AI Mental Health – Who’s Winning The SEO Race?

As someone who’s passionate about and interested in SEO, mental health, and AI, I wanted to explore the AI for mental health industry and find out who the winners and losers are in the topic space. Like many AI-related industries, it’s still in its infancy.

I think it’s interesting to see how this is playing out at an early stage and how each brand is set up for success moving forward. In this article I’ve tried to combine hard data with my opinions.

And at the end we’ll look at what I would do for each brand if I ran their SEO. Subtle hint: that would be my dream job (ditto working for a psychedelic biotech like Compass Pathways or Definium).

At A Glance

• AI therapy search interest peaked in August 2025 and has fallen since, while searches for traditional therapy are up 13% in the last 12 months. The decline is specific to AI, not mental health in general.
• On Google, small SEO-savvy apps win. Abby, Ash and Feelbetterbot rank on page one, while better-known names like Wysa and Woebot don’t. Apps still hold only about a third of page-one results.
• AI assistants recommend different apps. Wysa is named in 73% of AI recommendation answers despite not ranking on page one for core searches. Abby, a Google winner, is named in just 10%.
• AI assistants still recommend apps that have closed. 42% of AI recommendation answers name Woebot or Youper, both of which have shut down. Claude (71%) and Gemini (60%) do it most.
• AI cites different sources from Google. Review sites, the APA and NIH research dominate. Reddit holds 17% of page-one results but only 0.5% of AI citations.
• Ash is the all-rounder and Feelbetterbot is the riser. Ash is the only app that ranks on Google for both app and information searches and is also widely recommended in AI answers. Feelbetterbot went from 10 to 170 top-10 keywords in a year.

Data Sources And Methodology

• Ahrefs (US): search volume, two years of search history, SERP features and Google rankings. Rankings count top-10 keywords only.
• DataForSEO (November 2026): 27 questions about AI therapy apps, AI therapy and information, each asked three times on ChatGPT, Google AI Mode, Perplexity, Gemini and Claude (405 answers). Brand mention rates use the 240 answers to the 16 non-branded recommendation questions.
• Projections: “next 12 months” assumes searches stay at their average for the last six months. This is my own projection, not an Ahrefs forecast.
• What’s left out: Replika (an AI companion, not a mental health app), therapy platforms adding AI features (such as Talkspace) from the brand tables, and short-lived news-driven keywords. I haven’t used any traffic estimates.

Topic Overview

Before we get into who’s ranking for what, let’s break down this topic so we understand what’s within it.

Sub Topics

Broadly speaking we can bucket these subtopics into four areas:

1. AI therapy
2. Apps
3. Information
4. AI tools for therapists

The caveat with AI tools for therapists is that the people searching could be mental health professionals or clinicians looking for tools such as administration or note-taking software rather than mental health AI tools.

I also want to point out that the line between AI therapy and apps is quite blurry because apps tend to be about people searching for things like “best AI therapy app.” AI therapy is more about the concept of talking to an AI model as a therapist. For example the term “AI therapists”.

Let’s start by looking at the search volume within each subtopic:

AI Mental Health: Search Volume By Sub Topic

The table below shows the total monthly US search volume for the keywords I tracked in each subtopic, according to Ahrefs. Altogether that’s around 63,600 searches a month.

• People search for the idea more than the products. Core “AI therapist” and “AI therapy” searches are the biggest single topic, at around 14,600 a month.
• Branded app searches come second at around 11,000 a month, but almost a third of that is for Woebot, which closed its consumer app in June 2025.
• ChatGPT gets its own slice. Around 4,400 searches a month are about using ChatGPT for therapy, rather than a dedicated app.
• Grouped together, AI therapy makes up about 42% of searches, apps about 36%, information and debate about 11%, and tools for therapists about 10%.

Total US search volume by topic

Topic	Monthly searches	Chart
AI therapist / AI therapy (core)	14,590
Branded AI mental health apps	10,990
Mental health apps (adjacent)	6,800
AI tools for therapists	6,470
AI & mental health (informational)	5,680
ChatGPT for therapy	4,430
Free AI therapy	3,860
Best AI therapy apps	3,370
Mental health & therapy chatbots	2,300
AI journaling	1,950
Specialist AI support	1,810
AI therapy debate & safety	1,320

AI Mental Health – Which Brands Are Falling & Rising?

