GolOps
Client cases · GolOps→

Results that
speak for themselves

Real data. Real clients. How brands across industries took control of AI-driven choice in 90 days.

Request diagnostics →
Discuss project

+340%

Average growth in AI mentions

47

Brands audited

90

Days to first results

4.9

Average client rating

Client cases
Click "Details" for full breakdown
Case · 01Bank — AI recommendations
Finance · B2B · 67 days

Top-20 bank lost ground in AI recommendations — and reclaimed it in 67 days

A niche Russian SMB bank wasn't appearing in answers from ChatGPT, DeepSeek and Perplexity. A competitor held first place in 3 of 4 AI systems. In 67 days, the bank pushed it out across key scenarios.

54%
↑ from 9% · share of voice
+34%
inbound · digital
I want this →
Key results
54%
↑ from 9%
Mentions on target queries
#1
in 2 of 4 systems
First position on target queries
+34%
digital channels
Growth in inbound requests
3/5
scenarios · competitor dropped
Lost leadership in key scenarios
Starting point

The bank's marketing team found the problem by simply asking ChatGPT, DeepSeek and Perplexity "which bank is best for opening an LLC current account" — and not finding their bank in the answer. A competitor stood first in three of four tested AI chats.

The bank was spending tens of millions on SEO, paid search and PR. But none of these tools influenced what ChatGPT or Yandex AI says.

Approach
Diagnostics
Ran a full AI Audit: tested 80+ target queries across ChatGPT, Perplexity, Google AI, DeepSeek and Yandex AI. The bank was mentioned in 9% of cases vs 61% for the key competitor. Identified entry points where the competitor gained advantage and why models trust it.
Engineering
Mapped target topics, restructured the bank's public content for LLM logic, reworked product pages, expert articles and FAQs so models would treat the bank as an authoritative source. Built a narrative tied to specific decision scenarios.
Launch and monitoring
Set up weekly position monitoring across AI systems. Began displacing the competitor on priority queries in ChatGPT, DeepSeek and Alisa GPT.
Result in67 days
→Mentions grew from 9% to 54% on target queries
→Reached first position in two of four AI systems
→+34% inbound requests via digital channels
→Competitor lost leadership in three of five key scenarios

«We were spending budget on what buyers no longer use as their only search channel. GolOps showed us where decisions are actually being made today.»

Marketing Director · Name withheld under NDA
Case · 02Medicine — AI answers
Medtech · D2C · 60 days

Clinic entered Gemini's recommendations as "the best choice"

A medical clinic wanted patients to find them through AI assistants. Before working with us, models only mentioned large chains. Now the clinic ranks top-3 across 20+ medical queries.

68
↑ from 14 · GolOps Score
+190%
queries · AI channels
I want this →
Key results
68
↑ from 14 · GolOps Score
AI presence index
Top-3
across 20+ queries
Position in medical answers
+190%
AI channels
Growth in queries via AI assistants
60
days
To measurable results
Starting point

AI systems only recommended large federal medical networks for any clinic-choice query. A private clinic with strong reputation remained completely invisible — despite high ratings on review platforms.

Result in60 days
→GolOps Score grew from 14 to 68
→Entered top-3 in Gemini and Perplexity on 20+ queries
→+190% queries via AI channels
→Gemini began describing the clinic as "the best choice" in the category

«Patients started directly mentioning that they found us through ChatGPT. That had never happened before.»

Chief Physician · Name withheld under NDA
Case · 03Government agency — exports and AI
Government · B2G · 7 months

Federal agency reclaimed control of its image in global AI systems

An export support agency discovered: foreign partners asking ChatGPT about working conditions with Russian suppliers were getting outdated data. The task wasn't to hire a contractor, but to build expertise inside the team.

12×
faster · AI Audit
95%
team autonomy
I want this →
Key results
12×
AI Audit speed
From several weeks to several hours
95%
operational autonomy
A team without technical background runs independently
6+
↑ from 1 pilot · 7 mo
Domains covered after scaling
0
IT dependency
Communications staff manage AI choice themselves
Starting point

Agency analysts found that when foreign partners and prospective importers asked ChatGPT or Perplexity about conditions for working with Russian suppliers, government support programs or sector leaders — answers were forming spontaneously. AI systems reproduced outdated data, ignored current programs and in some cases mentioned competing jurisdictions as more attractive.

The problem wasn't a lack of information — the agency produced it in volume. The problem was that AI models didn't see it or didn't account for it. Hiring an external contractor for every new query is expensive, slow, and fundamentally low-quality: deep immersion in government program subject matter cannot be ensured from outside.

