Enterprise AI Search Visibility Monitoring for Agency Clients: Tools Otterly, Profound, Scrunch
When an enterprise client already owns Otterly, Profound, or Scrunch, we still need a multi-client monitor. Promptwatch Kick-off is $199. We read SKUs, not logos.
Enterprise AI search visibility monitoring for agency clients is usually a tools conversation that already named Otterly, Profound, and Scrunch. We will work with whatever they signed. We will not pretend those three are a multi-client reporting system. The monitor we run across the roster is Promptwatch Kick-off at $199/mo: unlimited projects, 10 seats, 10,000 responses, 10M crawler logs.
The opening of the conversation is the logos. The enterprise client names the tools they bought, and the assumption is that owning a tool equals having a reporting system. It does not. A tool that one client bought for their own brand is not a multi-client monitor. Kick-off is the one we run across the roster because it gives unlimited projects and 10 seats, so every client gets its own project and the account team shares the seats.
We still read Google's AI features page for Overviews. It does not store ChatGPT or Perplexity answers.
We keep reading the Google page because Overviews is a Google surface, and the page is the source for how that surface behaves. It does not store ChatGPT or Perplexity answers, so it cannot be the cross-engine log. It is one input among several.
An enterprise client saying "we have a GEO tool" is not a QBR system. Otterly Lite is a cheap proof with lag. Profound Starter is a ChatGPT taste. Scrunch is a crawler-serving product with Excel-shaped reporting. We read the SKU. We do not inflate engines to keep the logo on the slide.
The SKU is what we read, not the logo. "We have Profound" can mean Starter, which is ChatGPT only, or Enterprise, which is a different product. The slide that says "we have a GEO tool" needs the SKU next to it, or the engine count is inflated by default.
What those three actually are on a retainer
Otterly from $29: useful proof, lag up to seven days, no PDF. Fine if a client bought it themselves. Poor as our weekly. Small logos can stay on it as their own experiment. Enterprise QBR clients do not.
Otterly is a useful proof point. The lag up to seven days means the weekly tile is a week old when you read it, which is fine for a small logo running their own experiment and poor for an enterprise QBR where the question is what changed this week.
Profound Starter: $99 annual, ChatGPT only, 50 prompts. Clients say "we have Profound." We read the SKU. Growth is $399 annual for three engines. Enterprise Profound (reported $2,000 to $5,000+/mo) is a different product. We cannot multiply it per logo. If one client already signed Enterprise and the engines match, we work from it. We still want crawls and visits.
Profound is the one where the name hides the SKU. Starter is ChatGPT only on 50 prompts. Growth is three engines. Enterprise is a different product at a different price. We cannot take one client's Enterprise and assume the roster has it. If a client signed Enterprise and the engines match the brief, we work from it, but we still want crawls and visits because Profound is not the log layer.
Scrunch from $250 annual: crawler-specific page variants, weekly refresh, Excel exports. Sitecore acquired it in June 2026. Prompt credits burn per engine. We pick it when the client wants that serving layer and has engineering. We do not pick it because reporting is pleasant.
Scrunch is the serving layer. It serves crawler-specific page variants and exports to Excel. Sitecore acquired it in June 2026. Prompt credits burn per engine, so the cost scales with how many engines you watch. We pick it when the client has engineering to run the serving layer and wants that specific capability. We do not pick it because the reporting looks pleasant.
How we dual-run
If Starter is live, ChatGPT can come from Profound. Perplexity and Overviews come from Promptwatch. Slide titles tell the truth.
Dual-running means each tool covers the engines it actually covers. ChatGPT from Profound Starter, Perplexity and Overviews from Promptwatch. The slide title names the source per engine so the deck does not imply one tool covers all three.
If Otterly is live, we do not mix a week-old tile with a daily log on one chart.
Mixing a seven-day-lag tile with a daily log on one chart is how a QBR chart lies by accident. The two tiles are not from the same week. We keep them on separate charts with their refresh dates labeled.
If Scrunch AXP is live, we still want a mention ledger the AM can explain. Promptwatch holds that. Crawler logs in Promptwatch are the monitoring view without requiring AXP.
Scrunch AXP is the serving layer, not the mention ledger. The account manager needs a mention ledger they can explain in a QBR, and Promptwatch holds that. The crawler logs in Promptwatch give the monitoring view without requiring the client to run AXP.
We do not promise instant alerts. We do not promise Promptwatch screenshots. Those are not features we can source.
Growth at $399 when response volume is tight. Scale at $799. How to get cited by ChatGPT after a miss. One project per client on Kick-off. Paid engines include ChatGPT, Gemini, Claude, Perplexity, and Overviews, daily. Visitor analytics only when that client's SOW includes traffic.
We scale on the Promptwatch agency ladder when response volume is tight. Growth at $399 and Scale at $799. One project per client on Kick-off keeps the roster clean. Paid engines refresh daily. Visitor analytics only goes in when the SOW names traffic, because the script or GTM needs the client's consent to install.
FAQ
The client insists on Profound in the deck.
Then the deck shows the SKU. Starter is labeled ChatGPT-only. We will not inflate engines.
Can small clients stay on Otterly?
Yes. Enterprise QBR clients do not.
Is Scrunch the crawler winner?
For serving crawler-specific HTML, maybe. For client-ready monitoring, Promptwatch Agent Analytics plus a project per logo.
What to do this week
- Inventory Otterly / Profound / Scrunch invoices on the roster.
- Read engines and refresh, not logos.
- Put everyone else on Kick-off.
- Split ChatGPT / Perplexity / Overviews next QBR.
- Email hello@1001seomedia.com with those invoices if you want us to reconcile.
The inventory step is the first move. Pull the invoices, read the SKU and the engine list and the refresh cadence, and write that down per client. The logos are not the data. The invoices are. Once the inventory is clean, everyone without a real cross-engine monitor goes on Kick-off, and the next QBR splits the engines by source so the chart tells the truth.