Next, the brands themselves. The first table shows how many people search for each AI-first app by name each month. The second compares average monthly brand searches in November to January with August to October, to show who’s gaining interest and who’s losing it.

• Woebot still gets the most brand searches (3,550 a month), even though its consumer app closed in June 2025. People are still looking for it.
• Abby and Ash lead the apps that are still open, with around 2,270 and 1,820 searches a month.
• Feelbetterbot is the breakout. Its brand searches went from 24 to around 700 a month.
• Wysa nearly tripled (276 to 748), although its numbers bounce around a lot from month to month.
• Earkick and Sonia are fading fast, down 81% and 66%. Abby is down 38%.

Brand search volume: AI mental health apps (US)

Brand	Monthly searches	Chart
Woebot	3,550
Abby	2,270
Ash	1,820
Youper‡	1,400
Earkick	700
Wysa	660
Feelbetterbot	350
Sonia	100
Headspace Ebb	70

‡ Youper shut down in September 2026.

Which AI mental health brands are rising and falling in search (US)

Brand	Nov–Jan	Aug–Oct	Change
Rising
Feelbetterbot	24	699	▲ 2,771%
Wysa*	276	748	▲ 171%
Youper‡	1,353	1,509	▲ 12%
Falling
Earkick	1,027	196	▼ 81%
Sonia	129	43	▼ 66%
Headspace Ebb	78	32	▼ 59%
Abby	2,710	1,674	▼ 38%
Woebot†	3,672	3,236	▼ 12%
Ash	1,875	1,675	▼ 11%

Average monthly US searches for each brand’s branded keywords, Nov 2025–Jan 2026 vs Aug–Oct 2026. Source: Ahrefs.
* Wysa’s search volume is volatile month to month.
† Woebot shut down its consumer app in June 2025.
‡ Youper shut down in September 2026.

Search demand is only half the story, so I also looked at how each brand’s website is doing on Google. This table counts the keywords each site ranks for in Google’s top 10, comparing October 2025 with October 2026.

• Feelbetterbot is the standout, going from 10 to 170 top-10 keywords in a year.
• Ash nearly doubled (49 to 92), and Sonia more than doubled from a small base (7 to 16).
• Abby and Earkick took big hits. Abby is down 79% and Earkick 71%, although Earkick still ranks for more keywords than anyone else here.
• The closed apps are fading. Youper fell from 45 to 4 keywords and Woebot is down 39%. Wysa also slipped 24%.

Which AI mental health apps are gaining and losing Google rankings (US)

Brand	Oct 2025	Oct 2026	Change
Gaining
Feelbetterbotfeelbetterbot.com	10	170	▲ 1,600%
Soniasoniahealth.com	7	16	▲ 129%
Ashtalktoash.com	49	92	▲ 88%
Losing
Youper‡youper.ai	45	4	▼ 91%
Abbyabby.gg	696	145	▼ 79%
Earkickearkick.com	1,584	465	▼ 71%
Woebot†woebothealth.com	109	67	▼ 39%
Wysawysa.com	21	16	▼ 24%

Number of keywords each site ranks for in Google’s top 10 (US), October 2025 vs October 2026. Source: Ahrefs.
† Woebot shut down its consumer app in June 2025.
‡ Youper shut down in September 2026.

AI Mental Health – Future Outlook

So far we’ve looked a little bit at the history of AI for mental health in search and also at who’s doing well. What I want to do now is look at the outlook for the future as AI in general is moving super quickly. What are we expecting from the mental health niche?

Firstly we’ll look at the last 12 months for each subtopic, and a projection for the next 12 months (assuming searches stay at their average for the last six months).

