# Index: Which AI Search Visibility Tool Does AI Recommend? November 2026

> I asked 5 AI tools 18 buyer questions about AI search visibility tools, 3 times each. AI picked a tool in only 16% of 270 answers. Here is who it picked.

By Michael Zacs · Index Reports · Published Nov 4, 2026 · Last updated Nov 4, 2026
Source: https://exibran.com/blog/index-ai-search-optimization-tools-november-2026

## Quick answer

I asked ChatGPT, Gemini, Google AI Mode, Google AI Overviews and Claude 18 buyer questions about AI search visibility tools that name no tool, 3 times each. AI picked a tool in only 16% of the 270 answers. Otterly.AI, Peec AI and Semrush led, but their ranges overlap, so I call it a three-way tie.

## Key takeaways

- AI picked a tool in 16% of 270 answers (12% to 20%). 57% listed options and stopped. 27% named no tool.
- Otterly.AI (7%), Peec AI (6%) and Semrush (6%) lead, but their ranges overlap. It is a tie, not a ranking.
- Profound was mentioned first in 14% of answers but picked in 3%. Of the 42 picks, 22 were the tool the answer named first.
- The name effect is huge. When the question names the tool, AI picked it in 33% to 67% of answers. When it does not, 3% to 7%.
- Claude picked a tool in 31% of answers. Google AI Overviews picked in 9%, and Gemini and AI Mode in 7%.
- AI gives each tool a job title: budget, in between, enterprise, or "already use the suite".
- Only "best tool" questions get picks: 41% against 6% for everything else. And answers are unstable: 66 of 90 app and question pairs never picked a tool in any of 3 runs.

I asked five AI assistants which AI search visibility tool to buy. They mostly would not say.

*Image: Four numbers: AI picked a tool in 16% of answers, 57% listed options and stopped, 27% named no tool, 3 tools tied at the top*

*The Index in four numbers.*

> **The Short Answer**
>
> AI picked a tool in **16%** of 270 answers (range 12% to 20%). Otterly.AI, Peec AI and Semrush lead, but they are statistically tied. When a question names a tool, AI picks it far more often. This is the first Index. Method, data and limits are public.

These tools are also called AI search optimization tools. This Index measures [AI Pick Rate](/blog/ai-pick-rate): the share of AI answers whose own verdict chooses a tool, not just mentions it.

I run it for AI search optimization tools because that is the market I work in, and a tool buyer asks AI the same way your buyer does.

## How Was This Index Made?

### 30 buyer questions, 5 AI apps, 3 runs each

I asked ChatGPT, Gemini, Google AI Mode, Google AI Overviews and Claude every question three times on 3 November 2026, from the United States, in English. That is 450 answers, collected with my own tool, Exibran Radar.

Every headline number here uses only the **18 questions that included no tool's name**. That is 270 answers. The other questions name a tool or name Aiso, and I report them separately.

The [AI Pick Rate guide](/blog/ai-pick-rate) shows a 405-answer table. That table includes questions that name tools. This Index uses the cleaner 270.

The full method, the limits and how I fix mistakes are on the [methodology page](/methodology). The [raw data](/data/ai-search-optimization-tools-index-2026-11.xlsx) has every question and answer.

## Which Tool Did AI Pick Most?

### Otterly.AI, Peec AI, and Semrush, in a three-way tie

*Image: Bar chart of pick rates: Otterly.AI 7.4%, Peec AI 5.9%, Semrush 5.6%, Profound 3.0%, Scrunch AI 3.0%, Ahrefs 2.6%*

*Share of 270 neutral answers where AI picked the tool. The line is the 95% range.*

| Tool | Picked | Pick Rate | 95% Range | Mentioned |
| --- | --- | --- | --- | --- |
| Otterly.AI | 20 | 7% | 5% to 11% | 40% |
| Peec AI | 16 | 6% | 4% to 9% | 36% |
| Semrush | 15 | 6% | 3% to 9% | 37% |
| Profound | 8 | 3% | 2% to 6% | 36% |
| Scrunch AI | 8 | 3% | 2% to 6% | 19% |
| Ahrefs | 7 | 3% | 1% to 5% | 28% |

Ranges overlap, so I do not call Otterly.AI the winner. With three asks per question, a gap this small can be noise.