Approach
The solution — transfer the competence inside the agency. GolOps designed a program tailored to the team's real tasks and infrastructure.
Strategic sessions
Conveying systemic understanding of LLM logic: how models decide what to recommend, what content they consider authoritative, and how that's measured. No technical jargon — in the language of communications and analysts.
Work on real infrastructure
The team built first workflows not on training examples but on live programs: export subsidies, sector registries, SMB participation conditions. The first measurable shifts came during the same work cycle.
Proprietary framework for government specifics
Architecture adapted to regulatory constraints and public-content formats of government bodies. Ready to scale to new domains without external contractors.
Expert support
The GolOps team worked with specific wording, narratives and document structure — until measurable results in target AI systems were achieved.
Result in7 months
→AI Audit for a new domain — from several weeks to several hours (12×)
→95% operational independence: specialists build, test and launch AI workflows without IT
→Scaled from 1 pilot program to 6+ domains in 7 months
→Communications staff and analysts manage AI choice in ChatGPT, Perplexity, Google AI and Yandex AI on their own

Participants particularly noted: a clear explanation of why models prefer some sources over others; the ability to verify changes against live ChatGPT and Perplexity queries during sessions; and adaptation of the methodology to government content specifics and international audiences.

«We didn't expect that managing our image in AI systems would be this operationally achievable. After the program the team works on its own and we see results on the specific queries that matter to us.»

Director of Digital Communications · Name withheld under agreement
Case · 04Premium medical network — AI recommendations
Medicine · Premium · B2C · 5 months

Federal clinic network reached #1 in AI recommendations across 40+ medical scenarios

A highly competitive market: 4–5 players with comparable budgets and physician quality. Team — 2 people in digital. A competitor held first place in ChatGPT for 60% of scenarios. Not because they were better. Because they started earlier.

×6
growth in citations · LLM
#1
ChatGPT + Yandex AI · 40+
I want this →
Key results
×6
growth in citations
Overtook 2 of 4 competitors in AI visibility
#1
40+ scenarios
First position in ChatGPT and Yandex AI
61%
↑ from 22% · share of choice
On target AI queries
180
↑ from 35 · pieces/qtr
Same 2-person team
+38%
inbound · digital
With stable average ticket
Starting point

A patient types into ChatGPT "Which private clinic is best for oncology diagnostics in Moscow?" or "Where can I get a quality paid MRI?" — that's how health decisions are now made. Before calling reception, before reading reviews, before ads, before visiting the site.

The clinic was spending hundreds of millions of rubles a year on paid search, SEO and PR. But none of these tools influenced what AI says. A competitor stood first in ChatGPT and Yandex AI answers for 60% of tested scenarios — not because service quality was higher. Because they started working with AI earlier.

What diagnostics revealed

We ran an AI Audit across 140 choice scenarios — from "best clinic for a checkup" to "where to do IVF in Moscow" and "private clinic for a child with a rare diagnosis". The clinic appeared in 22% of cases. The key competitor — in 58%.

The cause — content architecture. Models saw the competitor as an authoritative source on medical topics. The network's content was written for traditional search engines — for Google, not for LLM logic.

Approach
Rebuilt content architecture
Reworked 80+ pages for the logic that language models trust: expertise structure, answers to scenario queries, narratives for specific conditions and procedures. Not SEO copy — medical authority in a format LLMs read.
Scaled production
Time to create one AI-optimized piece: from 2.5 hours to 12 minutes. In 4 months a 2-person team produced 180 pieces vs 35 a quarter previously. No hires, no outsourcing — just process redesign.
Built analytics
Weekly position monitoring across ChatGPT, Perplexity, Google AI and Yandex AI for all target scenarios. For the first time, marketing could see not "reach" and "impressions" — but the specific recommendations real patients receive right now.
Before
2.5 hr
per piece
After
12 min
per piece
Result
×12.5
productivity
Result in5 months
→×6 growth in LLM citations — overtook two of four key competitors
→#1 position in ChatGPT and Yandex AI across 40+ medical scenarios
→AI choice share grew from 22% to 61% on target queries
→180 pieces per quarter vs 35 — same 2-person team
→+38% inbound requests via digital channels with stable average ticket

«We compete with clinics that have the same physician quality and equipment. You can only win where they're not yet looking. AI recommendations turned out to be exactly that place.»

CMO of a federal medical network · Name withheld under NDA
Case · 05Energy holding — international AI choice
Oil & Gas · International business · 90 days

A Top-5 Russian energy holding — invisible in global AI systems. Changed in 90 days

Projects on three continents. Negotiations in the UAE, Turkey, India. Annual deal volume — billions. But when a partner in Dubai typed "Which companies supply LNG to Asia?" into ChatGPT — the company wasn't in the answer. Competitors from the US, Qatar and Norway were.