AI mental health search topics: last 12 months and outlook (US)

Topic	Last 12 mo	Next 12 mo
AI & mental health (informational)*	▲ 54%	▼ 10%
AI journaling	▲ 15%	● Flat
AI therapy debate & safety*	● Flat	▼ 11%
Best AI therapy apps	● Flat	▼ 8%
Specialist AI support	● Flat	● Flat
Free AI therapy	● Flat	▼ 6%
Mental health apps (adjacent)	▼ 10%	▼ 24%
AI therapist / AI therapy (core)	▼ 11%	● Flat
Mental health & therapy chatbots	▼ 22%	▼ 13%
AI tools for therapists	▼ 28%	▼ 8%
ChatGPT for therapy	▼ 32%	▼ 27%

US search volume for non-branded keywords in each topic. Last 12 months: Nov 2025–Oct 2026 vs the previous 12 months. Next 12 months: projected if searches stay at their average for the last six months. “Flat” = within ±5%. Source: Ahrefs.
* Excludes short-lived news-driven keywords that spiked briefly and then dropped to zero.

What I find really interesting is that the next 12 months look flat or declining pretty much across the board. My gut feel is:

1. Like many AI things people are more aware of what AI products and services are available.
2. Possibly quite a few people feel a sense of doubt when it comes to using AI for mental health.
3. We also need to consider this as part of the wider perception of AI in general, which in recent months has become more negative. For example people’s perception of how it is affecting the economy, the job market, and cybersecurity.

Given that I am very interested in mental health in general, I was a bit cautious about the projected drop in interest in AI mental health (because I know that, in general, people are having worse mental health and there’s more demand). To check my gut feel, I pulled data from Ahrefs, looking at the rise and fall and projections of non-AI mental health terms.

This data, at least at a high level, supported my theory. Searches for 17 general therapy terms (like “therapist near me”, “therapy” and “counseling”) grew 13% in the last 12 months, while AI mental health searches fell 8%. Looking ahead, traditional therapy searches stay flat while AI searches drop another 7%. Bear in mind these are Ahrefs search volume estimates, and the projection is my own (it assumes searches stay at their average for the last six months), so they’re not concrete figures.

AI therapy searches peaked in 2025. Therapy searches kept growing.

US monthly search volume, indexed so January 2025 = 100

AI mental health & therapyTraditional therapy & mental health

Line chart: AI mental health searches rose to 180 in August 2025, then fell to 111 by October 2026, while traditional therapy searches rose steadily to 131.

80 100 120 140 160 180 Jan 2025 Apr 2025 Jul 2025 Oct 2025 Jan 2026 Apr 2026 Jul 2026 Oct 2026 Woebot app closes Illinois bans AI therapy Peak: 180 111 131

80 100 120 140 160 180 Jan 25 Jul 25 Jan 26 Jul 26 Woebot closes Illinois ban Peak: 180 111 131

Show data table

Month	AI	Traditional
Jan 2025	100	100
Feb 2025	104	104
Mar 2025	116	108
Apr 2025	124	109
May 2025	140	110
Jun 2025	150	114
Jul 2025	168	119
Aug 2025	180	123
Sep 2025	179	125
Oct 2025	162	125
Nov 2025	148	126
Dec 2025	138	125
Jan 2026	141	125
Feb 2026	138	126
Mar 2026	138	128
Apr 2026	130	126
May 2026	130	126
Jun 2026	126	127
Jul 2026	125	128
Aug 2026	119	129
Sep 2026	114	130
Oct 2026	111	131

Three-month rolling average of US Google search volume, indexed to January 2025 (November 2024 to January 2025 average) = 100. AI: all non-branded AI mental health and therapy keywords with 90+ monthly searches, excluding short-lived news spikes. Traditional: 17 general terms such as “therapist near me”, “therapy” and “counseling”. Source: Ahrefs.

Who Owns Page One Of Google?

This table groups the 59 page-one results for eight core searches (like “ai therapist”, “mental health chatbot” and “will ai replace therapists”) by the type of site that ranks.

• Apps only get about a third of page one. AI therapy apps hold 32% of the results.
• Forums are the next biggest group. Reddit, Quora and LinkedIn take 17%.
• Trusted sources take a big share. Research, publishers, health authorities and therapy providers make up another 40% between them, which is what you’d expect for a health topic where Google leans hard on trust.

