The short answerIn our data, yes. Across 1,080 answers to 24 B2B SaaS agency questions (6 to 20 October 2026), questions that named a funding stage, a startup or a budget got 3.21 tracked brands per answer, against 4.63 for plain questions. First Page Sage was named in 30.9% of plain answers and 8.7% of stage or budget answers.
| Share of answers naming the agency (6 to 20 Oct 2026) | Plain questions (446 answers) | Stage or budget questions (574) | Series A questions (221) | Enterprise question (44) |
|---|---|---|---|---|
| Omniscient | 81.4% | 59.6% | 67.9% | 79.5% |
| Skale | 33.6% | 28.6% | 38.0% | 18.2% |
| First Page Sage | 30.9% | 8.7% | 6.3% | 59.1% |
| Directive Consulting | 27.4% | 14.1% | 12.2% | 70.5% |
| SimpleTiger | 22.9% | 15.7% | 20.8% | 13.6% |
| LoudFace | 21.7% | 16.6% | 23.1% | 18.2% |
| iPullRank | 20.4% | 7.1% | 7.2% | 52.3% |
| Siege Media | 19.7% | 12.2% | 17.2% | 43.2% |
| Tracked agencies named per answer | 4.63 | 3.21 | 3.44 | 5.64 |
If you research agencies by typing "best AI search agency for SaaS" into ChatGPT, you are reading a shortlist written for nobody in particular. The questions that named a stage or a budget got different lists.
How we ran the study
We track buyer questions for LoudFace in Peec AI, which sends the same prompt text to ChatGPT, Perplexity and Google AI Overviews once a day and records each answer, its sources and the brands it names.
For this study we split 24 of those questions into two groups.
- 10 plain questions. Category questions with no company detail, such as "Who is the best AI search agency for B2B SaaS companies in 2026?", "Best AEO agency for B2B SaaS 2026" and "Top B2B SaaS SEO agencies in 2026".
- 13 stage or budget questions. Five name a Series A ("Recommend an agency for AI search visibility for a Series A SaaS company."). Three name a Series B or an ARR band ("Agency for $5M to $20M ARR SaaS scaling organic growth"). Two say "startup". Three name a monthly budget ("Is $5,000 a month reasonable for an SEO and AI search agency for a SaaS startup? Which agencies are in that range?").
One more question, "Which AI search agencies work with SaaS companies that sell to mid-market and enterprise IT buyers?", we report on its own, because one question is a small sample.
Each question ran once a day for 15 days, 6 to 20 October 2026, in US English, on all three engines. That is 1,080 answers. On 16 of the Google runs no AI Overview text appeared (one of them returned only the heading), so every rate below counts the 1,064 answers with content: 446 for the plain questions, 574 for the stage or budget questions and 44 for the enterprise question.
"Named" means one of the 53 brands Peec tracks in our project appears in the answer, almost all of them agencies. An agency outside that list does not count. Low-budget answers named agencies outside it, such as HawkSEM and Arobis AI, and 21 of the 134 low-budget answers named none of the 53.
On Series A questions, First Page Sage falls from 30.9% to 6.3%
Four of the five agencies with the largest gaps in percentage points between the plain and Series A questions (First Page Sage, Directive Consulting, Omniscient and iPullRank) were each named in more than half of the answers to the enterprise question. The fifth, Derivatex, fell from 15.7% to 2.3% and was named in 6.8% of enterprise answers.
First Page Sage was named in 30.9% of answers to the plain questions (138 of 446). On the Series A questions it was named in 6.3% (14 of 221). On the Series B and scale questions it was named twice in 131 answers, 1.5%.
iPullRank followed the same path: 20.4% on the plain questions, 7.2% on Series A, 3.1% on Series B and scale. Directive Consulting fell from 27.4% to 12.2% on Series A.
First Page Sage's rate on each of the five Series A questions was 2.2%, 4.4%, 0%, 11.1% and 14.0%. On the three Series B and scale questions it was 2.3%, 2.3% and 0%. All eight sit below its pooled 30.9% on the plain questions. The groups overlap at the edges: on the 10 plain questions it ranged from 6.7% to 60.0%, so the two highest Series A questions sit above the two lowest plain ones (6.7% and 8.9%). No single question carries the result.
Who kept their place? Omniscient was still the most-named agency on Series A questions at 67.9%, down from 81.4%. Skale went up, from 33.6% to 38.0%, with the rise on ChatGPT and Perplexity. On Google AI Overviews it slipped from 45.9% to 40.8%. SimpleTiger and Siege Media slipped a few points.
