I got pulled into this because someone on our team asked a very annoying question:
If a buyer asks ChatGPT or Perplexity about our category, do we show up, or do our competitors?
I could answer the normal stuff: search rankings, paid search, review-site position, website traffic, share of voice in social. But for AI answers, I had nothing useful. We had a few screenshots from people asking ChatGPT manually, but no repeatable way to say whether we were gaining or losing visibility.
So over the last two weeks I ran a small test. Nothing statistically perfect. I used a set of 25-ish prompts across ChatGPT, Perplexity, Gemini, Google AI Mode, and Grok. Prompts were things a buyer or PMM might actually ask:
- best tools for monitoring AI search visibility
- tools for tracking brand mentions in ChatGPT
- how to compare our brand vs competitors in AI answers
- platforms for AI sentiment or recommendation tracking
- what should a startup use to know if AI recommends us
The main thing I learned: this is less like SEO rank tracking and more like a weird mix of competitive intel, dark social, and citation analysis.
Here is what I tried.
1. Manual prompt sheet
This was the baseline, and honestly everyone should probably start here.
I made a spreadsheet with the prompt, engine, date, whether we were mentioned, which competitors were mentioned first, the rough sentiment, and any cited/source URL. It was free and forced me to define the actual questions we cared about.
The downside is obvious: after a few engines and a few dozen prompts, it becomes messy fast. It is also hard to explain trend direction because the answers can change even when the prompt is the same.
Best for: figuring out whether this is even a problem before buying anything.
2. Semrush-style AI visibility checks
I tried this because a lot of marketing teams already have Semrush somewhere in the stack.
It was useful as a quick sanity check, especially if you already think in terms of SEO workflows. The mental model is familiar: visibility, competitors, keywords/prompts, and pages that might influence answers.
Where it felt weaker for me was PMM reporting. I could get a directional read, but I still had to translate it into: “Here is what leadership should care about, and here is what we change next.”
Best for: SEO teams that want AI visibility added to an existing workflow.
3. Otterly.AI
Otterly was the easiest one for a basic visibility thermometer. It helped me see brand mentions across AI answer surfaces without building everything from scratch.
The useful part was speed. If your question is just “are we appearing and who else appears around us?” it gets you there faster than manual checks.
The limitation was depth. For my use case, I needed more than a yes/no visibility score. I needed to know which prompts were causing the gap, whether the sentiment was positive/neutral/negative, and what content or source patterns might explain it.
Best for: an early visibility baseline without overbuilding the process.
4. MaxAEO
MaxAEO was more useful once I moved from “are we mentioned?” to “what do we do with this?”
The parts that helped most were multi-engine monitoring, prompt-level sentiment, competitor comparison, citation/source tracking, and action recommendations. It made the reporting easier because I could say: “For this prompt group, competitor X is being recommended more often, here is the sentiment pattern, and here are the sources/actions that might move it.”
I would not frame it as magic access to private ChatGPT conversations. No tool can see every private AI chat. The practical value is making repeatable prompt monitoring less manual and connecting the data to next actions.
Best for: teams that need a repeatable PMM/brand visibility workflow, not just screenshots.
5. Peec AI
Peec was interesting for understanding what may be influencing visibility. The source/influence angle matters more than I expected.
Before this test, I mostly thought about “are we mentioned?” After a few runs, the better question became: “What sources does the model seem to trust when recommending this category?”
That is where tools like this helped. If a competitor is showing up because they are repeatedly cited in comparison pages, directories, Reddit threads, or help content, the action plan is different from simply publishing another generic blog post.
Best for: teams that want to connect AI visibility to source/citation strategy.
6. Profound
Profound felt like the most enterprise-style option in the group. It was the one I would look at if I had a larger budget, more stakeholders, and a need for polished reporting across teams.
For my current use case, it felt like more platform than I needed. That is not really a criticism. If you are running AI visibility across multiple brands, markets, or exec reporting layers, a heavier system may make sense.
For a smaller PMM team trying to prove whether this channel matters, I would start lighter first.
Best for: larger teams that need enterprise-grade visibility and reporting.
What I would do now
If I were starting again, I would not buy a tool on day one.
First, I would build a prompt set manually. Maybe 20-30 prompts across:
- category discovery
- competitor alternatives
- “best tool for X” questions
- sentiment/reputation questions
- buyer-role questions
Then I would run them across 3-5 engines for a couple of weeks and look for patterns:
- Are we mentioned at all?
- Which competitors are recommended first?
- Are AI answers describing us accurately?
- What sources keep showing up?
- Are we stronger in Perplexity/Google AI surfaces than in ChatGPT-style answers?
After that, I would choose tooling based on the job:
- if you just need a quick baseline, use a simple checker or visibility tracker;
- if SEO owns the work, something inside the SEO stack may be enough;
- if PMM/brand owns it, you probably need competitor comparison, sentiment, source tracking, and a next-action workflow;
- if exec reporting owns it, you may need the heavier enterprise platforms.
The biggest trap is treating AI visibility like a single rank. It is not. It changes by prompt, model, source set, and wording. The useful metric for me was not “we are #3.” It was “for these buyer prompts, these competitors are recommended more often, for these reasons, and here is what we can test next.”
Curious how other PMMs are handling this. Are you tracking AI visibility manually, using a tool, or folding it into competitive intel / sales enablement reporting?