Who ranks on page one for AI therapy searches (US)

Type of site	Share of results	Chart
AI therapy apps	32%
Forums & community (Reddit, Quora, LinkedIn)	17%
Research & academic	14%
Publishers & media	10%
Health authorities & professional bodies	8%
Online therapy providers	8%
Other (vendor blogs, app store, video, clinician tools)	10%

Share of 59 page-one Google results across eight searches: “ai therapist”, “ai therapy”, “ai for mental health”, “best ai therapist”, “ai therapist free”, “mental health chatbot”, “chatgpt therapy” and “will ai replace therapists”. Source: Ahrefs.

AI Overviews For AI Mental Health

As with most topics now, AI Overviews are almost everywhere in this niche, appearing on 92% of the searches I looked at. The only real exception is tools for therapists, at 57%.

How often Google shows an AI Overview for AI mental health searches (US)

Search type	AI Overview	Chart
Apps (best AI therapy apps, mental health apps, AI journaling)	100%
Information (AI & mental health, debate & safety)	100%
AI therapy (AI therapist, free AI therapy, chatbots, ChatGPT)	94%
Tools for therapists (AI therapy notes, documentation)	57%

Percentage of non-branded keywords (90+ monthly US searches) whose Google results include an AI Overview. Overall: 92% (70 of 76 keywords). Keywords without recent SERP data were excluded. Source: Ahrefs.

GEO / AI SEO – Who’s Winning For AI Mental Health

So far we’ve looked at how each company is performing in terms of regular SEO so we can look at this holistically. Let’s now look at GEO or AI SEO, depending on what you want to call it.

Using the DataForSEO API, I ran various prompts across different LLMs to see how brands were showing up.

AI assistants and Google favour different AI therapy apps

Brand	AI answers	Google p.1
Wysa	73%	No
Earkick	45%	No
Ash	39%	Yes
Woebot†	37%	No
Youper‡	30%	No
Sonia	20%	Yes
Feelbetterbot	11%	Yes
Abby	10%	Yes

AI answers: share of 240 answers that mention the brand, from 16 non-branded app and AI therapy questions asked three times each on ChatGPT, Google AI Mode, Perplexity, Gemini and Claude (US, November 2026). Google p.1: whether the brand’s own site ranks in the top 10 for at least one of eight core AI therapy searches. Sources: DataForSEO, Ahrefs.
† Woebot shut down its consumer app in June 2025.
‡ Youper shut down in September 2026.

One thing that really jumped out at me is that AI assistants are still recommending apps that no longer exist. Woebot shut down its consumer app in June 2025 and Youper closed in September 2026, yet 42% of the recommendation answers I collected named at least one of them.

How often AI assistants recommend AI therapy apps that have closed

Platform	Answers naming Woebot or Youper	Chart
Claude	71%
Gemini	60%
Google AI Mode	33%
Perplexity	29%
ChatGPT	19%

Share of 48 answers per platform that name Woebot or Youper, from 16 non-branded app and AI therapy questions asked three times each (US, November 2026). All platforms: 42% (102 of 240 answers). Woebot shut down its consumer app in June 2025; Youper shut down in September 2026. Source: DataForSEO.

Woebot is the bigger problem: it was named in 37% of answers, well over a year after it closed. Claude and Gemini were by far the worst for this. For anyone working on GEO, it’s a good reminder that AI answers can lag a long way behind reality, and that what an AI assistant says about your brand (or your competitors) may be badly out of date.

The next table is a summary of which sources were cited across these prompts. It was interesting to see a mix of very authoritative domains like psychology.com and apa.org with some of the apps we’ve looked at in this article.

Wysa is the most-cited app website, appearing in 11% of answers. But it’s even further ahead on brand mentions: as the table above shows, it’s named in 73% of recommendation answers.

What’s also striking is how little AI citations overlap with Google. Only 20% of citations come from sites that rank on page one for the same topics, and Reddit, which takes 17% of page-one results, makes up just 0.5% of AI citations.