The engines also named fewer tracked agencies per answer on the questions that named a stage: 3.44 on the Series A questions and 3.34 on Series B and scale, against 4.63 on the plain questions. Agencies we do not track may have taken some of those places. We did not measure that, so read the gap as fewer tracked agencies. It does not prove a shorter list.
Series B questions: First Page Sage and iPullRank fall further
Moving from Series A to Series B did not bring First Page Sage or iPullRank back. First Page Sage fell further, to 1.5%, and iPullRank to 3.1% on the Series B and scale questions. Directive Consulting recovered part of its drop, from 12.2% on Series A to 19.8%.
The tracked agencies named most often on these questions were Omniscient (78.6%), Skale (36.6%), Powered by Search (35.9%, 47 of 131), and Directive Consulting and Omnius (19.8% each, 26 of 131).
Powered by Search's rate comes mostly from one question. 31 of its 47 namings were answers to "Best organic growth agency for Series B SaaS companies" (72.1% of that question's answers). On the Series B pipeline question it was named in 11.6% and on the ARR question in 24.4%. Three questions with no stage in them named it more often than either: "Top B2B SaaS SEO agencies in 2026" in 51.1% of answers, the enterprise question in 34.1%, and "Best B2B SaaS organic growth agency 2026" in 31.1% (14 of 45). That last one is similar in wording to the Series B question but names no stage, and it named Powered by Search far less often. The two questions still differ in more than the stage, so with this design we cannot separate the wording of a question from its stage, and the effect sits in one question.
One enterprise question names First Page Sage and iPullRank again
The one question about mid-market and enterprise IT buyers named 5.64 tracked agencies per answer, against 4.63 for the plain questions and 3.21 for the stage or budget questions. On ChatGPT it was 7.6 and on Perplexity 4.87. Google AI Overviews did not follow: 4.36 per enterprise answer, against 4.90 on its plain questions.
First Page Sage and iPullRank, at 6.3% and 7.2% on Series A, were named in 59.1% and 52.3% of enterprise answers. Both were above their Series A rates on every engine. Against the plain questions the picture is uneven: Perplexity named iPullRank in 1 of 15 enterprise answers, below its 15.3% plain rate, and Google AI Overviews named First Page Sage in 28.6% (4 of 14), close to its 26.0% plain rate. On ChatGPT, First Page Sage was named in all 15 enterprise answers.
Directive Consulting went from 12.2% on Series A to 70.5%. XQL was named in 56.8% and Siege Media in 43.2%.
XQL shows how far one question can move an agency. It was named in 4.7% of answers to the plain questions, in 0.5% of the stage or budget answers, and in 56.8% of the answers to this question, all of them on ChatGPT and Perplexity. With one enterprise question, we cannot say whether that holds for enterprise buyers in general.
Skale moved the other way: 38.0% on Series A, 18.2% on the enterprise question. That drop came from ChatGPT (45.3% to 13.3%) and Perplexity (28.0% to none of 15). On Google AI Overviews it held, 40.8% on Series A and 42.9% (6 of 14) on the enterprise question.
Budget splits the answer in two
The budget questions did not behave as one group.
At $10k to $20k a month, the answer looks like the plain question. On "We have $10k to $20k a month. Which AI search agencies are worth it at that budget for B2B SaaS?", Omniscient was named in 95.6% of answers, First Page Sage in 37.8% and iPullRank in 35.6%.
On the three low-budget questions, the plain-question leaders mostly disappear. Two of them set a budget under $10k a month (a retainer under $10k, and $5,000 a month). The third asks about a startup that "can't afford an enterprise retainer" and names no figure. Across the three, the engines named 2.08 tracked agencies per answer. Omniscient was named in 7.5% (10 of 134): 11.1% on the two questions with a figure (10 of 90) and in none of the answers to the third. 21 of the 134 answers named none of the 53 brands we track. The tracked agencies named most often were PipeRocket Digital (20.1%), SimpleTiger (16.4%), Optimist (16.4%), TripleDart (15.7%) and Derivatex (13.4%).
We counted dollar figures on one low-budget question. On "Which AEO agencies are best for a startup that can't afford an enterprise retainer?", whose wording contains no dollar figure, 43 of the 44 answers with content contained a dollar sign followed by a number. That was 15 of 15 on ChatGPT, 14 of 15 on Perplexity and 14 of 14 on Google AI Overviews. The count includes general market ranges as well as an agency's own price.
On 18 October, Perplexity answered that question with HawkSEM (a local GEO package from $1,550 a month), Arobis AI (a $2,500 a month starter plan) and Optimist (from $3,000 a month). Only Optimist is one of our 53 tracked brands. ChatGPT answered the $5,000 question by naming Derivatex first, with its published $5,000 a month entry plan.