**Mentioned** is how often the tool appeared anywhere in an answer.

Profound is mentioned about as often as Peec AI and Semrush, and picked about half as often or less. *Being known is not the same as being chosen.*

## How Crowded Does the AI Visibility Market Look to AI?

### 167 tools, about 4 per answer

Across the 270 neutral answers, AI named 167 different products. 66 of them appeared in only one answer.

The top six took 49% of all mentions, and the average answer mentioned 4 of them. A small product can be known and still be part of a long tail.

## How Often Did AI Choose Anything at All?

### In 16% of answers

Of the 270 answers, 42 picked a tool. 154 listed options and stopped. 74 named no tool. A buyer who asks AI "which tool should I use" usually gets a list and a shrug.

## Does Saying the Tool's Name Change the Answer?

### Yes, by a lot

*Image: Paired bars showing pick rate when the question names the tool versus when it does not, for six tools*

*Share of answers where AI picked the tool, when the question names it and when it does not.*

| Tool | Picked When The Question Names It | Picked When It Does Not |
| --- | --- | --- |
| Otterly.AI | 67% (10 of 15) | 7% |
| Ahrefs | 60% (9 of 15) | 3% |
| Profound | 55% (33 of 60) | 3% |
| Scrunch AI | 53% (8 of 15) | 3% |
| Peec AI | 51% (23 of 45) | 6% |
| Semrush | 33% (10 of 30) | 6% |

Some of those questions are "X vs Y", which forces a verdict, so the jump is partly built in.

Still, the gap is the point. If buyers do not already search for you by name, AI rarely brings you up as the pick.

## Do the AI Platforms Agree?

### No. Claude picks far more than Google's AI

*Image: Bar chart of how often each AI app picked a tool: Claude 31%, ChatGPT 22%, Google AI Overviews 9%, Gemini 7%, Google AI Mode 7%*

*Share of answers where each app picked a tool. 54 neutral answers per app.*

Each app had 54 neutral answers, so each range is wide. Claude was collected through its API with web search on, and AI Overviews through a scraper.

They may differ from what a person sees in the apps.

## Which Questions Get a Pick (Recommendation)?

### Only the "best tool" ones

When the question asks for the best or the right software, AI picked one in 41% of answers (31 of 75, range 31% to 53%).

For every other kind of question, it picked in 6% (11 of 195, range 3% to 10%). That covers how-to, cost, accuracy and "should we" questions.

The highest was the startup on a small budget question, at 67% (10 of 15). Eight questions got no pick at all, including the cost, accuracy and reporting ones.

## Who Gets Mentioned First?

### Profound and Otterly.AI, but first is not the same as picked

Radar saves the tools in the order each answer mentions them. In 196 of the 270 neutral answers, AI mentioned at least one tool.

This table shows how often each of the top five came first, next to how often it was named anywhere and how often it was picked. All shares are out of 270 answers.

| Tool | Mentioned First | Mentioned Anywhere | Picked |
| --- | --- | --- | --- |
| Profound | 14% (38) | 36% (97) | 3% (8) |
| Otterly.AI | 12% (32) | 40% (109) | 7% (20) |
| Semrush | 7% (19) | 37% (101) | 6% (15) |
| Peec AI | 5% (13) | 36% (96) | 6% (16) |
| Ahrefs | 2% (6) | 28% (76) | 3% (7) |

Profound was mentioned first more often than any other product, but AI picked it in only 8 answers. Otterly.AI was named first in 32 and picked in 20.

Profound and Otterly.AI are close enough on "mentioned first" that I would call it a tie. Ahrefs shows the opposite pattern: mentioned in more than a quarter of answers, but first in only 6.

Being first does help. In 22 of the 42 answers that picked a tool, the pick was the tool the answer named first. In the other 20, AI picked a tool that came later in the list.

The order is the sequence in the saved answer, not a ranking AI gave, and it changes from ask to ask. Treat it as a clue, not a result. The order is in the "Tools named, in order" column of the data file.

