Profound Review: What $399/Month Actually Buys You

Profound goes much deeper than basic AI visibility monitoring, but the value of its $399 Growth plan is not the headline Visibility Score. The useful parts start when you break performance down by platform, prompt, query fanout, and citation source.

This review covers Profound's $399 Brand Growth plan, not its separate Agency Growth plan. We tested the core workflows inside a live workspace to see which metrics lead to useful decisions, where the product needs context, and which features justify the price.

Quick facts

The pricing distinction matters. The $399 Growth plan reviewed here is a brand plan. Agency Growth starts at $99 per month because the base subscription is designed around prospect audits, Agency Mode, and consolidated billing. Full ongoing client workspaces are separate $399 monthly add-ons.

How we tested Profound

To test Profound, we used a live workspace tracking our tested brand across 10 AI video-generation prompts grouped into one topic. We reviewed a five-day window and 150 executions to assess how the product's core reporting workflows behave with a consistent dataset.

The prompts covered broad and specific buying questions around AI video generation, including general tools, short-video tools, image-to-video tools, and platform-specific use cases.

We evaluated Profound against the $399 Growth plan, which allows 100 tracked prompts, three Answer Engines, daily execution, and 9,000 analyzed responses per month. Our test used 10 of those prompt slots.

Across the selected five-day period, Profound recorded 150 executions, with 15 executions displayed for each prompt. We used that dataset to inspect Visibility Score, Share of Voice, Average Position, platform differences, Prompt Analysis, Query Fanouts, and citation behavior.

The results therefore describe our configured prompt set, not the whole software category or every question a potential buyer might ask.

Ten prompts over five days are enough to evaluate how Profound's metrics and screens behave, but not enough to establish industry benchmarks or long-term visibility trends. We use the results to judge the product, not to define what a normal visibility score should be.

Visibility Score vs Share of Voice: how these numbers help your brand

Profound's Visibility Score tells you how often your brand appears in tracked AI answers, while Share of Voice shows its competitive presence within those answers. Read together, they separate basic coverage from competitive strength and give each metric a clear reporting role.

Visibility Score is primarily a coverage metric. If your brand appears in 50 of 100 qualifying responses, its Visibility Score is 50%.

Share of Voice adds competitive context. AI answers often mention multiple companies, so appearing frequently does not automatically mean your brand owns an equally large share of the conversation.

Our tested brand had a 41.4% Visibility Score and ranked #6 among 98 results. Its Share of Voice was 6.3%, also ranking #6 among 98 results. The first brand in the leaderboard recorded 86.7% Visibility Score and 13.1% Share of Voice.

The 41.4% and 6.3% figures are not two versions of the same metric. Their denominators differ. Visibility tells you how broadly the brand enters relevant responses. Share of Voice adds the competitive environment inside those responses.

That gives each metric a different job:

• Visibility Score: Is the brand appearing in more of the AI conversations you track?
• Share of Voice: Is it gaining presence relative to other companies appearing in those answers?
• Rank: How does the result compare with the rest of the tracked set?

The limitation is that Visibility Score does not tell you whether the brand appeared first, fifth, positively, negatively, or as a cited source. Profound separates those questions into other metrics and screens.

Use Visibility Score as the coverage trend and Share of Voice as the competitive check. If either moves significantly, break the result down by platform and prompt before deciding what needs to change.

Average Position: the metric Profound gets wrong

Profound's Average Position looks like a ranking metric, but it measures something narrower: where your brand appears only in AI answers that mention it. That makes it useful for prominence, but unreliable as a standalone measure of whether your brand is actually winning.

Our tested brand had an Average Position of 5.4 but an Average Position Rank of #43. Two assets at the top of the leaderboard had an Average Position of 1, followed by results at 1.5 and 2.

The calculation itself is not the main problem. The problem is how Profound presents Average Position Rank as a competitive leaderboard without incorporating how frequently each brand appears.

Average Position measures where a brand sits when it is mentioned. It does not combine that placement with Visibility Score.

A brand with limited visibility can therefore rank very highly on Average Position if its appearances tend to occur near the top. Another brand can appear much more consistently but rank lower because it usually appears further down an answer.

That creates two separate optimization problems:

• Visibility: getting the brand into relevant AI answers.
• Prominence: moving the brand higher when it appears.

Average Position measures the second.

This is why #43 should not be read as the 43rd strongest brand overall. It means 42 tracked assets had a better average placement inside responses where they appeared.

Use Average Position at the prompt level to identify queries where your brand already appears but tends to sit behind other recommendations. Do not use Average Position Rank as a headline competitive KPI without reading it alongside Visibility Score.