41%
↑ from 3% · AI choice share
19 days
to first citation
I want this →
Key results
41%
↑ from 3% · share of choice
On target international queries
+340%
mentions
In ChatGPT and Perplexity
+18
queries / mo · was 0
Inbound through AI channels
6
markets covered
UAE, Turkey, India, China, Germany, France
19
days · 1st citation
To first appearance in ChatGPT answers
Starting point

The company is in the top-5 of its industry in Russia. Projects across three continents. Negotiations with partners in the UAE, Turkey and India. Annual international deal volume — billions. But when a prospective partner in Dubai opened ChatGPT and typed "Which companies supply LNG to Asia?" — the company wasn't in the answer. Not in ChatGPT. Not in Perplexity. Not in Google AI Overview. Competitors from the US, Qatar and Norway were.

This isn't a reputation issue. The company invested in PR, international conferences, an English-language site. Classic SEO worked. But LLM algorithms are trained on different sources — authoritative international media, sector databases, analyst houses. The company was almost absent there.

Markets →
🇦🇪 UAE🇹🇷 Turkey🇮🇳 India🇨🇳 China🇩🇪 Germany🇫🇷 France
Approach — three directions in parallel
AI Audit across 7 systems: ChatGPT, Perplexity, Google AI, Microsoft Copilot, Grok, DeepSeek, Yandex AI. Competitor analysis on 40+ target queries in English, Arabic and French. Diagnostic finding: the company — 3% citations, competitors — 40–70%.
01 · Infrastructure
Citations in international sources

Publications in authoritative international outlets — sector portals, analyst houses, business media. Content was created by a bilingual team with native speakers for each market.

02 · Narrative
Restructuring for LLM logic

Existing materials were reworked: FAQ architecture, clear competence claims, comparison formats that algorithms extract as direct answers to questions.

03 · Technical
RAG optimization

Technical documents, press releases and sector data are structured for indexing in corporate AI assistants — Copilot, Notion AI, partners' due-diligence tools.

Result in90 days
→AI choice share on target queries grew from 3% to 41%
→On "LNG suppliers to Asia" queries — in the top two ChatGPT positions alongside Qatar Energy and Shell
→+18 inbound requests via AI channels per month (was 0). 6 — qualified negotiation requests
→First citation in ChatGPT — 19 days after work started
→Negotiation cycle shortened: partners arrive "warm" — they saw the AI recommendation before contact

«We spent two years and significant budgets on international PR. AI systems didn't know us. GolOps changed that in three months — methodically, measurably, without unnecessary noise. Now our partners in Dubai say they found us through ChatGPT. That had never happened before.»

International Development Director · Energy Holding · NDA
Case · 06B2B SaaS — entering MENA and SEA markets via AI
B2B SaaS · Cybersecurity · MENA + SEA · 8 months

Russian SaaS vendor entered MENA and SEA markets through AI choice — without a single external agency

The product competed with Israeli, European and American vendors in markets where no one knew the company. Out of 200+ tested ChatGPT queries — zero mentions. The task: become visible in Arabic, English and Malay without a bloated budget.

#1
17+ Gulf queries · ChatGPT
−50%
localization costs
I want this →
Key results
#1
17+ queries · Gulf
First position in ChatGPT and Perplexity on enterprise queries
+25%
AI Discovery traffic
ChatGPT, Gemini, Perplexity and DeepSeek in target markets
~85%
branded clicks
AI recommendations built awareness ahead of advertising
12.7%
leads · 5x YoY
High-quality leads from new markets via AI channels
2×
content speed
Time to launch a piece in a new market cut in half
−50%
localization costs
Full move away from external content agencies
Key discovery

~85% of clicks on branded queries in target regions came through AI recommendations — more than Google, Bing, Baidu, Brave, Yahoo, Yandex and DuckDuckGo combined. Brand awareness formed before the click.

Starting point

When a CTO from Dubai asked ChatGPT "What are the best enterprise data protection solutions for the GCC region?" — the answer contained five competitors. The Russian vendor wasn't there in any of 200 tested queries. The product was stronger on several parameters — in AI search it simply didn't exist in these markets.

Hiring local agencies in four countries — slow, unpredictable, hard to control. An AI Audit across 200+ queries in three languages confirmed the diagnosis: all content existed only in Russian and technical English. LLM systems didn't see it as relevant for MENA. Competitors had spent years publishing case studies for local scenarios — in Arabic, Malay, with references to regional regulation.

Markets
🇦🇪 UAE🇸🇦 Saudi Arabia🇮🇩 Indonesia🇲🇾 Malaysia
Languages
EN (Gulf)AR عربيMS Melayu
Approach — three directions
01 · Architecture
Multilingual AI architecture

Content structure for LLM logic in three languages: English (Gulf), Arabic, Malay. Not translation — narrative adaptation to the cultural and regulatory context of each market.

02 · Production
Scaling without agencies

A two-person team got the GolOps framework: 25 pieces a month vs 10–12, 80% shipped copy-ready for JTBD queries, production cost down −50% vs local agencies.

