LoudFace
Marketing

Does Naming ChatGPT in Your Question Change Which Agencies AI Recommends? 1,788 Answers From ChatGPT, Perplexity and Google AI Overviews (2026)

We read 1,788 answers from ChatGPT, Perplexity and Google AI Overviews to 20 agency questions. When the question named an engine, every engine named fewer tracked agencies. On ChatGPT, Omniscient appeared in 65.0% of those answers against 96.4% of the others, and Perplexity named Derivatex in 25.0% against 8.3%.

The short answerYes. Across 1,788 answers from ChatGPT, Perplexity and Google AI Overviews (23 October to 21 November 2026), questions that named an AI engine got fewer of the 53 tracked agencies on all three engines. ChatGPT named 4.33 tracked agencies per answer against 6.52, Perplexity 1.73 against 3.45, and Google AI Overviews 3.18 against 4.12.

EngineTracked agencies per answer: engine namedNo engine namedChangeAnswers naming no tracked agency (named vs not)
ChatGPT4.336.5234% fewer0.4% vs 0%
Perplexity1.733.4550% fewer30.0% vs 10.3%
Google AI Overviews3.184.1223% fewer4.2% vs 1.1%
All three3.084.7011.5% vs 3.8%

The agencies that lose the most are the ones that lead the plain answers. iPullRank, Siege Media, First Page Sage and Omniscient each fall on every engine separately. On ChatGPT, iPullRank went from 74.4% of answers to 42.1% and Omniscient from 96.4% to 65.0% when the question named an engine. A few smaller specialists move the other way: Perplexity named Derivatex in 25.0% of engine-named answers against 8.3%, ChatGPT in 12.9% against 1.7%, and Google AI Overviews in 15.9% against 7.4%.

The engines' own searches follow the question. When the question named an engine, ChatGPT put an engine name into its own web searches in every answer where its searches were recorded. Engine-named answers gave a smaller share of their citations to listicles, on all three engines.

ChatGPT named 4.33 tracked agencies per answer when the question named an AI engine, against 6.52 when it did not

0246ChatGPT, engine named4.33ChatGPT, no engine6.52Perplexity, engine named1.73Perplexity, no engine3.45AI Overviews, engine named3.18AI Overviews, no engine4.12
ChatGPT, engine named4.33
ChatGPT, no engine6.52
Perplexity, engine named1.73
Perplexity, no engine3.45
AI Overviews, engine named3.18
AI Overviews, no engine4.12
LoudFace
Source: LoudFace study of 1,788 Peec AI answers, 23 Oct to 21 Nov 2026; 53 tracked agencies

Tracked agencies per answer, engine named vs not. Source: LoudFace study of 1,788 Peec AI answers, 23 October to 21 November 2026.

How we ran the study

We track buyer questions for LoudFace in Peec AI, which runs each question once a day on ChatGPT, Perplexity and Google AI Overviews and records the answer, the sources it cites, the searches the engine ran and the brands it names.

For this study we took 20 agency questions we track and split them by one thing: whether the question names an AI engine. We fixed the two groups before reading any data.

8 questions name an engine. Six name ChatGPT: "Best agency for getting cited in ChatGPT", "Which agencies will help a B2B SaaS show up in ChatGPT answers?", "Can you recommend an agency that helps B2B SaaS companies get mentioned in ChatGPT answers?", "Best agency for ChatGPT and Perplexity citations", "Best agency for getting mentioned in Perplexity and ChatGPT" and "Best agency for getting B2B SaaS cited in ChatGPT and Claude". Two name Google's AI: "Best agency to get a SaaS product cited in Google AI Overviews" and "Agency that optimizes SaaS sites for Gemini and AI Overviews".

12 questions ask for the same kind of agency without naming one. "Best AI search agencies", "Best GEO agencies", "Top GEO agencies 2026", "Best AEO agencies 2026", "Best AI visibility agencies", "Best LLM SEO agencies", "Best agency for AI search optimization for SaaS", "Agency that helps SaaS companies rank in AI search", "AI search marketing agency for SaaS companies", "Who is the best AI search agency for B2B SaaS companies in 2026?", "Generative engine optimization agency for B2B SaaS" and "What agencies specialize in generative engine optimization for SaaS? Give me a shortlist I can actually contact."

