AI search visibility tools: features, evidence and pricing models compared
The tools in this category price four different ways, publish very different amounts of evidence, and mostly agree on nothing. This page compares the pricing models rather than the price tags, says who each suits, applies the same tests to CiteGraph and Profound, and gives you a checklist for validating any of them.
By CiteGraph · Updated 2026-09-07 · Prices checked on the vendors’ own pages
Disclosure: this page is written by CiteGraph. Competitor prices and plan facts are from each vendor’s own pricing page, checked 23 August 2026, and linked below. Where a vendor does not publish something we say “not published”; where we did not test something we say “not verified”. The measured evidence includes us, and we come last.
Four pricing models, and who each one suits
Per prompt tracked
Suits: Teams that know exactly which ten or fifteen questions matter and want them checked daily.
Watch for: The bill scales with curiosity. Adding a second product or a new angle means buying more prompts, and daily single-run checks show a lot of movement that is sampling noise.
Example: Otterly.AI (15 prompts on Lite, 4 engines, daily).
Per project or brand
Suits: Founders and agencies who want a fixed, repeated measurement per domain and a plan per domain.
Watch for: The question count per project is capped, so breadth is traded for depth. Check whether repeated runs and margins are included or whether it is one run dressed as a rate.
Example: CiteGraph (10 questions × 3 to 5 runs × 4 engines per project, weekly).
Base fee plus action credits
Suits: Teams that want the tool to act on their behalf, typically posting replies in communities, and are comfortable with that being done in their name.
Watch for: The visible price is not the price. Credits run out, per-action fees sit on top, and the actions themselves carry reputational risk that the invoice does not show.
Example: CrowdReply ($99 base plus $50 of credits on Starter, replies billed from $7 to $10 each).
Annual billing, quoted enterprise
Suits: Larger marketing teams that budget yearly and want a vendor conversation.
Watch for: Monthly figures are often the yearly rate divided by twelve; the real commitment is the year. Ask what the monthly-billed price is, if one exists.
Example: Profound ($99 and $399 a month billed yearly; Enterprise by sales).
Published prices, in plain text
Entry and top published tiers as listed on each vendor’s pricing page on 23 August 2026 (CrowdReply checked 25 August 2026).
| Tool | Entry plan | Higher published tier | What the entry plan includes |
|---|---|---|---|
| Otterly.AI | Lite: $29/mo | Premium $489/mo (Enterprise from $1,000/mo) | Lite covers 15 search prompts across 4 engines with daily tracking and unlimited team members. |
| Profound | Starter: $99/mo billed yearly | Growth $399/mo billed yearly (Enterprise custom) | Starter and Growth are published; Enterprise is quoted by sales. |
| Peec AI | Starter: not published on the pricing page | not published | Plan names are listed — Starter, Pro, Advanced, Enterprise — but the page did not render prices or quotas when we checked. |
| CrowdReply | Starter: $99/mo | Enterprise from $499/mo | Starter tracks 20 prompts across 2 AI models and includes $50 of engagement credits; Reddit comments are billed on top at $7–$10 each. |
| CiteGraph | Starter: $29/mo | Scale $149/mo | 10 buyer questions × 3 runs × 4 engines per scan, every rate published with its margin. 7-day trial, card required, nothing charged up front. |
Evidence: what AI actually recommends in this category
The one measurement in this comparison that anyone can reproduce. In the week of 2026-08-17 we asked 10 buyer questions about AI visibility tools to 3 production engines, 60 answers in total, and counted which tools were named and which had their own site cited. The full board and the exact questions are on the AI SEO tools leaderboard, re-run weekly.
