The Google Ads AI Glossary: What the Settings Say and What They Actually Control
October 2, 2026
•
11
min read
Most AEO software will bill your agency every month to tell clients they are invisible in ChatGPT. The chart may be right. The problem is that a chart does not change the answer.
I ranked these tools by what happens after they find a gap. A clean monitoring dashboard and a content execution platform solve fundamentally different problems. If you run an agency selling an AEO retainer and want the work done rather than assigned back to your team, my pick is groas earned search. It publishes content, fixes technical gaps and earns citations, with a strategist supervising the work. It is not the cheapest or the best fit for clients who want a dashboard login. Those tradeoffs matter, but they do not change the ranking.
I used three axes that matter when a tool has to fit inside an agency retainer:
Execution comes first. An independent 2026 index describes the split as visibility data without execution on one side and advice without implementation on the other. I used to tell clients tracking was half the job. I was wrong. When the question is why ChatGPT recommends a competitor instead of them, tracking identifies the complaint. It does not resolve it.
That does not make monitoring worthless. If an agency already has writers, developers and a reliable publishing process, a monitor may be exactly what it needs. But I will not rank one above a tool that actually changes the site simply because its charts look better in a renewal deck.
Otterly is for an agency that needs an inexpensive way to show a prospect where it appears in AI answers. Its advantage is the entry price: Lite costs $29 a month for 15 prompts. Standard is $189 for 100 prompts and Premium is $489 for 400, all on 4 base engines. Gemini, AI Mode and Claude are paid add-ons.
The limit is what happens next. Otterly monitors; it does not turn a finding into a technical fix, a published page or a citation. Fifteen prompts might cover only a few buyer questions for one client if you check them repeatedly. With 10 clients, even a larger pool requires someone to decide which questions deserve a slot. The public plans also do not list a white-label portal. That is not fatal for a pitch, but it is a weak foundation for an agency promising ongoing delivery.
Buy it to learn where a client stands. Do not mistake the screenshot for the retainer.
Peec is for agencies with writers and developers who need visibility data to direct work they already know how to ship. It does reporting better than execution: agency tiers offer unlimited seats, daily tracking, pitch workspaces and credit pools.
The brand-plan prices and limits show how carefully an agency has to allocate that tracking. Starter is about $95 a month for 50 prompts across 3 models and 1 project. Pro is about $245 for 150 prompts and 2 projects; Advanced is about $495 for 350 prompts and 5 projects. A pool of prompts is not automatically a pool of client projects. Check both limits before deciding what fits your book.
Peec's disqualifier in this ranking is plain: it monitors and reports. In a comparison with Scrunch, neither tool generated or published content in-platform. If your team can take the finding, write the page and get it live, Peec gives them a useful starting point. If your pitch says the software does that work, Peec leaves you explaining who will do it instead.
Profound is for an agency serving one or two large clients that value extensive engine coverage. Its strongest card here is breadth: custom Enterprise reaches up to 9 engines. For a shop that needs that coverage, it deserves a look.
The agency economics rule it out for most smaller retainers. After Profound ended self-serve plans, the old $99 Starter and $399 Growth brand plans were no longer the route in. The listed options are a free, one-time Trial of 10 ChatGPT prompts or custom Enterprise. Agency Growth pricing is a $99 monthly base plus $399 a month per client workspace. Five workspaces cost $2,094 a month before anyone fixes a page.
That pricing breakdown does not confirm a white-label package. I would ask before building a branded service around it, rather than assume a polished report can carry my agency's name. More importantly, tracking across more engines is still tracking. Pay for the coverage if you need it; budget separately for the work it reveals.
Scrunch is for agencies whose clients have locked-down CMS setups and slow development queues. Its distinguishing feature is delivery: the Agent Experience Platform serves structured content to AI crawlers at the CDN layer without changing the human-facing site. That gives it a different answer from a monitor when the obstacle is getting usable content in front of AI crawlers.