The window is 30 days, 23 October to 21 November 2026, one answer per question, engine and day. Peec returned 1,800 runs; 1,788 carry an answer (Google AI Overviews showed no answer on 12 runs). That is 719 answers to engine-named questions (240 ChatGPT, 240 Perplexity, 239 Google AI Overviews) and 1,069 to the others (360, 360 and 349).

What counts as named. We use the unit of Peec's visibility metric, which "measures the percentage of AI responses that mention your brand": an agency is named in an answer when Peec's brand detection lists it. Peec tracks 55 brands in our project. We left out two, as in our earlier studies in this series: LoudFace, which publishes this study, and GEO Agency, whose name is a generic phrase. That leaves 53 tracked agencies. "Tracked agencies per answer" is the number of those agencies an answer names, averaged over answers.

A check on the detector. We searched the text of every answer, links removed, for 13 agencies this article relies on, matching each name as a whole word. For all 13 the text search and Peec's detection agree within 0.3 points in both groups.

All three engines name fewer tracked agencies when you name an engine

Per engine, answers to engine-named questions carried 23% to 50% fewer tracked agencies on average than answers to the others. The medians move the same way: ChatGPT's typical answer named 4 tracked agencies instead of 7, Perplexity's 1 instead of 3.5, and Google AI Overviews' 3 instead of 4.

Perplexity shows the biggest change. In 30.0% of its answers to engine-named questions it named none of the 53 tracked agencies, against 10.3% when no engine was named. Its answers stay short either way (a median of 48 words against 49), so the drop in tracked agencies comes with no drop in answer length.

Here is one Perplexity answer to "Best agency for getting cited in ChatGPT", read on 20 November. In 48 words it said no agency is independently verified as best and named two firms, Red-engage and Catalyst. Peec tracks neither. On the same day ChatGPT answered "Best AI search agencies" with a ten-row table: iPullRank, NoGood, First Page Sage, Siege Media, Omniscient Digital, Directive, Avenue Z, Single Grain, WebFX and Skale.

In the answers we read, part of the drop comes from engines naming firms outside the tracked list, on top of naming fewer firms. Kalicube, which Peec does not track either, sat in ChatGPT's top five for "Best agency for getting cited in ChatGPT" the same day. We count the 53 tracked agencies throughout, and they include the agencies that lead the plain answers.

The four category leaders lose the most

Share of answers naming each agency, all three engines:

AgencyEngine named (719 answers)No engine named (1,069)Change (points)
iPullRank27.8%51.8%-24.0
Siege Media24.9%48.7%-23.8
First Page Sage25.6%47.0%-21.4
Omniscient48.7%69.6%-20.9
NoGood9.9%29.4%-19.5
Directive Consulting12.1%29.3%-17.2
WebFX1.5%14.1%-12.6
Skale10.8%22.7%-11.9
Animalz11.3%22.3%-11.0

Each of the nine falls on every engine separately too:

AgencyChatGPT (named vs not)PerplexityGoogle AI Overviews
iPullRank42.1% vs 74.4%26.3% vs 45.8%15.1% vs 34.7%
Siege Media42.5% vs 75.0%17.9% vs 41.1%14.2% vs 29.5%
First Page Sage45.8% vs 69.4%16.7% vs 32.8%14.2% vs 38.4%
Omniscient65.0% vs 96.4%42.5% vs 57.2%38.5% vs 54.7%
NoGood7.9% vs 36.9%4.2% vs 19.7%17.6% vs 31.5%
Directive Consulting27.5% vs 49.4%0.8% vs 16.7%7.9% vs 21.5%
WebFX3.8% vs 26.4%0% vs 10.6%0.8% vs 5.2%
Skale16.7% vs 39.4%2.9% vs 12.8%13.0% vs 15.8%
Animalz23.8% vs 42.8%0% vs 11.7%10.0% vs 12.0%

Omniscient still has the highest share of the four in both groups. But ChatGPT named it in 96.4% of answers to plain questions and in 65.0% once the question said "ChatGPT" or another engine. A buyer who asks ChatGPT the plain question sees it almost every time. A buyer who names an engine sees it in about two ChatGPT answers of three.