## How Stable Are the AI Answers?

### Not very

I asked every question three times on each app, which gives 90 app and question pairs. In 66 of them, AI never picked a tool in any of the 3 runs.

In the other 24, it picked in only 1 of 3 runs for 12 pairs, in 2 of 3 for 6 pairs, and in all 3 for 6 pairs. None of those last 6 gave the exact same pick three times.

The tool lists moved too. Two runs of the same question shared about 35% of the tools they mentioned. This is why I report ranges and ties, and why one screenshot proves little.

*Image: Grid of 90 dots: 66 never picked in 3 runs, 12 picked in 1 of 3, 6 picked in 2 of 3, 6 picked in all 3*

*Each dot is one question on one AI app, asked three times.*

## How Does AI Describe Each Tool?

### It gives each one a job title

*Image: A scale from budget to enterprise with Otterly.AI at the budget end, Peec AI in the middle and Profound at the enterprise end, and Semrush and Ahrefs placed off the scale as tools for people who already use an SEO suite*

*Phrases quoted from saved answers. These are AI claims, not facts.*

I ran simple word rules over the sentences about each product. The counts are approximate, and they show what AI says.

- **Otterly.AI:** budget or small-team wording in about 44 of 109 answers that mentioned it.
- **Peec AI:** budget wording in about 17 of 96, often set against enterprise tools.
- **Profound:** enterprise or premium wording in about 36 of 97.
- **Semrush:** "if you already use an SEO suite" wording in about 51 of 101.
- **Ahrefs:** the same suite wording in about 33 of 76.

Real objections were thin. Ahrefs was called keyword-derived rather than real AI conversations in 3 Claude answers.

Peec AI was called monitoring only in 4. AI mostly sorts tools by who they suit, not by what is wrong with them.

## Which Sites Does AI Cite the Most?

### Reddit first, then the tools' own sites

*Image: Bar chart of the most cited sites: reddit.com 20%, tryprofound.com 14%, otterly.ai 12%, semrush.com 12%, youtube.com 11%*

*Share of the 270 answers that cited each site. 241 answers showed sources.*

Reddit appeared in 20% of answers. The own sites of Profound, Otterly.AI, Semrush and Peec AI came next, with YouTube in the middle and LinkedIn close behind.

AI is reading what vendors say about themselves, and what buyers say about them.

## What Kinds of Sites Does AI Cite?

### Tool sites lead, forums and videos follow

Across the 241 answers that showed sources, AI cited 1,147 sites, counted once per answer.

| Site Type | Citations | Share | Answers Citing It |
| --- | --- | --- | --- |
| Tool vendor sites | 478 | 42% | 80% |
| Other sites (small blogs, agencies, tools I could not match) | 453 | 39% | 74% |
| Video and social (YouTube, LinkedIn) | 71 | 6% | 20% |
| Forums and Q&A (Reddit, Quora) | 66 | 6% | 24% |
| Publishers and SEO media | 40 | 3% | 16% |
| Research and code | 22 | 2% | 7% |
| Big platforms | 17 | 1% | 6% |

Reddit is a Google story here. It appeared in 53% of Google AI Overviews answers and 52% of AI Mode answers that showed sources.

ChatGPT, Gemini and Claude cited it 0 times in 138 answers. The tools' own sites led on every app: 93% on Gemini, 90% on Claude, 81% on ChatGPT, 75% on AI Overviews and 62% on AI Mode.

The tool vendor share is a floor, because some "other" sites are also tools I could not match by name.

Radar saves the site, not the page link, so I cannot say which pages or formats AI cited. The Google section below shows what the pages look like in search.

## Does AI Mention the Tools Whose Sites It Cites?

### The big six, yes. Smaller tools, often not

Of 478 citations to a tool's own site, AI mentioned that software in 335 answers (70%) and picked it in 43 (9%).

For the six biggest tools, almost every citation came with a mention. Smaller tools were different.

Their sites were cited 321 times, and the tool was named 192 times (60%) and picked 18 times (6%). In the other 129 answers, AI cited the tool's site and left the tool out.

