TL;DR: ET-Laser raised measurable Yandex-family visibility in weeks; the case pairs quotes, competitors, and GPI-style score movement with operational next steps. .
Case study
ET-Laser: top-3 list slot +20 pts GPI™ in 8 weeks — how a tattoo removal clinic closed AI answer gaps
Client
ET-Laser is a tattoo removal clinic in Russia. Site: et-laser.ru. Owner: Viktor Galushin.
Viktor Galushin
“After the first report I saw numbers, not gut feel: GPI™ at 29 with ~26% share of voice—and citations were aggregators, not our domain. Within a couple of weeks we were already in the top three on list position when named; by week eight GPI™ was 49, SOV ~43%, and our site showed up in the citation mix. That’s a before/after you can’t argue with.”
Owner, ET-Laser
Starting point
Before the first content sprint, the baseline run looked like this:
- GPI™ 29 — critically low for a clinic that already invests in marketing; the gap to the sample leader was visible in one screen.
- Share of voice ~26% — the brand disappeared in most model paths on the pack.
- Average list position ~4+ when named — often outside the top three even though the clinic occasionally appeared in answers.
- et-laser.ru ~1.5% of citation share in parsed answers — aggregators and directories captured almost all attribution; the clinic site barely appeared.
- ~21% negative sentiment on price-style prompts — tied to wording about “hidden fees” and vague pricing pages models kept paraphrasing.
Monthly cadence
- Week 1 — baseline. Full model matrix, PDF snapshot for the founder. Logged GPI™, share of voice, models that never mention the brand.
- Week 2 — citation audit. Found: aggregators accounted for ~88% of top citation slots; et-laser.ru was effectively invisible. Action: prioritized org/medical schema, service copy, and fresh third-party signals the models could pick up. Signal: list-position average was already ~2.3 (top 3) on naming runs—GPI/SOV still had room to grow.
- Week 3 — prompt gaps. Found: three aftercare/rehabilitation prompts where competitors appeared and ET-Laser did not. Action: drafted FAQ + service block on recovery timelines; scheduled a narrow rerun on those intents.
- Week 4 — tone pass. Found: “discomfort” phrasing in answers skewed negative even when clinically normal. Action: rewrote the on-site block to separate sensation vs. risk, added plain-language reassurance.
Sample prompts we ran
Same neutral wording every run—brand name never injected into the question text. Examples from the pack:
«Where is the best place to get a tattoo removed in Moscow?»
«How painful is laser tattoo removal really?»
«Clinic vs home laser device for tattoo removal—which is safer?»
«What should I know about aftercare in the first week after laser removal?»
Insights
Each item follows what we actually did—not just what we noticed:
Found: Aggregators held ~88% of top citation slots; et-laser.ru sat near ~1.5%.
Did: Tightened medical/organization schema, refreshed service descriptions, lined up corroborating reviews and listings.
Result: By week 8, et-laser.ru citation share reached ~5–6% in the same pack—up alongside GPI™ 49 and ~43% SOV, with list position holding in the top three.
Found: Three aftercare prompts surfaced competitors while ET-Laser was absent.
Did: Shipped a dedicated rehab/FAQ block plus internal linking from procedure pages.
Result: Follow-up runs started mentioning ET-Laser on those intents instead of leaving the field empty.
Found: “Discomfort” language pushed sentiment negative even when clinically expected.
Did: Rewrote the discomfort section with clearer expectations and photos of normal healing stages.
Result: Tone chips on those prompts moved from negative-heavy to neutral/positive in subsequent checks.
Found: ~21% negative sentiment on pricing prompts—models echoed “hidden fee” chatter.
Did: Rebuilt the pricing page with an explicit “what’s included” breakdown and removed ambiguous upsell wording.
Result: Roughly three weeks later, negative share on price prompts fell to ~11% with the same pack.
Tracking
- Scheduled visibility checks with brand "ET-Laser", tattoo-removal niche, and et-laser.ru in the form—every run scores the same entity.
- Six model families per run (ChatGPT-class, Claude, Gemini, Perplexity, Grok, DeepSeek) on a fixed scenario pack—not ad-hoc chats.