The sources AI assistants cite for AI therapy questions

Website	Answers citing it	Chart
psychology.com	42%
apa.org	25%
choosingtherapy.com	16%
ncbi.nlm.nih.gov (PMC)	15%
wysa.com	11%
talkspace.com	9%
lovon.app	9%
ilty.co	9%
abby.gg	7%
feelbetterbot.com	7%

Share of 324 answers from ChatGPT, Google AI Mode, Perplexity and Claude that cite each website (Gemini doesn’t return citations). Only 20% of all citations come from sites that rank on Google’s page one for the same topics, and Reddit, which takes 17% of page-one results, makes up just 0.5% of AI citations. 27 questions, each asked three times per platform, US, November 2026. Source: DataForSEO.

AI Mental Health Brands – Who Is Competing For What

To pull everything together, the table below shows where each brand (plus a few of the non-app sites) is competing. The first three columns are Google: whether the site ranks on page one for app searches (like “ai therapist” and “best ai therapist”), information and safety searches (like “will ai replace therapists”) and “chatgpt therapy”. The last column shows how often the brand is named or cited in AI answers across all five platforms.

• Ash is the all-rounder. It’s the only AI app that ranks for both app and information searches on Google and is also strong in AI answers.
• Wysa wins without Google. It’s nowhere on page one for these searches, yet it’s the app AI assistants recommend most.
• Abby and Feelbetterbot are Google specialists. Abby is strong on page one but weak in AI answers, and Feelbetterbot mostly gets named for “free” and “no sign-up” questions.
• Sonia has a foothold in both. It ranks for information searches and gets some AI mentions.
• Reddit and the authorities go different ways. Reddit is all over page one but rarely cited by AI, while the APA and NIH are strong in both.

Who competes where in AI mental health search

Player	Google apps	Google info	Google ChatGPT	AI answers
Ash	Yes	Yes	–	Strong
Wysa	–	–	–	Strong
Abby	Strong	–	–	Weak
Feelbetterbot	Yes	–	–	Niche*
Sonia	–	Yes	–	Some
Talkspace	Yes	–	Yes	Some
Reddit	Yes	Yes	Yes	Rare
APA / NIH	Yes	Yes	Yes	Strong

Google: whether the site ranks on page one (US) for searches about finding an AI therapy app (“ai therapist”, “ai therapy”, “best ai therapist”, “ai therapist free”), information and safety (“ai for mental health”, “will ai replace therapists”) or “chatgpt therapy”. AI answers: how often the brand is named or the site is cited across 405 answers from ChatGPT, Google AI Mode, Perplexity, Gemini and Claude, November 2026. Sources: Ahrefs, DataForSEO.
* Feelbetterbot is named mainly for “free” and “no sign-up” questions.

What I’d Do If I Ran SEO For..

This is going to be super high level as I’m not going to sit here creating a full SEO strategy (with that said, I’d love to work for these kinds of companies so if you’re hiring…)

Here are a few things that come to mind..

Sonia Health

• AI Imagery
  • IMO the AI images (photos) used aren’t doing the site any favours. Yes, they’re branded in terms of colours, and yes I get the point that the platform is AI, but I don’t think people will really resonate with them
• Consent
  • The banner being small and in the bottom right means some traffic potentially isn’t recorded in GA4 (i.e. because the visitor can easily use the site without interacting with the banner)
• General
  • ‘HIPAA Compliant’ – add a tooltip explaining what this means
  • Improve schema (e.g. homepage is currently ‘FAQPage’)
  • Logo in middle of top navigation is confusing – remove ‘Home’ in nav and replace with slightly larger logo
  • Remove the ‘Coming soon’ link in navigation
  • Remove any ‘Coming soon’ links in the blog hub
  • Text too small in some images (desktop and mobile) – rethink this for accessibility
• EEAT
  • Have a dedicated author bio for each team member
• SEO Basics
  • Improve internal linking – looks like blog pages don’t do any at all
  • Title tags – the /enterprise page tags are branded – probably best to make a hybrid as the search volume for something like ‘Sonia for Employers’ is going to be super low
• Research Content
  • Rather than just link out to the papers, have a blog post which summarises each one and link out from there
  • For the /science hub page, have some summaries of everything
• Content Strategy
  • The site is kicking ass for some big main terms, but is missing an opportunity for supporting non-commercial/guide content
  • Create comparison page/s – Sonia AI vs competitors
• Digital PR / Link Acquisition
  • There’s a reasonable base here but huge opportunity to improve this
  • Almost all links are to the home page
  • Non-commercial content additions will help but like I’ve said for other brands on this page, there is HUGE scope for passive and proactive digital PR
• Product
  • I’m not seeing any mention of pricing, just a link to the free download
  • As a possible customer, I want to know that before I download it

Earkick

I really have no idea what’s going on with this site and product – everything seems to be on the blog. subdomain and what I assume is their product (‘Chat With Panda’) – that subdomain doesn’t even open.