*Image: Bars for Siftly, Dageno AI, SE Ranking, Searchable, LLM Pulse and Trysight showing how often AI cited the tool's site but did not mention the tool*

*Answers that cited the tool's own site, split by whether the tool was mentioned.*

| Tool | Answers Citing Its Site | Mentioned | Picked |
| --- | --- | --- | --- |
| Siftly | 15 | 3 | 0 |
| Dageno AI | 13 | 7 | 0 |
| SE Ranking | 11 | 5 | 0 |
| Searchable | 10 | 1 | 0 |
| LLM Pulse | 9 | 3 | 0 |
| Trysight | 5 | 0 | 0 |

I did not open the cited pages, so I cannot say whether each one listed the tool or only talked about the market.

Still, a site AI keeps citing without mentioning the tool is a gap worth closing.

## What Does Google Show for the Same Questions?

### Mostly articles about the tools, not the tools

On 4 November 2026, I searched all 30 questions in a logged-out private window and recorded the top 3 normal results for each.

That is 90 results, and Google showed an AI Overview on every search.

| Result Type | Results |
| --- | --- |
| Blog posts | 34 |
| "Best tools" lists | 30 |
| Videos | 10 |
| Forum threads | 6 |
| Tool websites | 5 |
| Review pages | 3 |
| News and other | 2 |

Only 5 of 90 were a vendor's own site. Most of the rest were articles written by other companies, often by companies that sell in the same space.

Buyers meet the software through someone else's list, and AI reads those same lists.

## What Happened to Aiso?

### AI never brought it up and knew it when asked

I added Aiso (getaiso.com), a smaller SaaS product, as a test of a challenger. In the 405 answers to questions that did not name it, Aiso appeared **0 times**.

When three questions named it, AI named it in 42 of 45 answers.

When asked "Aiso vs Profound", AI picked Aiso in 10 of 15 answers, usually alongside Profound. Asked about Aiso on its own, AI picked it in 2 of 15 and 1 of 15 answers.

The name is also a problem. Search results for the Aiso questions mixed it up with AISO and AISEO, the general terms for AI search optimization, and that spilled into the AI Overview.

I have no relationship with Aiso, yet. Nobody asked for this report, and nobody paid for it.

## What Does the Data Suggest You Do?

### Three plain moves

1. **Get buyers to search for you by name.** The name effect is the biggest gap in the data. Reviews, comparisons, podcasts and posts that put your name next to the problem all help.
1. **Be on the pages AI reads.** Reddit threads, your own site, YouTube and LinkedIn showed up most. Check what they say about you today.
1. **Say who you are for, in one plain line.** AI sorts tools by role. If your line is vague, AI picks the role for you. Eg., Write for the “best X for [who]” question.

These are readings of one snapshot, not guarantees.

## What Comes Next?

I plan to run this again in mid-December 2026 with the same questions, so the two months can be compared.

I will add Perplexity if I can collect it properly. Under the rules on the [methodology page](/methodology), a move under 7 points will be reported as no change.

If a product is tagged wrongly, or you are a vendor and want to check an answer, email [chat@exibran.com](mailto:chat@exibran.com).

## How Does Your Own SaaS Product Show Up?

I do this for B2B SaaS teams at Exibran. You send your product name, and I send back the real answers from ChatGPT, Gemini, and Perplexity, word for word, with your pick rate and its range.

## Frequently asked questions

### Which AI search visibility tool does AI recommend most?

In my November 2026 test, Otterly.AI, Peec AI and Semrush led, picked in 7%, 6% and 6% of 270 answers. Their ranges overlap, so no one is clearly ahead.

### Why did AI pick a tool in only 16% of answers?

Most answers list options and stop. 57% listed tools without choosing one, and 27% mentioned none. AI often leaves the choice to the buyer.

### Do these results apply to my category?

Not directly. This is one category, one date, one country and one language. The method works for any category, and the numbers will differ.

### Can I see the raw answers?

Yes. The data file has every question, every answer text and my tags. The method is on the methodology page.