Omniscient was named in 48.7% of answers when the question named an engine, against 69.6% when it did not

0%20%40%60%Omniscient, with engine48.7%Omniscient, no engine69.6%iPullRank, with engine27.8%iPullRank, no engine51.8%Siege Media, with engine24.9%Siege Media, no engine48.7%First Page Sage, with engine25.6%First Page Sage, no engine47%
Omniscient, with engine48.7%
Omniscient, no engine69.6%
iPullRank, with engine27.8%
iPullRank, no engine51.8%
Siege Media, with engine24.9%
Siege Media, no engine48.7%
First Page Sage, with engine25.6%
First Page Sage, no engine47%
LoudFace
Source: LoudFace study of 1,788 Peec AI answers, 23 Oct to 21 Nov 2026

Share of answers naming each agency, all three engines. Source: LoudFace study of 1,788 Peec AI answers, 23 October to 21 November 2026.

A few specialists gain

AgencyEngine namedNo engine namedChatGPT (named vs not)PerplexityGoogle AI Overviews
Derivatex17.9%5.8%12.9% vs 1.7%25.0% vs 8.3%15.9% vs 7.4%
Discovered Labs10.2%6.8%16.3% vs 12.2%3.8% vs 2.5%10.5% vs 5.7%
PipeRocket Digital6.8%1.2%7.1% vs 0.6%1.3% vs 0%12.1% vs 3.2%
MADX Digital5.0%1.0%4.2% vs 0.8%4.6% vs 0%6.3% vs 2.3%

Derivatex is the clearest case. Perplexity named it in a quarter of its engine-named answers and in 8.3% of the others. Each engine's top ten changes too. Perplexity's engine-named top ten brings in Derivatex, RevenueZen, MADX Digital and Discovered Labs and drops Directive Consulting, Skale, Graphite and Animalz. ChatGPT's brings in Discovered Labs and Minuttia and drops NoGood and WebFX. Google AI Overviews' brings in Derivatex and Revv Growth and drops Directive Consulting and Onely.

Engine-named questions still draw mostly the same names: ChatGPT and Google AI Overviews keep 8 of their top 10, Perplexity 6. The change is in how many of them make each answer, and in two to four names in each top ten.

The makers say so. OpenAI's help center says "ChatGPT search typically rewrites your query into one or more targeted queries that it sends those providers." Google's Search Central documentation says "Both AI Overviews and AI Mode may use a "query fan-out" technique", which it describes as "issuing multiple related searches across subtopics and data sources".

Peec records the searches ChatGPT and Perplexity ran for many of their answers (none for Google AI Overviews). When the question named an engine, ChatGPT put an engine name into its searches in all 151 answers where searches were recorded, and Perplexity in 141 of 143. When the question named no engine, ChatGPT did so in 30 of 239 answers and Perplexity in none of 216.

One pair from 20 November shows the difference. For "Best AI search agencies", ChatGPT searched "best AI search optimization agencies GEO AEO agencies 2026 generative engine optimization". For "Best agency for getting cited in ChatGPT", it searched "best generative engine optimization GEO agencies ChatGPT citations 2026 case studies". Same category, plus the engine's name and "case studies".

Across all answers, engine-named questions gave a smaller share of citations to listicles. Share of each engine's citations that went to pages Peec classes as listicles:

EngineEngine namedNo engine named
ChatGPT58.1%70.7%
Perplexity56.0%77.1%
Google AI Overviews55.4%59.0%

Google AI Overviews gave 14.1% of its engine-named citations to pages Peec classes as how-to guides, a class outside its top five on the plain questions. And ChatGPT and Perplexity cited the platforms' own documentation: engine-named answers cited Google's Search Central documentation 24 times (20 on ChatGPT, 4 on Perplexity) and OpenAI's documentation 3 times (2 on ChatGPT, 1 on Perplexity). Answers to the plain questions cited neither.

If the four leaders owe much of their presence to big listicles, an engine that cites fewer listicles would name them less often. That fits the data, but we did not test it page by page.

Does it hold question by question?

Mostly, yes. ChatGPT's tracked agencies per answer ranged from 3.07 to 6.73 across the eight engine-named questions (median 4.20) and from 4.53 to 7.87 across the twelve others (median 6.84). Its weakest plain question still named more agencies than its median engine-named question.