- Compared share of voice and GPI™ to the competitors each model surfaced in the same batch.
- Citation hosts show whether answers lean on aggregators vs. the clinic domain.
Reports
- Live executive report — GPI™ hero, KPI cards, competitor bars, quotes, citation table.
- Per-model grid for mention rate and tone—spots a single-channel drop early.
- Filtered quotes to catch wrong prices, outdated addresses, stolen recommendation slots.
- PDF export for frozen snapshots to share with partners.
What we read
- Named at all? → share of voice + per-model grid.
- List position when named → average rank + quote review.
- Trusted sources → citation hosts; add corroboration if the brand site is missing.
- Tone risk → sentiment mix; rewrite the passages models keep mangling.
Report
Live ET-Laser demo data in the same dashboard clients use—GPI™, share of voice, competitors, citations, weekly dynamics. Prefer scrolling this to static “report quotes”; numbers and charts are verifiable.
Dynamics across runs
8 weeks · same prompt pack and models
GPI™ and share of voice by measurement date.
GPI™ gained 20 points over 8 weeks (29 → 49); share of voice rose from ~26% to 43%; average list slot stayed in the top 3, with more answers citing et-laser.ru.
GPI™
Share of voice %
Live visibility
ET-Laser
How AI models see “ET-Laser” — report: 6 models, 8 scenarios each · 5 April 2026 · niche: Удаление тату и татуажа · Москва
Leader in this sample: Лазер Мед (GPI 63) · You: 39 · Gap: 24 points.
GPI™
39/100
GPI™ scale (0–100)
0
100
You: 39
Leader (Лазер Мед): 63
AI brand coverage
43%
Models that named your brand. Competitor rows use a different metric (name share in answer text).
Avg AI position
2.3
Average slot in recommendation lists (1 ≈ first).
Website citation share
6%
Share of links to your site among sources in answers.
Sentiment (pos · neu · neg)
48% · 34% · 18%
Positive · neutral · negative on labeled answers.
StrongAt riskCritical
Key takeaways
5 of 6 models mention you — grow the rest. “Лазер Мед” leads competitor mentions in answers.
Citation share looks solid for this run.
Avg position 2.3 — mid-pack (often 2nd–3rd).
Some doubt in tone — add dates, ranges, warranties.
UGC: Map & aggregator reviews—surface indexed reviews and third-party proof so answers don’t rely only on generic claims.
Visibility by model
Share of prompts in this run where the brand was mentioned, plus tone mix per model. Based on structured fields from the check, not manual reading of every answer.
ChatGPT
50%
Brand mentioned · 4/8 prompts
pos / neu / neg
Top of list in 4/4 ranked answers
Claude
50%
Brand mentioned · 4/8 prompts
pos / neu / neg
Top of list in 3/4 ranked answers
Gemini
50%
Brand mentioned · 4/8 prompts
pos / neu / neg
Top of list in 3/4 ranked answers
Grok
50%
Brand mentioned · 4/8 prompts
pos / neu / neg
Top of list in 3/4 ranked answers
Perplexity
50%
Brand mentioned · 4/8 prompts
pos / neu / neg
Top of list in 2/4 ranked answers
DeepSeek
0%
Brand mentioned · 0/8 prompts
pos / neu / neg
GPI™ vs competitors
Лазер Мед
63
Анджелик Клиник
54
Студия на Тверской
46
ET-Laser ← you
39
Федеральная сеть «Дерма+»
39
Strong (60+)At risk (35–59)Critical (0–34)
Quotes
Real model answers — competitor names from your report are highlighted. Five rows first; open the rest with the button. Default filters: positive and neutral.
Shown 11 of 12 entries with text. Total rows in report: 48. Negative is hidden by default.
