Case study · Agenzy on Agenzy

Case study · Agenzy on Agenzy

Agenzy: 80,000+ EUR of inbound pipeline from AI, and first of 33 agencies in AI answers

We ran our own GEO playbook on our own brand. Ask ChatGPT which GEO agency to hire and Agenzy is the answer it gives most often, and those answers now send us leads every week.

300,000+ AI chats analysed · 7,500+ prompts tracked · 120+ businesses audited

Vertical Agenzy case-study hero visual, first of 33 agencies in AI answers about GEO

We sell AI search visibility, so in April 2026 we put Agenzy into the same daily tracking we run for clients. The instrument is Peec AI. It runs 88 Lithuanian buyer prompts every day on ChatGPT, Gemini and Google AI Overview, with 33 agencies tracked. None of those prompts names Agenzy, and 28,000+ AI answers have been analysed. Across the full month of August 2026 Agenzy was named in 22.8% of those answers, first of 33, against 16.6% for the next agency. On ChatGPT alone that figure is 41.8%, which is 3.6 times more often than any other agency. Share of voice grew from 8.6% in April, the first month we measured, to 23.0% in August. The pipeline those answers created has passed 80,000 EUR.

80,000+ EUR

80,000+ EUR

80,000+ EUR

inbound pipeline from ChatGPT and other LLMs, arriving every week

#1 of 33

#1 of 33

#1 of 33

agencies named in AI answers about GEO, full month of August 2026

41.8%

41.8%

41.8%

of ChatGPT answers name Agenzy, 3.6 times more often than any other agency

8.6% to 23.0%

8.6% to 23.0%

8.6% to 23.0%

share of voice across ChatGPT, Gemini and Google AI Overview, April to August 2026

What the engines say across thousands of answers

Across 8,547 AI answers in August 2026, Agenzy was the agency named most often, first of the 33 tracked. On ChatGPT the gap is widest: Agenzy is named in 41.8% of answers, 3.6 times more often than any other agency, which is what the leaderboard below shows.

ChatGPT names Agenzy in 41.8% of GEO agency answers, 3.6 times more than the next agency

Share of ChatGPT answers naming each agency, August 2026. Source: Peec AI, 88 non-branded prompts.

Open ChatGPT and ask it, in Lithuanian, which GEO agency in Lithuania it would recommend. You can check this without us.

Where we started

April 2026 is the first month we measured, and Agenzy sat at 11.0% visibility, third of 33 agencies, with a share of voice of 8.6%. Not one of the 88 tracked prompts names Agenzy, so nothing in that number comes from asking the engine about us. When an engine did name us, it named us third on average.

What we changed

We ran three layers from April 2026 onward. They are the same three layers we build for a client.

The technical layer

The technical layer

We added schema that tells an engine what Agenzy is and what it sells. We opened the site to every AI engine and gave them a plain map of what is on it. We rewrote our own questions and answers in the words buyers actually type.

The authority layer

The authority layer

We published our own research on our own domain: the Lithuanian AI visibility study 2026 and the AI share-of-voice study. We wrote posts for the buyer questions that returned no brand at all, and we earned mentions on sites we do not own.

The data layer

The data layer

We track 88 buyer questions every day across three engines, and none of them names Agenzy. We read the answers themselves and keep the competitor list current, which is how the tracked set grew to 33 agencies. We test a change against a control, so it has to prove itself before it stays.

What happened, month by month

Agenzy went from 11.0% visibility in April 2026 to 22.8% in August, and from third of 33 agencies to first. The rank flipped in July and held through August. Average position, meaning where Agenzy sits in an answer that names it, improved from 3.1 to 1.9. At 1.9, Agenzy is the first or second brand in the answers that name it.

Agenzy AI visibility by month, 11.0% in April 2026 to 22.8% in August, rank third to first

Visibility on 88 non-branded Lithuanian buyer prompts. Source: Peec AI, April to August 2026. April is the first month measured.