So I’d want to fix some basics before even thinking about anything else!

Abby

Abby lost a LOT of keywords in the last year, although since the last Google update they seem to be starting to recover

• Cookies / Consent Missing
  • I assume because they’re US based, they’ve neglected this – but from the UK, I see no consent banner or options, despite the site using GA4, GAads etc.
  • This is a pretty big issue – it’s a potential GDPR breach, and not a good look in the health niche
• EEAT
  • If possible, attribute an author/editor to each blog article rather than just ‘The Abby Team’ and link to a proper author page
• Content
  • Pad out some of the blog articles – there’s a ‘Why Do I Cry So Easily? 9 Reasons’ page. It’s OK but pretty short. If this was made purely for a Google Discover type platform I’m OK with it but otherwise feels like content for the sake of content
  • Be consistently declarative. E.g. in FAQs, some have this format – Question > Yes/no > Explanation. Some, like ‘Does Abby replace therapy’ don’t. If a question has a clear yes/no answer, give that immediately then the explanation, if needed.
• Product
  • Allow signup by Google as well as Apple
  • Be clear about pricing – e.g. there’s an FAQ ‘Is it free?’ but it has quite a woolly answer ‘Your first conversation is free. No credit card, no appointment. We think that’s all it will take. If you want to keep going, every plan starts with a 7-day free trial.’ I’d rather you’re just open about the pricing.
• Digital PR/ Link Acquisition
  • Same story as most other brands on this page – a decent base with some great referring domains, but a consistent digital PR strategy would do wonders long term

Feel Better Bot

• Cookies / Consent Missing
  • Same situation as Abby
• EEAT
  • What I find off-putting as a potential user is that there is nothing about who or what is behind the product
  • In the privacy page it says “FeelBetterBot is operated by Better Bot LLC, a New Mexico limited liability company. If you have privacy questions, use the feedback form on the main page.” – which isn’t at all reassuring
  • There’s no contact page, no email, no About us
  • Personally on this basis I wouldn’t use the product
  • I did some more research and it looks like a one-person project. The intent behind it seems legit, but I think the founder is missing out big time by making it so faceless

Wysa

• Consent Banner – Change Position And Size
  • The current one is too small and hides away (on desktop) in the bottom left corner
  • As a result, I’d imagine quite a lot of GA4 data isn’t being collated as it’s too easy for the user to move around the site without accepting or declining
  • I also feel the ‘Decline All’ button is a bit dark-patterny. I get the idea of bringing ‘Accept All’ to the visual forefront, but greying the Decline All button and border colour is a bit iffy (IMO)
• Improve Contact Page
  • The ATF here is too big, so I have to scroll right down to see the form or contact details
  • Then there are multiple email addresses, and the contact form being in one column next to contact details isn’t easy on the eye
  • I’d also like to try streamlining the contact form – there are 12 mandatory form fields for the demo form
  • I’d have a standalone ‘Demo’ page (linked to from the ‘Book Demo’ button in nav). This would house the demo request form as well as FAQs about the demo, SLA for a reply etc. That’s to separate the very specific commercial intent of getting a demo, with more general contact needs
• Move blog from Wysa.io to Wysa.com
  • I’m assuming there was a reason (possibly CMS) but it’s a very strange experience clicking ‘Blog’ in the .com nav only to be taken to Blogs.Wysa.io – especially as the root domain Wysa.io 301s to Wysa.com
• Improve UX on Wise Up Homepage
  • Super cool page with links to their magazine, but two CTAs with two descriptions are pushing the magazine issues down too far

Ash (Talk To Ash)

