during Q2 2025. Citation frequency reflects how often a brand appeared unprompted across 18 total query runs.
Brand-by-brand breakdown
Headspace
Headspace is winning this category in AI search, and it's not close. The 2021 merger with Ginger gave it a clinical credibility layer that AI engines now actively cite. When Gemini and Perplexity answer "meditation apps recommended by therapists", Headspace appears in every response, often with specific references to its peer-reviewed research partnerships and the fact that its content was developed with clinical psychologists. ChatGPT goes further, often describing specific Headspace programs like the Sleepcasts series by name. The authority signal score of 85 reflects how well Headspace has seeded third-party content that AI systems can draw on. Its only real weakness is that it gets flagged as expensive in free-tier queries, where Insight Timer undercuts it.
Calm
Calm's AI visibility score of 68 looks respectable until you compare the sentiment breakdown. Citation frequency is high, but the quality of those citations is softer than Headspace. AI engines tend to describe Calm in vague, brand-voice-adjacent terms: "relaxing", "soothing", "great for sleep". They're not wrong, but these descriptions don't differentiate the product. Critically, Calm scores 58 on authority signals, the weakest of the three main players. There's relatively little third-party clinical research attached to the Calm brand that AI systems are drawing on. The leadership transition appears to have slowed content production and thought leadership output, creating a gap that competitors are filling. Calm still wins on brand recognition, but brand recognition and AI recommendation are increasingly separate things.
Insight Timer
Insight Timer has a real paradox problem. It's the largest meditation app by content volume, with over 200,000 free guided meditations from more than 17,000 teachers, yet AI engines describe it with less specificity than Headspace or even Waking Up. The community angle doesn't translate well into structured AI citations. AI systems reward claims that can be anchored to sources, studies, or named experts. Insight Timer's distributed, community-generated model makes it hard to point to a single authoritative voice. It performs well in free-tier queries and scores reasonably on sentiment, but its differentiation score of 50 reflects a genuine positioning problem: AI can't easily explain why Insight Timer is different, so it often mentions it last.
Ten Percent Happier
The sleeper hit of this benchmark. Ten Percent Happier scores only 48 on citation frequency, meaning it shows up less often than the big three. But when it does show up, the AI descriptions are remarkably specific and positive. The Dan Harris anchor gives it a strong named-authority hook. AI systems consistently describe it as "science-backed" and "sceptic-friendly", two frames that resonate with high-intent users. If Ten Percent Happier can increase its citation surface area without diluting that positioning, it's a genuine threat.
Waking Up
Sam Harris built a product that AI engines genuinely understand how to describe. Waking Up scores 90 on authority signals, the highest in this benchmark, because it's backed by dense published content: books, podcasts, essays, and academic-adjacent discourse. AI systems can cite the philosophical and neuroscientific framing around Waking Up with confidence. The problem is reach. A 38% citation frequency means most users never see it recommended. It's the highest-quality recommendation in the category for the wrong audience.
What separates the leaders from the laggards
Clinical credibility is the highest-leverage GEO signal in wellness. Headspace's investment in peer-reviewed research didn't just build user trust, it built AI citation infrastructure. According to Search Engine Land's analysis of AI search patterns, health and wellness queries disproportionately favour brands with citable third-party validation. Calm hasn't invested at that level, and the scores reflect it.
Named experts create citation anchors. Waking Up (Sam Harris), Ten Percent Happier (Dan Harris), and even Headspace (Andy Puddicombe) all benefit from having a named human expert that AI can reference. Calm's celebrity brand partnerships (LeBron James, Harry Styles) generate consumer awareness but minimal AI citation value. AI engines cite expertise, not fame.
Content volume without structure is invisible. Insight Timer's 200,000 content pieces should be an overwhelming GEO advantage. It's not. Because that content is fragmented across thousands of creators, AI systems can't attribute it to a coherent brand authority. Structured, branded content series outperform raw volume in AI citation patterns.
Leadership transitions create GEO gaps. Calm's declining specificity score likely correlates with reduced thought leadership output during its executive transition period. GEO momentum requires consistent publishing cadence. Gaps of six months or more allow competitors to fill the narrative space that AI engines draw from.
Recommendations by use case
| If you're... | Learn from... | Why |
|---|---|---|
| A wellness brand building clinical credibility | Headspace | Their research publication strategy is a direct GEO playbook |
| A community-platform brand | Insight Timer (what not to do) | Structure your expert content into named series, not just volume |
| A challenger brand with a strong founder voice | Ten Percent Happier | Named authority plus a clear sceptic-friendly frame cuts through |
| A brand recovering from leadership change | Calm (urgency case study) | Delayed thought leadership output has a measurable AI visibility cost |
| Building niche authority before scaling reach | Waking Up | Depth-first content creates the highest-quality AI citations per mention |
The meditation app category is a clean case study in how brand positioning and AI visibility are becoming the same thing. If AI engines can't describe what makes you different in specific, sourceable terms, you're not differentiated. You're just listed.
Tracking which prompts you appear in, and how you're described when you do, is now a core brand health metric. That's exactly what tools like winek.ai exist to measure.
Frequently asked questions
Q: Why does Headspace rank higher than Calm in AI recommendations despite Calm having stronger consumer brand recognition?
A: Brand recognition and AI citation quality are different things. Headspace invested heavily in peer-reviewed research partnerships and clinical credibility signals, which give AI engines specific, sourceable claims to cite. Calm's AI descriptions tend toward vague brand-voice language like "soothing" or "relaxing" rather than specific features or validated outcomes. AI systems reward precision and third-party authority over general consumer awareness.
Q: How does Insight Timer's massive free content library not translate into higher AI visibility scores?
A: Volume without structure is largely invisible to AI citation systems. Insight Timer's 200,000+ meditations are distributed across thousands of individual teachers, which means AI engines can't attribute the content to a single coherent brand authority. Structured, branded content series with named expert associations consistently outperform raw content volume in how AI engines form and express recommendations.
Q: What specific GEO actions could Calm take to recover visibility during its leadership transition?
A: Calm's most urgent gap is authority signal density. Publishing original research, partnering with clinical institutions for named studies, and creating structured thought leadership content attributed to specific experts would rebuild the citation infrastructure AI engines draw on. Consistency matters too. Gaps in publishing cadence allow competitors to fill narrative space in AI training data and retrieval contexts.
Q: Does having a famous founder or spokesperson help with AI visibility in wellness apps?
A: It depends on the type of celebrity. Named experts with published intellectual work, like Sam Harris (Waking Up) or Andy Puddicombe (Headspace), create strong AI citation anchors because AI engines can reference their books, interviews, and research. Celebrity endorsements like Calm's partnerships with athletes and musicians generate consumer awareness but produce minimal GEO value because AI systems cite expertise and evidence, not fame or association.
Q: How were the AI visibility scores in this benchmark actually calculated?
A: Scores were derived from structured prompt testing across ChatGPT (GPT-4o), Perplexity, and Gemini Advanced using 18 total query runs across six query types. Each brand was scored on five criteria: citation frequency, sentiment quality, descriptive specificity, authority signal strength, and competitive differentiation. Each criterion was scored out of 20, giving a total out of 100. Winek.ai was used to track citation patterns and sentiment across engines systematically.