Month

Visibility

Rank of 33

Share of voice

Average position

April 2026 (first month measured, work already under way)

11.0%

#3

8.6%

3.1

May 2026

15.0%

#2

11.8%

2.4

June 2026

15.8%

#2

14.4%

2.1

July 2026

21.0%

#1

19.6%

1.8

August 2026 (full month)

22.8%

#1

23.0%

1.9

August in full, across 8,547 AI answers

Agenzy was named in 22.8% of answers, Insite in 16.6%, APG Media in 16.2%, Digital Star in 16.2% and AISEO.lt in 10.2%. Share of voice, which is the share of all agency mentions that belong to one agency, was 23.0% for Agenzy.

ChatGPT, where the gap is widest

On ChatGPT in August 2026, Agenzy was named in 41.8% of answers, AISEO.lt in 11.6%, Digital Star in 11.0%, Insite in 9.9% and APG Media in 9.6%. Two of every five ChatGPT answers to a Lithuanian GEO buyer question name Agenzy, usually first, and 3.6 times more often than any other agency.

Agenzy visibility by engine in August 2026 across ChatGPT, Google AI Overview and Gemini

Where Agenzy stands per engine, August 2026. Source: Peec AI, 88 non-branded prompts.

What this is worth in money

80,000+ EUR

inbound pipeline from ChatGPT and other LLMs, arriving every week

4

inbound GEO deals opened in a single week in August 2026

Leads arrive from ChatGPT and the other LLMs every week. They range from small online shops to enterprises, and the pipeline they created has passed 80,000 EUR. Four inbound GEO deals opened in a single week in August 2026.

One of them was a Baltic software company. They told us on the first call how they found us. They asked Claude for AI visibility agencies in the Baltics and Agenzy was the answer. That is the whole business case in one sentence.

Across our client base, AI search has generated 500,000+ EUR for our clients in the past 3 months. The instrument there is the clients’ own attribution, at checkout and on their forms.

How we measured this, and how you check it

Instrument.

Peec AI, project “Agenzy”. 88 recommendation-framed Lithuanian prompts across the awareness, consideration, purchase and vertical stages, run daily on ChatGPT, Gemini and Google AI Overview, with 33 agencies tracked. 28,660 AI answers have been analysed in the project to date. Visibility is the share of answers naming a brand. Share of voice is a brand’s share of all brand mentions. Rankings are against the 33 tracked agencies.

What is excluded.

No tracked prompt names Agenzy or a competitor. Twelve prompts that do name agencies were added on 23 August 2026 for a separate diagnostic, and every figure here excludes them. A prompt that names the brand answers itself.

Where we are not first.

On Google AI Overview we are fourth and on Gemini we are sixth, on the same 88 prompts. We say so rather than charting only ChatGPT.

The stress test.

On 5 August 2026 we ran a separate 75-prompt experiment, built to be unwinnable by construction. It used zero branded prompts and buyer language only, across three engines including Perplexity, for 2,080 answers in one collection cycle. Agenzy came out at 35.2% visibility, #1 of 15 tracked brands, average position 1.7, and 63.9% on ChatGPT alone. It was a single-day snapshot across three engines. The ranking is against the 15 brands tracked in that experiment.

The measurement setup behind every figure on this page

88

Lithuanian buyer prompts, none of which names Agenzy

3

AI engines: ChatGPT, Gemini and Google AI Overview

Daily

every prompt runs on every engine, every day

33

agencies tracked, and 28,660 answers analysed to date

Peec AI project record, pulled 1 September 2026.

Verify it yourself.

Ask ChatGPT today, in Lithuanian: “kokią GEO agentūrą Lietuvoje rekomenduotumėte?”

Who we are.

Agenzy is the trading name of Genz way, MB, company code 307589052, registered in Vilnius, Lithuania, at agenzy.lt.

Our tooling.

Agenzy is an official Peec AI partner (Berlin, 2026). The subscription and the tooling are ours.

Questions buyers ask us about this

How long until we see the first signal, and what does three months look like?

Movement on tracked prompts usually starts in the second or third month of GEO work. Agenzy's own numbers moved every month: 11.0% in April, 15.0% in May, 15.8% in June, then a jump to 21.0% in July. Technical changes are picked up by AI crawlers within weeks. Content and outside mentions take longer, because an engine has to read them first.