Our recommendation, which this study did not test: if you want budget-limited buyers to find you, put your price where an engine can read it. We publish ours on the pricing page: engagements start from $5k a month. For a list of who else does, see which AI search agencies publish their pricing.
Do ChatGPT, Perplexity and Google AI Overviews react the same way?
No. All three named fewer tracked agencies and named First Page Sage and iPullRank less often, but by different amounts.
| Plain questions vs stage or budget questions | ChatGPT | Perplexity | Google AI Overviews |
|---|---|---|---|
| Tracked agencies named per answer | 5.66 vs 4.04 | 3.35 vs 2.24 | 4.90 vs 3.36 |
| First Page Sage named | 43.3% vs 10.3% | 23.3% vs 8.7% | 26.0% vs 7.1% |
| iPullRank named | 24.7% vs 10.8% | 15.3% vs 2.1% | 21.2% vs 8.7% |
| Omniscient named | 92.0% vs 69.2% | 72.0% vs 51.8% | 80.1% vs 57.6% |
| Top five places shared by both groups | 4 of 5 | 1 of 5 | 4 of 5 |
Perplexity changed its shortlist the most. Its top five on the plain questions was Omniscient, Directive Consulting, RevenueZen, First Page Sage and Derivatex. Its top five on the stage or budget questions was Omniscient, Skale, Omnius, SimpleTiger and Minuttia. Only Omniscient survived.
ChatGPT and Google AI Overviews kept most of their top five and changed further down the list.
This fits what we found on 15 October: in the 500 answer sets where all three engines named an agency, 71% of the agencies they named appeared in only one engine's answer. That study also found the engines read different sources. Our reading, which this study does not test, is that different sources explain why each engine reacts to a buyer's detail in its own way.
Which SEO agency is best for a Series A startup?
On the five Series A questions, the engines most often named Omniscient (67.9% of answers), Skale (38.0%), LoudFace (23.1%), SimpleTiger (20.8%) and NoGood (18.1%), followed by Animalz and Siege Media at 17.2% each. That is what the engines say, which is not the same as which agency fits your company.
A Series A founder or marketing lead building a shortlist from AI answers can use this in three ways.
- Ask the way your company is. Put your stage and your budget into the question. The plain questions returned lists in which First Page Sage was named in 30.9% of answers. On the Series A questions it was named in 6.3% and iPullRank in 7.2%.
- Ask more than once, on more than one engine. On 19 October we found that ChatGPT named an identical list on consecutive days 2.9% of the time. The engines also differed in how far their lists moved on the questions that named a stage. One answer is one draw.
- Check the price before the call. Some low-budget answers quoted starting prices, such as Derivatex's $5,000 a month entry plan on ChatGPT. Read the pricing page yourself and confirm the scope, because an engine repeating a starting price does not tell you what that price buys. Our questions to ask an AI search agency before you sign cover the rest.
What this changes about measuring AI visibility
If you track your brand on generic questions only, you are measuring a different list from the one a buyer with a stage or a budget sees. Across all three engines, LoudFace was named in 21.7% of answers to the plain questions and 23.1% of answers to the Series A questions, while First Page Sage fell from 30.9% to 6.3% on the same two groups.
- Track the questions your buyers ask, with their detail in them. A Series A founder and an enterprise IT buyer get different lists. Track a version of your main question for each buyer you sell to.
- Report each engine on its own. Perplexity shared one of five top places between the plain and qualified questions. ChatGPT and Google AI Overviews shared four. A blended number hides that.
Then work on the pages behind the qualified question. In an earlier study of 2,895 answers, 624 of 631 namings of LoudFace were in answers that had read our site. That study measured one brand, so applying it here is our inference: to be named on a Series A question, be on the pages an engine reads when it answers one.
Limits of this study
- Tracked brands only. We count the 53 brands in our Peec project. Low-budget answers named agencies outside that list, so a lower count of tracked agencies does not prove a shorter list.
- Different wording. The stage or budget questions are worded differently from the plain ones. Some of each gap may come from wording rather than the stage or the budget. First Page Sage stayed below its 30.9% plain-question rate on all eight Series A and Series B questions, which argues against one question driving the drop. Powered by Search's Series B rate, by contrast, comes mostly from one question, and a similarly worded question without a stage named it far less often. The two differ in more than the stage, so we cannot separate wording from stage there.
- One enterprise question. The enterprise result rests on 44 answers to a single question.
- One run per engine per day, in US English, from Peec's accounts. Logged-in users with memory or a different location may see other answers.
- The dollar count covers one question and counts any dollar figure, including general market ranges.
- 15 days. The answers change daily, so a later window will move the exact numbers.
Questions that name a stage or a budget got different shortlists from the plain questions. Track the version your buyers ask, engine by engine, because that is the list they will call from.