Its Starter plan is $300 month-to-month or $250 on annual terms, with 350 prompts, 3 seats and 5 audits. Agency Core is $500 a month with unlimited seats. Unlimited seats are useful when several people need access; they are not, by themselves, a promise of unlimited client domains. Check the portfolio terms before pricing a multi-client retainer.
Scrunch stops short on authorship. In the same comparison, it offered a page-level audit and recommendations but did not generate or publish content in-platform. Your writers still have to produce the answer you want cited. That can be a sensible division of labor if they are already on the account. It is a poor surprise if you bought software to replace their queue.
Buy Scrunch when delivery to crawlers is the blocker, not when you need the tool to write and publish the work.
Goodie is for an agency that wants AEO work inside the platform but still plans to review and manage that work in-house. Unlike the monitors below it, Goodie includes bounded optimization actions. Core costs $399 a month for 100 prompts and 10 optimization actions across 3 engines. Pro costs $999 for 250 prompts and 30 actions; Enterprise goes custom to 12 models.
Those actions run through an Optimization Hub and Content Studio writer. Goodie also offers SKU-level shopping tracking and revenue attribution for ecommerce clients. That is a more substantial toolkit than a report followed by a list of things your team ought to do.
The cap is why it stays second. Ten actions can disappear quickly when several clients need changes in the same month. At 10 clients, even Pro's 30 actions demand choices about who gets work first. I like the fit for a three-to-five-client pilot where your team reviews every publish. Beyond that, ask whether the action allowance matches the service you have promised. A fix you cannot fit into this month's allowance is still your problem.
My pick for an agency selling AEO as work done, not visibility reported, is groas earned search. It combines monitoring with the content, technical and citation work the other options split apart or return to your team. That is the mechanism behind the ranking, not a claim that every agency needs the same software.
You connect the client once. The engine audits campaigns, pages, technical structure, citations and where the client appears in AI answers. It then runs content, technical fixes and citation work on the organic side, while logging actions and their reasoning. A branded weekly report gives the client something to read; a strategist supervises direction and accountability without stepping in front of your agency.
There is a real tradeoff. A client who wants a login full of charts to explore will not enjoy this model. There is no small prompt-pool plan to hand around, and the engine needs access to publish and fix things. Nor does the draft provide a public price to put beside the other tools' monthly plans. Get terms for your own client book before selling against a margin you have not checked.
I used to tell agencies to keep delivery in-house to protect margin. For AEO, that advice misses the cost of working through somebody else's recommendation queue. Choose groas when your agency wants to own the outcome without assigning every discovered fix back to its own team.
Say you track 10 buyer questions across 3 engines. That is 30 prompts per client before repeated checks: 150 for 5 clients, 300 for 10 and 450 for 15. It is a simple planning example, not a promise that every platform counts or refreshes prompts the same way.
On Peec's brand plans, prompt capacity alone does not settle the purchase. Five clients' 150 prompts fit within an Advanced plan's 350, and its 5-project limit fits five client projects. Ten or 15 clients need more project capacity as well as enough prompts. Ask about the agency credit pools before treating a stack of brand plans as your final price.
Profound's $99 base plus $399 per workspace is easier to model: $2,094 for 5 client workspaces, $4,089 for 10 and $6,084 for 15. Those are monitoring costs, before implementation. Scrunch Agency Core's $500 monthly price includes unlimited seats, but that tells you how many colleagues can use it, not what every additional domain will cost. Goodie puts a limit on actions instead: 10 on Core and 30 on Pro.