How long until we see the first signal, and what does three months look like?

Movement on tracked prompts usually starts in the second or third month of GEO work. Agenzy's own numbers moved every month: 11.0% in April, 15.0% in May, 15.8% in June, then a jump to 21.0% in July. Technical changes are picked up by AI crawlers within weeks. Content and outside mentions take longer, because an engine has to read them first.

How do you know a lead came from AI?

Agenzy measures AI leads in three layers. Layer one is analytics, where chatgpt.com, perplexity.ai and the other assistants show up as traffic sources in their own right. Layer two is the form, where we ask every lead where they heard of us. Layer three is the sales call, where a buyer often names the exact question they typed into the assistant.

How do you know a lead came from AI?

Agenzy measures AI leads in three layers. Layer one is analytics, where chatgpt.com, perplexity.ai and the other assistants show up as traffic sources in their own right. Layer two is the form, where we ask every lead where they heard of us. Layer three is the sales call, where a buyer often names the exact question they typed into the assistant.

What exactly is the KPI?

The KPI is visibility on a frozen set of buyer prompts: the share of AI answers to those prompts that name your brand. Agenzy's own reading is 22.8% across 88 prompts in August 2026. The set is agreed before the work starts and never changed afterwards, because changing the prompts moves the number without changing what the engines say.

What exactly is the KPI?

The KPI is visibility on a frozen set of buyer prompts: the share of AI answers to those prompts that name your brand. Agenzy's own reading is 22.8% across 88 prompts in August 2026. The set is agreed before the work starts and never changed afterwards, because changing the prompts moves the number without changing what the engines say.

Do you guarantee results?

Agenzy does not guarantee a result, and any agency that guarantees one in AI search is guessing. What we do instead is measurable. We show your baseline before you sign anything, we freeze the prompt set so the number cannot be gamed, we run treated prompts against a holdout so a change has to prove itself, and we read the numbers with you every month.

Do you guarantee results?

Agenzy does not guarantee a result, and any agency that guarantees one in AI search is guessing. What we do instead is measurable. We show your baseline before you sign anything, we freeze the prompt set so the number cannot be gamed, we run treated prompts against a holdout so a change has to prove itself, and we read the numbers with you every month.

Can I see this for my own brand before signing anything?

Yes, and it costs nothing. Agenzy runs a free AI visibility audit: 30 buyer prompts across three engines, delivered in 48 hours, showing where your brand appears today, which competitors the engines name instead and which pages they quote. It is the same instrument the numbers on this page come from. You get the baseline whether or not you work with us.

Can I see this for my own brand before signing anything?

Yes, and it costs nothing. Agenzy runs a free AI visibility audit: 30 buyer prompts across three engines, delivered in 48 hours, showing where your brand appears today, which competitors the engines name instead and which pages they quote. It is the same instrument the numbers on this page come from. You get the baseline whether or not you work with us.

Why does an agency's own result matter to me?

Because it is the one case an agency cannot outsource or select. Agenzy is measured on the same 88 non-branded prompts, on the same three engines, on the same daily schedule as a client. If our own GEO work did not move our own numbers, no client engagement would be worth buying. Ask any GEO agency for its own reading and its own prompt set.

Why does an agency's own result matter to me?

Because it is the one case an agency cannot outsource or select. Agenzy is measured on the same 88 non-branded prompts, on the same three engines, on the same daily schedule as a client. If our own GEO work did not move our own numbers, no client engagement would be worth buying. Ask any GEO agency for its own reading and its own prompt set.

Free AI visibility audit

Find out what AI says about you before your buyers do

Send us your website and we send back where your brand appears today. Thirty buyer prompts, three engines, in 48 hours.

300,000+ AI chats analysed · 7,500+ prompts tracked · 120+ businesses audited

No time?

Call us directly - +37060420012

Book a time directly

Pick a slot in our calendar and we talk for 20 minutes about what your own numbers look like.

Pick a time on our calendar →

120+ businesses audited

We use your details only to answer this one question. Read how we handle them in our privacy policy.