The groups are not identical. Five of the eight engine-named questions mention SaaS or B2B, against six of the twelve plain ones, and among the plain questions, the SaaS-qualified ones name fewer tracked agencies on every engine. So we split each group by question type. The drop holds in all six engine-by-type comparisons:

Question typeChatGPT (named vs not)PerplexityGoogle AI Overviews
Questions naming SaaS or B2B (5 vs 6)4.37 vs 5.671.53 vs 2.483.44 vs 4.01
Generic questions (3 vs 6)4.24 vs 7.362.07 vs 4.422.76 vs 4.24

Results per question (tracked agencies per answer, 30 answers per engine unless shown, and ChatGPT's three most-named agencies out of 30 answers):

QuestionChatGPTPerplexityGoogle AI OverviewsChatGPT's top three
Best agency for getting cited in ChatGPT4.170.872.40First Page Sage 30, Siege Media 21, Graphite 19
Which agencies will help a B2B SaaS show up in ChatGPT answers?6.733.474.67Omniscient 29, iPullRank 22, Skale 20
Can you recommend an agency that helps B2B SaaS companies get mentioned in ChatGPT answers?3.070.732.83Derivatex 16, Breaking B2B 15, Omniscient 13
Best agency for ChatGPT and Perplexity citations4.132.433.00First Page Sage 26, iPullRank 18, Omniscient 16 (Siege Media also 16)
Best agency for getting mentioned in Perplexity and ChatGPT4.432.902.87First Page Sage 26, Omniscient 23, iPullRank 22
Best agency for getting B2B SaaS cited in ChatGPT and Claude4.231.233.59 (29)Omniscient 27, Discovered Labs 14, Directive Consulting 12
Best agency to get a SaaS product cited in Google AI Overviews4.601.603.07Omniscient 28, Siege Media 21, Skale 13
Agency that optimizes SaaS sites for Gemini and AI Overviews3.230.603.03Veza Digital 20, PipeRocket Digital 12, Omniscient 8
Best AI search agencies7.533.632.37Omniscient, First Page Sage, Siege Media, iPullRank 30 each
Best GEO agencies6.904.104.90 (29)
Top GEO agencies 20267.475.105.63
Best AEO agencies 20267.334.934.45 (29)
Best AI visibility agencies7.073.203.31 (29)
Best LLM SEO agencies7.875.574.88 (24)
Best agency for AI search optimization for SaaS5.572.603.72 (29)
Agency that helps SaaS companies rank in AI search6.000.874.60
AI search marketing agency for SaaS companies6.773.833.97
Who is the best AI search agency for B2B SaaS companies in 2026?5.330.974.27
Generative engine optimization agency for B2B SaaS5.832.433.62 (29)
What agencies specialize in generative engine optimization for SaaS? Give me a shortlist I can actually contact.4.534.173.83

Perplexity is the exception to a clean split. Two plain questions ("Agency that helps SaaS companies rank in AI search" at 0.87 and "Who is the best AI search agency for B2B SaaS companies in 2026?" at 0.97) sit as low as its engine-named ones. Its group averages still differ by half.

Limitations

  • 20 questions, one project. These are the agency questions we track in Peec AI, in US English, over 30 days. Other categories may behave differently.
  • Wording differs beyond the engine name. Engine-named questions also lean on phrases like "getting cited". This is an observational comparison of two groups of real tracked questions. We did not rewrite one question with and without the engine.
  • The tracked list is not the market. We count 53 agencies Peec tracks. Answers we read also named untracked firms (Kalicube, Red-engage, Catalyst), so "fewer tracked agencies" is not the same as "fewer firms".
  • Google AI Overviews searches are not recorded. The search-query finding covers ChatGPT and Perplexity only.
  • Peec's page classes. Listicle and how-to shares use Peec's own classification of each cited page.

If you are shortlisting an agency

The way you ask changes how many tracked agencies you see and which ones. Two cheap checks before you trust one answer:

  1. Ask both ways. Ask for the best AI search agencies, then ask which agency can get you cited in ChatGPT. On these questions the second kind named fewer tracked agencies on every engine and was more likely to surface a specialist.
  2. Ask on more than one engine. Perplexity named no tracked agency at all in 30.0% of its answers to engine-named questions. One answer with no tracked agency in it says more about the question than about the market.

Our earlier studies found the same pattern from other angles: the way you phrase a question and asking on a different day both change the list.

If you want the engines to name you

On ChatGPT, questions that name no engine get the four leaders in 69.4% to 96.4% of answers, and the specialists rarely appear. Buyers who name an engine see a different mix there: iPullRank falls to 42.1%, Siege Media to 42.5% and First Page Sage to 45.8%, while Derivatex rises from 1.7% to 12.9%, PipeRocket Digital from 0.6% to 7.1% and MADX Digital from 0.8% to 4.2%.