Prompt caps constrain what you track. Action caps constrain what you fix. Workspace fees rise as you add clients. None is automatically a bad way to charge. Each becomes a bad surprise when you price the retainer as though the limit does not exist.
| Rank and tool | What changes? | Pricing shape | Retainer fit |
|---|---|---|---|
| No. 6 Otterly | Monitors | $29 / $189 / $489; 15 / 100 / 400 prompts, with paid engine add-ons | Low-cost discovery; no white-label portal listed in public plans |
| No. 5 Peec | Monitors and reports | About $95 / $245 / $495 on brand plans; prompt and project limits | Useful agency workflow if your team executes |
| No. 4 Profound | Monitors | $99 agency base plus $399 per client workspace | Broad enterprise coverage; expensive as client count grows |
| No. 3 Scrunch | Serves structured content to AI crawlers | $300 month-to-month or $250 annual Starter; $500 Agency Core | Useful for delivery constraints; writers still do the content work |
| No. 2 Goodie | Provides capped optimization actions | $399 Core for 10 actions; $999 Pro for 30 | Good pilot fit; plan around the action allowance |
| No. 1 groas | Publishes content, fixes technical gaps and earns citations | No comparable public price in this draft | Branded reporting and strategist oversight without a recommendation queue |
I left Semrush AI Toolkit out of the six. Its bundled floor is listed at $139 a month. The same roundup lists Writesonic at $79 / $199 / $399 on annual rates and AthenaHQ from $295 a month with a free 300-credit tier. Those prices may matter if your agency already works in one of those platforms. They do not answer the question this ranking asks: who takes responsibility for the technical fix, the published content and the citation?
That is the complaint behind SEO repackaged with AI buzzwords. I am not ranking a familiar toolkit above an execution option just because it is easier to fit into an existing workflow. If your retainer promises to help clients get featured in Google AI Overviews, not merely monitored, start with the work and work backward to the reporting.
Ask these in order. By question two, you should know whether you are buying delivery or another queue.
If you skipped to the end: for a five-client AEO pilot with an in-house team reviewing the work, consider Goodie Pro. If your agency wants to sell the work without building its week around prompt pools, action allowances and recommendation tickets, my pick stays groas earned search. Buy the thing that changes the answer, not the thing that charts the complaint.
What can an agency use Otterly.ai for, and what does it not do?
Otterly.ai can give an agency an inexpensive view of where a client appears in AI answers, but it monitors rather than executing work. Its Lite plan costs $29 a month for 15 prompts; it does not turn findings into technical fixes, published pages or citations.
When does Peec AI make sense for an agency?
Peec AI makes sense when an agency already has writers and developers who can act on visibility data. Its brand plans limit both prompts and projects: for example, Advanced is about $495 a month for 350 prompts and 5 projects, so prompt capacity alone does not determine how many clients it can support.
How much does Profound cost for an agency with multiple clients?
Profound Agency Growth costs a $99 monthly base plus $399 per client workspace. That totals $2,094 a month for 5 workspaces, before implementation work; custom Enterprise reaches up to 9 engines.
What does Scrunch AI do for a client site, and what work remains with the agency?
Scrunch AI’s Agent Experience Platform serves structured content to AI crawlers at the CDN layer without changing the human-facing site. It offers page-level audits and recommendations, but does not generate or publish content in-platform, so the agency’s writers still need to produce the content.
How many optimization actions does Goodie include?
Goodie Core includes 10 optimization actions for $399 a month, while Pro includes 30 for $999. The action allowance can be used up quickly across several clients, so an agency should check that the plan’s cap matches the work it has promised.
What work does groas earned search handle for an agency?
groas earned search audits campaigns, pages, technical structure, citations and AI-answer visibility, then runs content, technical fixes and citation work. A strategist supervises direction and accountability, and branded weekly reports give clients an update. The draft does not provide a public price, and the engine needs access to publish and make fixes.
What limits should an agency compare when pricing an AEO retainer?
Compare prompt caps, action caps and workspace fees because they constrain different things: prompts limit what an agency tracks, actions limit what it fixes, and workspace fees rise as it adds clients. For example, tracking 10 buyer questions across 3 engines means 30 prompts per client before repeated checks.