| Tool | Named in answers | Own site cited |
|---|---|---|
| Otterly | 70% | 27% |
| Peec AI | 55% | 20% |
| Profound | 50% | 23% |
| Semrush | 45% | 10% |
| Ahrefs | 22% | 10% |
| Scrunch | 20% | 10% |
| AthenaHQ | 13% | 0% |
| SE Ranking | 12% | 5% |
| HubSpot AEO Grader | 12% | 10% |
| AirOps | 10% | 3% |
| CiteGraph (us) | 0% | 0% |
CiteGraph and Profound, by the same criteria
The two tools this page is most likely to be read against, judged on the same six points. Profound’s cells hold only what its public pages state; we did not run its trial.
| Criterion | CiteGraph | Profound |
|---|---|---|
| Pricing model | Per project: $29, $69, $149 a month, billed monthly, 7-day trial; $12 single scan | Annual billing: $99 and $399 a month billed yearly; Enterprise by sales |
| Prompt sampling | 10 buyer questions per project, written from the site, editable, fixed between scans | not published |
| Engines | ChatGPT, Claude, Gemini, Perplexity via official APIs with web search | not verified |
| Repeated runs and margins | 3 runs (5 on Growth and Scale); margin printed on every rate | not published |
| Evidence kept | Every answer verbatim; named and cited counted separately | not verified |
| What it tells you to do | Blind spots with the winning sentence; cited pages graded; action plan in order; 15 site readiness checks with fixes | not verified |
Where Profound is the better choice on what is public: a larger team that budgets yearly and wants a vendor relationship with a sales process behind it. Where CiteGraph is: a founder or small agency that wants a measurement it can check line by line, on a monthly bill, with a plan attached.
A reproducible checklist for validating citations across assistants
Run this on any tool, including ours, before you rely on its numbers. It takes about half an hour.
- Take one prompt from the tool’s set. Confirm it is a question a buyer would type, not a keyword.
- Ask that question yourself in each engine the tool claims to cover, with web search on, and save the answers.
- Open the tool’s record for the same question. It should show the verbatim answers, not a summary.
- Compare sources: the domains the tool lists should match the ones you saw, with redirect links unwrapped to the real host.
- Check that “named” (in the text) and “cited” (as a source) are two different numbers.
- Ask the same question again an hour later. If the answer changed and the tool shows one run, the tool is showing you a coin flip.
- Find the margin of error on any rate. If there is none, ask how many answers the rate came from and compute one yourself.
- Pick one question you lose and ask what the tool wants you to do about it. If the answer is a score rather than a page and a sentence, it is a dashboard, not a plan.
The method behind each of those steps, with worked error bars and a downloadable record schema, is in How to measure brand citations in ChatGPT and other AI assistants. The six criteria as a buyer’s framework are in How to choose an AI citation tracking tool.
Common questions
How do AI search visibility tools charge?+
Three models are common: per prompt tracked (you pay for how many questions are monitored), per project or brand (you pay per domain and get a fixed set of questions), and per action (a base fee plus credits for things the tool does on your behalf, such as posting replies). The model decides what a second product or a bigger question set costs you, so it matters more than the entry price.
Which AI visibility tool is cheapest?+
By published entry price, Otterly.AI Lite and CiteGraph Starter are both $29 a month as of the date on this page. They are not the same thing: one prices per prompt with daily tracking, the other per project with repeated sampling and margins. Compare the number of sampled answers you get for the money.
Does Profound publish its prices?+
Yes. As of the check date on this page, Starter is $99 a month billed yearly and Growth $399 a month billed yearly, with Enterprise quoted by sales. Peec AI lists plan names without prices.
How can I validate a visibility tool's numbers myself?+
Take one question from its prompt set, ask the same engine yourself with web search on, and compare: does the tool show the verbatim answer, the same sources, and separate named-versus-cited counts? Then ask for the margin of error. The checklist on this page walks through it.
Is this comparison neutral?+
No, and it says so at the top. It is written by CiteGraph. Every fact about another vendor is a published price or plan from that vendor's own page, dated and linked, and everything we did not verify is marked not verified. The measured table includes us coming last.
See which pages answer your category — and what they say about you.
Check your site →