The Platform Matrix: where AI engines disagree about you

Profound's Matrix View shows whether an apparently healthy overall visibility score is being carried by one answer engine while another barely mentions you. That split matters because platform-level gaps tell you where to investigate instead of optimizing against one blended number alone.

Our five-day Matrix View exposed a much larger platform gap than the overall Visibility Score suggested.

The tested brand recorded:

• 82% Visibility Score on the strongest engine
• 22.9% on the second engine
• 19.1% on the weakest engine

That is a 62.9 percentage-point difference between the strongest and weakest platform.

The first brand in the matrix was far more consistent, recording 96%, 83.3%, and 80.9% across the three engines. Another competitor showed the opposite pattern to our tested brand, with only 26% visibility on one engine but 72.3% on another.

That is what makes the Matrix View useful. An overall Visibility Score can tell you there is a problem, but it cannot tell you whether the weakness exists everywhere or is concentrated on one platform.

Those situations require different investigation. Weak visibility across every engine suggests a broader problem. Strong performance on one engine and weak performance on another gives you a much narrower place to look.

The Matrix View is diagnostic, not causal. It shows where the divergence occurs, but not which prompts, citations, or sources produced it.

Use the matrix to find engines where your brand materially underperforms its own results elsewhere. Then move into Prompt Analysis and citation data instead of treating AI search as one channel.

Prompt Analysis: where you lose, and what Profound won't tell you

Profound's Prompt Analysis is where an aggregate visibility score becomes useful for optimization. It breaks performance down to individual tracked questions, showing which prompts consistently surface your brand, where competitive rank drops, and which queries deserve investigation before your team changes content.

Our Prompt Analysis contained 10 prompts, one topic, and 150 executions across the five-day view. Results varied substantially even though every prompt belonged to the same topic.

One prompt produced 78.3% Visibility Score and rank #5, while another returned 33.3% visibility and rank #10.

That is a 45 percentage-point visibility gap hidden inside the same topic-level dataset.

Visibility and competitive rank also did not move together. One prompt generated 75% visibility, 10.3% Share of Voice, and rank #2, while another produced 66.7% visibility, 8.7% Share of Voice, and rank #6.

Prompt Analysis therefore gives a content team a practical way to separate:

• prompts where the brand already performs well;
• prompts where it appears but sits behind competitors;
• prompts with weak visibility;
• gains and losses hidden by the overall average.

The limitation is in the word analysis. The table identifies where performance changes, but it does not establish why.

A weak prompt does not prove that a missing page, citation gap, competitor action, or specific content issue caused the result. You still need to inspect the underlying responses and sources.

Prompt selection matters too. A precise dashboard built around low-value questions is still a low-value dataset.

Use Prompt Analysis for prioritization, not diagnosis. Find the important prompts where visibility or rank is weakest, then inspect what sits behind those answers before changing content.

Query Fanouts: the closest thing to keyword data in AI search

Profound's Query Fanouts show the search queries answer engines generate behind a tracked prompt, making them the closest thing to keyword discovery for AI search. The feature helps teams see how models reinterpret intent before deciding what content to create or update.

A prompt tells you what the user asked. Query Fanouts show additional searches generated while an answer engine researches that request.

In our five-day test, Query Count ranged from 11 to 20 across the 10 tracked prompts. Two prompts produced 20 queries, while the lowest result generated 11.

Average Queries Per Execution ranged from 1.3 to 2.1 at prompt level. At platform level, the most active engine averaged 2.3 queries per execution, while another averaged 1.0.

That makes Query Fanouts useful for three tasks:

• Finding content gaps: recurring fanouts can surface relevant subtopics missing from an existing page.
• Expanding coverage: the generated searches reveal angles beyond the wording of the original prompt.
• Comparing engine behavior: different platforms can research the same question differently.

The limitation is explicit inside Profound. Query fanout data is not guaranteed. The interface warns that coverage varies and some executions may not return fanout data.

That means the queries should not be treated as a complete log of everything an answer engine searched.

Use Query Fanouts as a research layer, not as a replacement for traditional keyword data. Look for recurring queries around commercially important prompts and use them to identify missing content angles worth investigating.

Citations: the most useful screen in the product

Profound's Citations view is the most actionable screen in the product because it connects AI answers to the exact domains and pages shaping them. Instead of stopping at visibility, it shows where influence comes from and which external pages are worth investigating first.

Our tested brand had 0% Citation Share and no Citation Rank during the five-day period, while the project still contained roughly 1.1K citations across the tracked responses.