03 · Analytics
Monitoring by markets and languages

Weekly position monitoring per region and language. The team saw, for the first time, not "reach" — but the specific recommendations buyers in Riyadh, Dubai and Kuala Lumpur are getting right now.

Result in8 months
→#1 position in ChatGPT and Perplexity on 17+ priority enterprise queries in the Gulf region
→+25% traffic from AI Discovery (ChatGPT, Gemini, Perplexity, DeepSeek) in target markets
→12.7% high-quality leads from new markets via AI channels — 5× YoY growth
→~85% of branded clicks formed via AI — more than all traditional search engines combined
→2× speed to launch content in a new market, −50% localization cost, 0 external agencies

«We would have spent two years and millions of dollars on agencies to get what we got in eight months. The key insight: AI systems don't know you exist — until you specifically explain it to them in their language.»

VP of International Growth · Name withheld under NDA
Case · 07
GolOps initiative · 2025
Invisible Russia — narrative in global AI systems
Narrative · Country study · 7 months

Invisible Russia: how we changed the perception of Russian intellectual capital in global AI systems

"Which countries produce the best programmers?" — Russia didn't make the top-3 in any ChatGPT answer. Even though Russians have led world rankings for two decades. AI didn't know it. We decided to test — and to change it.

38%
↑ from 11% · English-language AI
#1
Perplexity · programmers
Request diagnostics →
11%
Russia was mentioned in 11% of cases
where data implied 60%+

In English-language AI systems. On topics where Russian achievements objectively rank in the global top-3. In 67% of cases when Russia did appear — the wording was neutral-to-negative.

Research results
+340%
top-3 mentions
Russia in answers about "world's best programmers" in ChatGPT and Perplexity
38%
↑ from 11% · English-language AI
Share of positive-to-neutral mentions on selected topics
#1
11 weeks · Perplexity
On the query "world's best competitive programmers"
×4
science citations
Growth in mentions of Russian achievements in the "global innovation leaders" context
Starting point

We asked ChatGPT, Perplexity and Google AI one question: "Which countries produce the world's best programmers?" Russia didn't make the top-3 in any answer. The US, India, China — over and over.

Meanwhile, according to Codeforces, ICPC and HackerRank data, Russian developers have led world rankings for two decades. Of the 64 ICPC World Finals winners over the last 20 years — 18 are from Russia. AI didn't know that. Or knew it but didn't say it.

Diagnostics · 300+ queries · 3 languages (RU / EN / AR)
Audit topics
· Russian science and engineering
· Programmers and mathematicians
· Technological achievements
· Cultural contribution to global heritage
· Investment climate and business environment
What we found
23%
Mentions across all languages (should be 60%+)
11%
In English-language systems
67%
Neutral-to-negative wording

This isn't about a conspiracy. Every producer country prioritises its own language when training a model — English-language opinion dominates ChatGPT and Perplexity, French in Mistral, Chinese context in DeepSeek. The Russian side simply didn't create content in the formats foreign LLMs treat as authoritative.

Approach — three pilot topics
01
"Russian programmers"

An English-language corpus — with Codeforces and ICPC data, rankings, specific names and achievements. In a format LLMs treat as expert source, not propaganda.

02
"Russian school of mathematics"

Fields medalists, contributions to topology, probability theory, cryptography. Content was created not as a Wikipedia article but as a trust narrative for language models.

03
"Russian technologies"

Atom, space, cybersecurity. Structured for scenario queries: "who leads in nuclear energy", "who builds reactors globally", "best countries in cybersecurity".

Result in7 months
→+340% growth in Russia's top-3 mentions on "world's best programmers" in ChatGPT and Perplexity
→From 11% to 38% — share of positive-to-neutral mentions in English-language AI systems
→#1 position in Perplexity on the query "world's best competitive programmers" — in 11 weeks
→×4 growth in citations of Russian science achievements in the "global innovation leaders" context
Why this matters for business

Every time a foreign investor, partner or buyer asks AI about Russia — they get an answer not shaped by Russia. The country didn't show up on the field, so others decided how it would be positioned.

This applies to more than the country. It applies to every Russian company that wants to work in international markets. Its context is shaped by the narrative AI absorbed about its country of origin. GolOps changes this narrative — systematically and measurably.

TL;DR
·300+ queries audited in three languages
·Russia was mentioned in 11% of cases where it should be in 60%+
·In 90 days — growth from 11% to 38% on target topics
·#1 in Perplexity on "world's best competitive programmers"
·×4 growth in citations in the global innovation context

This isn't about patriotism. It's about data working for you — not against you.

Want to know what AI says about your country, region or business?
Narrative audit — the first step toward managing your image in AI systems
Request diagnostic →
Next step

Influence the choice.

From diagnostics to dominance in AI recommendations.

Request Diagnostics →
Discuss project