So track both phrasings. A prompt set with no engine-named questions overstates how often the leaders reach buyers who ask that way and understates the specialists. And write for the search the engine actually runs: when the buyer names an engine, ChatGPT's own searches include an engine name, so publish pages that answer the engine-named question directly. Our guide to the searches behind AI answers covers how to read them. If you want help being named on both kinds of question, see our GEO agency service, or start with a free AI visibility audit.

Frequently asked questions.

Answers to the questions readers ask most about this topic.

Does naming ChatGPT in a question change which agencies ChatGPT recommends?
In our study it did. On questions that named an AI engine, ChatGPT named 4.33 of the 53 tracked agencies per answer, against 6.52 on questions that named none (23 October to 21 November 2026). It named Omniscient in 65.0% of engine-named answers and 96.4% of the others.
Which engine changed the most?
Perplexity. It named 1.73 tracked agencies per answer on engine-named questions against 3.45 on the others, 50% fewer, and named no tracked agency at all in 30.0% of its engine-named answers, against 10.3%. Its answers kept the same length, a median of 48 words against 49. Google AI Overviews changed the least: 3.18 tracked agencies per answer against 4.12, 23% fewer.
Which agencies lose when a buyer names an engine?
The leaders of the plain answers. With an engine named, ChatGPT named iPullRank in 42.1% of answers against 74.4%, Perplexity in 26.3% against 45.8% and Google AI Overviews in 15.1% against 34.7%. Omniscient fell from 96.4% to 65.0% on ChatGPT, from 57.2% to 42.5% on Perplexity and from 54.7% to 38.5% on Google AI Overviews. Siege Media fell from 75.0% to 42.5%, from 41.1% to 17.9% and from 29.5% to 14.2% on the same three engines.
Which agencies gain when a buyer names an engine?
A few specialists. Derivatex appeared in 12.9% of ChatGPT's engine-named answers against 1.7% of the others, in 25.0% against 8.3% on Perplexity and in 15.9% against 7.4% on Google AI Overviews. PipeRocket Digital rose from 0.6% to 7.1% on ChatGPT, from 0% to 1.3% on Perplexity and from 3.2% to 12.1% on Google AI Overviews. MADX Digital rose from 0.8% to 4.2% on ChatGPT, from 0% to 4.6% on Perplexity and from 2.3% to 6.3% on Google AI Overviews.
What do the engines search for when the question names an engine?
An engine name, in the searches Peec recorded. When the question named an engine, ChatGPT put an engine name into its web searches in all 151 answers where Peec recorded searches, and Perplexity in 141 of 143. Engine-named answers gave a smaller share of citations to listicles: 58.1% of ChatGPT's citations on engine-named questions against 70.7% on the others. The study compares two groups of questions, so it shows a pattern and does not prove a cause.
How was the study run?
We took 20 agency questions we track in Peec AI, 8 that name an AI engine and 12 that do not, and read every daily answer from ChatGPT, Perplexity and Google AI Overviews for 30 days: 1,788 answers. An agency counts as named when Peec's brand detection lists it; a text search of every answer agreed within 0.3 points for the 13 agencies we report on most.
Written by
Founder & Head of Growth

Arnel Bukva is the founder of LoudFace, a B2B SaaS organic growth agency that ships SEO and AEO (Answer Engine Optimization) programmes for B2B SaaS and fintech companies, Series A to C. His work focuses on AI-cited content systems that move pipeline rather than vanity traffic, with named client outcomes including Toku (consistently the top-cited vendor on stablecoin payroll prompts in AI search) and TradeMomentum (a major climb in organic impressions). One of the earliest Webflow users (2017), he has spent the past several years at the intersection of technical SEO and AI search, building the prompt-graph methodology LoudFace uses across every client engagement.

On the record
Published
Nov 22, 2026
Category
Marketing
Reading time
12 min read
A 30-minute strategy call

Ready to grow your business?

Your visibility in AI answers now
ChatGPT, Perplexity, Gemini and AI Overviews
Which pages they read
The sources behind the answers in your category
What we would do first
A written scope within 48 hours
Prefer email? Write to contact@loudface.co
We reply within two hours during working hours