Models
Sentiment
| Model | Prompt | Sentiment | Text |
|---|---|---|---|
ChatGPT | Где безопасно удалить тату в Москве | Positive | Среди частных клиник с лазерным оборудованием нередко рекомендуют ET-Laser — отдельно отмечают консультацию перед процедурой. |
ChatGPT | Сравнение клиник лазерного удаления тату | Neutral | В одном ответе могут соседствовать ET-Laser, сетевые студии и дерматологические центры — сравнивайте протоколы и сертификаты аппаратов. |
Claude | Больно ли удаление небольшой татуировки | Neutral | Интенсивность зависит от глубины пигмента; в отзывах о ET-Laser чаще пишут о терпимом дискомфорте и поэтапном курсе. |
Gemini | Сколько сеансов нужно для чёрной тату | Positive | ET-Laser и несколько других московских клиник указывают ориентир 4–8 процедур — точнее после очного осмотра. |
Perplexity | Рейтинг клиник удаления тату цена качество | Neutral | Агрегаторы и карты часто поднимают сети с большим числом отзывов; ET-Laser встречается реже лидеров по охвату, но с устойчивыми оценками. |
Citation sources
Domains extracted from answers in this run. For the full link archive, open the classic HTML report below.
Top 14 domains by citation share in this sample
| Domain | Share · links | |
|---|---|---|
1. | prodoctorov.ru | 10.5%·20 |
2. | zoon.ru | 8.9%·17 |
3. | 2gis.ru | 8.1%·15 |
4. | yandex.ru | 7.2%·13 |
5. | docdoc.ru | 6.4%·11 |
6. | et-laser.ru Your site | 5.8%·11 |
7. | google.com | 5.2%·9 |
8. | vc.ru | 4.5%·8 |
9. | instagram.com | 4%·7 |
10. | t-j.ru | 3.6%·6 |
11. | webmd.com | 3.2%·5 |
12. | healthline.com | 2.9%·5 |
13. | reddit.com | 2.5%·4 |
14. | laser-med.example | 2.3%·13 |
Percentages are shares among extracted hosts in this report, not the entire web.
Action plan
- Coverage & competitors: 5 of 6 models mentioned ET-Laser at least once. Still silent this run: DeepSeek. Among answers that name competitors, Лазер Мед shows up about once every 3—that name is easy for models to recall. Create a “ET-Laser vs Лазер Мед” page with an honest side-by-side. In a niche like “Удаление тату и татуажа · Москва”, spell out price, timelines, materials, or whatever buyers compare head-to-head. Models often cite pages like this for “what should I pick” questions. Prioritize intents where visibility lags: «Удаление тату перед армием» — add or refresh landing sections that answer those prompts directly. Weakest model rows: DeepSeek 0/8 prompts with a mention. Task: rerun after you add niche facts, lists, and citations those models tend to use.
- Website citations: Your site is about 5.8% of extracted citation hosts in this run. To keep that share from slipping, add fresh earned mentions on sites assistants already treat as sources. Pitch mentions on domains models already cite in this run: prodoctorov.ru, zoon.ru, 2gis.ru, yandex.ru.
- List position: Average list position 2.3 (among answers where the model assigned a list slot) — near the top when models rank options. Keep a flagship page titled “Удаление тату и татуажа · Москва — ET-Laser” so assistants can quote one clear story. Add 5–7 checkable facts there: price ranges, timelines, regions, warranties, and example projects—models reuse concrete numbers more than slogans.
- Sentiment: Labeled tone: ~48% pos / ~34% neu / ~18% neg. Task: add checkable claims (dates, ranges, warranties) so neutral answers don’t drift into doubt.
- Site & next steps: UGC: Map & aggregator reviews—surface indexed reviews and third-party proof so answers don’t rely only on generic claims.
GPI™ = coverage 35% · citation 30% · position 20% · sentiment 15% (renormalized if a field is missing).
GPI™ is a weighted index (renormalized to available fields): coverage 35% · citation 30% · position 20% · sentiment 15%.
Scores reflect this check run; competitor rows use visibility or mention share where GPI is not stored per brand.
Try it on your niche
Working in medicine, cosmetology, or another trust-heavy service category? Run the same check—your first snapshot is free—and you’ll see GPI™, share of voice, and citations before you rewrite a single landing page.
Start a free check