The domain leaderboard showed how concentrated that influence was. The first domain accounted for 8.94% Citation Share, followed by domains at 4.79%, 4.75%, 3.57%, and 3.48%. Our capture contained 217 cited domains.

The page-level view was more actionable.

The most-cited individual page had 4.79% Citation Share, and Profound marked our tested brand as Not Mentioned on it. The next results included one page at 2.58% where the brand was mentioned and another at 2.43% where it was not. The table contained 377 cited pages.

The Mentioned on Page column separates two situations:

• a frequently cited page already contains your brand;
• a frequently cited page does not contain your brand.

The second group gives a GEO team specific pages to investigate for publisher outreach, digital PR, updated product information, or content work.

That is much more useful than simply knowing Citation Share is low. Citation Share tells you there is a problem. Top Citation Pages shows where an opportunity may exist.

There is an important limitation. A page being frequently cited does not prove that adding your brand to it will cause AI answers to mention you. Citation data identifies influence and opportunity, not causation.

Start with Top Citation Pages, identify influential sources where your brand is absent, and prioritize the opportunities you can realistically affect.

Agent Analytics: check your CDN before you buy

Profound's Agent Analytics measures AI crawler and referral activity from server or CDN logs, but the feature only becomes useful once your infrastructure is connected. Before buying for this capability, check the setup screen because support and feature depth vary sharply by provider.

Our setup screen showed 12 integration options plus a custom connection, but support was not equal across them.

Two integrations were restricted to customers using an Enterprise infrastructure plan. Two others were marked Limited features. Another enterprise CMS integration was marked Beta and restricted to its cloud-hosted version.

The same screen placed four popular website-building and hosting platforms under "Not yet supported."

That is more useful than a generic integration logo list. Profound can include unlimited Agent Analytics domains on Growth while your actual site still requires an infrastructure upgrade, has limited functionality, or cannot yet connect.

Our capture also showed one enterprise-only CDN integration as Detected for the test domain, demonstrating that Profound checks the current infrastructure during setup.

We inspected the setup workflow but did not complete the connection, so we did not collect first-party crawler or referral measurements. We therefore do not use unmeasured traffic data to judge Agent Analytics.

Check your actual CDN, hosting platform, and plan before treating Agent Analytics as part of the $399 value. Infrastructure requirements can materially change both implementation effort and total cost.

Profound Agents: credits are the real pricing lever

Profound's Agents can turn visibility data into automated research, content, and optimization workflows, but the $399 Growth plan does not include unlimited automation. Its 400 monthly Agent credits are the practical usage ceiling, so workflow complexity matters as much as the feature list.

The Agents interface supports natural-language workflow creation, manual building, and templates. Our account showed templates for FAQ generation, article translation, product-page optimization, and video metadata optimization.

The pricing constraint matters more than the templates.

• Starter: 100 Agent credits/month
• Growth: 400 Agent credits/month
• Agency Growth: 400 Agent credits/month per client workspace

A credit is not the same as one Agent run. Consumption depends on workflow complexity. A simple workflow with fewer steps uses fewer credits than a workflow that pulls data, analyzes it, creates content, and passes the output into additional actions.

This means 400 credits cannot be translated into a fixed number of monthly workflows.

Profound shows estimated credit consumption before execution and actual consumption after a run. That lets teams assess whether recurring workflows fit inside the included allowance.

The 100-prompt limit controls how much monitoring Growth can run. The 400-credit limit controls how much automated execution you can perform with the resulting data.

If Agents are occasional, that allowance may be sufficient. If automation becomes part of a recurring content or GEO workflow, credits can become the real boundary between Growth and a larger plan.

Before relying on Agents, build the workflows you expect to use and check their estimated credit cost.

Final Verdict

Profound earns its $399 Growth price through the diagnostic layers underneath its headline visibility metrics. Platform Matrix shows where engines diverge, Prompt Analysis identifies the questions driving those differences, and Query Fanouts expose additional searches generated behind them.

Citations is the strongest part of the product. Domain and page-level citation data gives GEO teams specific external sources to investigate instead of stopping at a visibility percentage.

Average Position needs more caution. It measures prominence when a brand appears, but its leaderboard does not incorporate appearance frequency. It should not be treated as an overall competitive ranking.

Agent Analytics can add crawler-level evidence, but its value depends on your infrastructure. The setup screen shows meaningful differences between fully supported, restricted, limited, beta, and unsupported integrations.

Agents extend Profound from analysis into execution, but the included 400 monthly credits create a separate usage ceiling.

The $399 Growth plan makes the most sense for teams that will actively investigate platform differences, prompt-level performance, query fanouts, and citation sources. If you mainly need a simple AI visibility score, much of Profound's depth will go unused.