Case study: AI mental-health startup — 28× clicks

AI mental-health startup: 28× clicks

From February through April 2026, I built the complete SEO and GEO strategy with this AI mental-health startup and worked with its team to implement it in full. All planned pages were created and published during the three-month engagement. In Google Search Console, manually corrected non-branded clicks increased from 277 in the first complete 28-day window (February 21-March 20) to 7,767 in the later September 8-October 5 window. Impressions increased from 41,790 to 281,799.

The task

An early-stage AI startup in mental health was targeting the US, UK, Canada and Australia. The engagement needed to build non-branded organic demand in a YMYL category where trust, expert review and careful claims matter.

What I did

I built the complete SEO and GEO strategy: keyword and content-gap analysis, site architecture, page planning, technical recommendations, link-acquisition planning and AI-citation guidance. From February through April 2026, I worked with the startup team to implement the roadmap in full and create and publish every planned page.

The numbers

28-day click growth: 28× · 28-day impression growth: 6.7× · non-branded clicks: 22,503 · non-branded impressions: 1,011,918. The strategy and planned pages were fully implemented by the end of April, and non-branded growth followed. Comparing equal 28-day windows, manually corrected non-branded clicks increased from 277 during February 21-March 20 to 7,767 during September 8-October 5, a 28× increase. Impressions increased from 41,790 to 281,799, or 6.7×. Across February 21-October 5, the corrected totals were 22,503 clicks and 1,011,918 impressions.

How this result is evidenced

Period

Strategy and implementation: February-April 2026. Google Search Console observation window: February 21-October 5, 2026. The first 28-day comparison window (February 21-March 20) falls during implementation; the last (September 8-October 5) shows the growth after the completed rollout.

Baseline

The manually corrected first 28 days produced 277 non-branded clicks and 41,790 impressions, averaging 9.9 clicks and 1,492.5 impressions per day.

Scope

Full SEO and GEO strategy, demand and content-gap analysis, site architecture, technical recommendations, page creation, link-acquisition planning and AI-citation guidance.

Source

Google Search Console Web report. Google's AI-classified Non-branded filter is available for this property from February 21, 2026. I audited all exposed query rows in each comparison window and subtracted visible brand, parent-company and founder variants: 35 clicks and 1,068 impressions from the baseline; 48 clicks and 801 impressions from the final window. Across the full observation period, the correction removed 504 clicks and 7,274 impressions.

Implementation

Dima built the strategy and worked with the startup team to implement it in full during the three-month engagement. All planned pages were created and published by the end of April 2026.

Attribution limits

The timing shows that the growth followed the completed strategy and page rollout. Google Search Console remains observational rather than a controlled experiment, so it cannot isolate the contribution of every workstream. Google's branded-query classification may be imperfect; anonymized queries cannot be audited individually.

How the screenshot reconciles with the case metrics

The screenshot shows Google's AI-classified Non-branded totals before manual cleanup. I removed visible brand, parent-company and founder variants from both equal 28-day windows.

Clicks

GSC filter: 312; 7,815
Visible brand removed: -35; -48
Case metric: 277; 7,767

Impressions

GSC filter: 42,858; 282,600
Visible brand removed: -1,068; -801
Case metric: 41,790; 281,799