• CRO
  • Update testimonials on homepage (last one was July 2025 – this makes me, as a possible customer, think “this product isn’t updated”)
  • The ‘Download now’ CTA links to a page with a QR code – preferable would be direct to the page I can download it on
  • The page isn’t branded either, so the journey feels disjointed – it almost feels like the app. subdomain could be a third party
• General
  • In the testimonial block, ditch the pink highlights of the key points – it made me think they’re hyperlinks
  • In the block below ‘what people are saying’, make the M and K black font
  • I’m not a fan of the ‘Read the Original’ outlinks on /news items. I’d suggest putting these at the end of the article as a text link rather than a washed-out button.
• EEAT
  • Have a proper contact page not just email addresses in the footer
  • Ensure all team members have a proper bio (Daniel doesn’t, for example)
• Schema
  • Doesn’t look like this is being used
• Content
  • ‘Research’ in top navigation takes me to /clinical-research. As it’s prominent in the navigation I’m assuming this page is significant
  • However, it’s not structured well. There is a massive hero image, a logo scroller, then very large staff images before I get to links to research pages
  • This a) makes the journey to research content difficult, and b) misses an opportunity to highlight key findings
  • There are quite a few blog posts which are just summaries of YouTube videos. You could flesh these out by adding a transcript, or bringing them to life more. Also, embed the YouTube video rather than just linking out
  • Be declarative in FAQs – e.g. ‘Is Ash safe and private’ – the answer should begin with ‘Yes, Ash is safe and private..’
• SEO Basics
  • All blog articles should have internal links added
• Digital PR / Link Acquisition
  • 301 this page as it’s 404 and has a nice link from Reason.com

Woebot

Woebot shut down its consumer app in June 2025 and now focuses on enterprise, so these points are about the site that’s still live.

• Have a proper blog/news strategy
  • ‘News’ is just some links out to external mentions and interviews and seems out of date
  • ‘Blog’ is nicely designed but the posts are basically just summary pages with a YouTube embed (last one was 2025). I’d be willing to bet a bunch of those aren’t indexed and are stuck in crawled or discovered not yet indexed
• Call To Action Strategy
  • Other than a ‘Join Interest List’ button it’s not obvious what I’m supposed to do on the page
  • The form it takes me to is a Google Form – this doesn’t look professional
  • There’s also a ‘Sign up to Woebot Health updates’ link in the footer. It’s basically hidden there, I can almost guarantee it’s not clicked on. And again it links to a Google form
• GTM/GA4
  • From what I can see in dev tools/network, not much is being tracked above and beyond the out of the box GA4 stuff
• SEO Basics
  • Improve title tags
  • Add internal linking
• Navigation
  • I’m assuming there’s some reason so much stuff isn’t in the nav
  • Header and footer navs need planning out properly and redoing
  • Add internal site search function to header
• Digital PR / Link Acquisition
  • I’ve only had a cursory look through Ahrefs but although Woebot has a great backlink profile (helped by being early in the market), there don’t seem to be many new backlinks
  • As with all mental health AI apps and services, there are tons of opportunities to continue to build authority

So, Who’s Winning?

It depends on where you look.

On Google, the smaller apps that are good at SEO are ahead. Abby, Ash and Feelbetterbot all rank on page one for core searches, and Feelbetterbot has grown faster than anyone else over the past year. AI answers tell a different story. Wysa is named in nearly three-quarters of recommendation answers even though it doesn’t rank on page one, while Abby barely features.

If I had to pick the brand best placed overall, it’s Ash. It’s the only app that ranks on Google for both app and information searches and is also regularly recommended by AI assistants.

The picture for the category as a whole is less rosy. Interest in AI therapy peaked in summer 2025 and has been falling since, while searches for traditional therapy keep growing. AI assistants are also still recommending apps that closed over a year ago, which shows how young this space still is. The brands that get the basics right now (trust signals, consent, clear pricing, and content that earns links and citations) will be best placed if demand picks back up.

Disclosure – Use Of AI When Creating This Article

• Claude was used for:
  • Connecting to the Ahrefs MCP
  • Connecting to the DataForSEO API
  • Collating data into Google Sheets
  • Getting data into HTML charts and giving me the code
  • Drafting the At A Glance summary, section explainers and conclusion, which I then reviewed and edited
• Me:
  • Everything else