Method in brief
- Prompts: 18
- Prompt types: treatment, category, location, persona and bottom of funnel
- Engines: ChatGPT, Claude, Gemini, Perplexity, Grok, Mistral and DeepSeek
- Mode: augmented
- Web search: on
- Runs per prompt: 3
- Responses per measurement: 378
18 prompts × 7 engines × 3 runs = 378 responses.
Context
The clinic had relevant information about treatments, practitioners and locations. The engines, however, did not always recognise every unit as part of the same entity.
In certain questions each location was treated in isolation. In others, competitors with more consistent local signals appeared more often.
Baseline measurement, T0
- Date
- 15 January 2026
- Citation rate
- 4.8%
- Mention rate
- 18.5%
- Recommendation rate
- 8.2%
- Share of voice
- 3.9%
- Engines recommending
- 1 of 7
- Average position
- 6.1
Reading the baseline
The brand had low presence in questions combining treatment and location.
The signals for the primary entity, the locations and the practitioners were not consistent enough between the site and external sources.
Work carried out
Structure
January 2026
- Reviewed the structure of the location pages.
- Clarified the relationship between brand, location, treatment and practitioner.
- Corrected structured data.
- Standardised name, address and contact.
- Reduced location pages with duplicated content.
Information
January and February 2026
- Rewrote treatment pages.
- Created content specific to each location.
- Added frequently asked questions.
- Explained the consultation and follow-up process.
- Improved decision pages.
Notability
February 2026
- Reviewed the main local profiles.
- Corrected inconsistencies in directories.
- Strengthened coherence between the site and external sources.
- Organised the references associated with each location.
Authority
February 2026
- Improved team pages.
- Identified the practitioners responsible.
- Reviewed sources for clinical claims.
- Clarified content authorship.
Links
January to March 2026
- Linked treatments to locations.
- Linked practitioners to clinical areas.
- Linked informational content to appointment booking.
Final measurement, T1
- Date
- 15 April 2026
- Citation rate
- 25.4%
- Mention rate
- 43.7%
- Recommendation rate
- 29.1%
- Share of voice
- 13.7%
- Engines recommending
- 5 of 7
- Average position
- 3.0
Observed change
- Date
- 15 January 2026
- Citation rate
- 4.8%
- Mention rate
- 18.5%
- Recommendation rate
- 8.2%
- Share of voice
- 3.9%
- Engines recommending
- 1 of 7
- Average position
- 6.1
- Date
- 15 April 2026
- Citation rate
- 25.4%
- Mention rate
- 43.7%
- Recommendation rate
- 29.1%
- Share of voice
- 13.7%
- Engines recommending
- 5 of 7
- Average position
- 3.0
| Metric | T0 | T1 | Change |
|---|---|---|---|
| Citation rate | 4.8% | 25.4% | 20.6 pp |
| Share of voice | 3.9% | 13.7% | 9.8 pp |
| Recommendation rate | 8.2% | 29.1% | 20.9 pp |
| Engines recommending | 1 of 7 | 5 of 7 | up 4 |
| Average position | 6.1 | 3.0 | closer to the start |
Result
Over the period measured:
- Citation rate went from 4.8% to 25.4%.
- The change was 20.6 percentage points.
- Share of voice went from 3.9% to 13.7%.
- The brand went from recommended by 1 engines to 5.
- Average position went from 6.1 to 3.0, meaning the brand began to appear closer to the start of the answers.
Conclusion
The case shows how much a coherent local entity matters.
AI engines retrieve information from different sources. When the site, the local profiles, the treatments and the practitioners use inconsistent structures, recommendation becomes unstable.
Methodology and definitions
T0 and T1 use exactly the same prompts, engines, mode, web-search setting and number of runs, and the same definitions and formulas. Without that correspondence the change would not be computed.
- Unit of observation. Case × measurement × prompt × engine × run.
- Citation rate. Responses with a qualifying citation ÷ total valid responses × 100.
- Mention rate. Responses with a mention ÷ total valid responses × 100.
- Recommendation rate. Responses with a recommendation ÷ total valid responses × 100.
- Share of voice. Recommendations of the brand ÷ total recommendations of every brand observed × 100.
- Engine recommending. An engine with a recommendation rate of 25% or higher across the prompts in the case.
- Average position. Simple mean of the brand's ordinal position, counted only in responses where it appears. A lower figure is better.
Limitations
- Results apply only to the prompts, engines, mode and period measured. They do not extrapolate to other questions or dates.
- Results in AI engines vary by version, mode, location and moment of collection.
- A citation is not an endorsement, and a recommendation does not measure service quality.
- No effect on sales, revenue, enquiries or pipeline is claimed, because none was measured.
Nature of the evidence
Baseline and final exports from the Visibility Tracker, together with dated captures of the profiles and pages analysed.
Back to all cases · Cross-engine consistency of AI visibility for Portuguese consumer-service brands, 2026