Our read
Four jobs are documented: client measurement, prompt discovery, crawler-readiness fixes and new-business audits. Detailed agency examples exist for Profound (Jordan Digital Marketing) and Scrunch (Proper Propaganda, AlchemyLeads, Clapping Dog Media). Peec AI and Evertune appear in this sector as OBSERVED agency lines without a described task, so they show adoption only. The sources do not separate these tools by quality, so test your own client prompts rather than pick by logo count.
Which tool for which job?
| Your job | Documented example | The reference to read | What it documents |
|---|---|---|---|
| Pilot on one client, then roll out | Jordan Digital Marketing ran Profound on a single-client pilot, then deployed it across its client portfolio to see how clients appear across AI platforms. 1 | ||
| Find the questions users ask AI | Jordan Digital Marketing used Prompt Volumes to find what expecting parents ask chatbots for a nursery furniture client, then wrote short-form content for those prompts. 1 | ||
| Restructure pages for AI crawlers | Jordan Digital Marketing used AI crawler analytics to restructure a cybersecurity client's pages into machine-readable formats. 1 | ||
| Run the agency's own GEO program | Proper Propaganda tracked about 700 prompts, restructured its pages and used citation analytics to choose third-party placements, for itself and for clients. 2 | ||
| Benchmark clients and prepare pitches | AlchemyLeads benchmarks clients against competitors and audits a prospect's AI visibility with Scrunch before the first discovery call. 3 | ||
| Audit client sites for AI readability | Clapping Dog Media's founder used Scrunch's Site Audit on client sites, found a healthcare client's insurance icons unreadable to AI and replaced them with text. 4 |
Which GEO jobs do agencies actually document?
Four jobs show up in named agency sources: measuring how a client appears in AI answers, finding the prompts people ask, fixing pages so AI crawlers can read them, and using visibility data to win new business. Every detailed example comes from two tools, Profound and Scrunch. For Peec AI and Evertune, the public record in this sector shows agency names but no described task.
As of 2026-11-07, usedby lists 20 Agency & Studio companies with a tool in this category: 4 at CONFIRMED and 16 at OBSERVED. CONFIRMED means a case study, quote or press release; OBSERVED means the vendor shows the company on its own pages. That is a map of public lines, not market share, and it says nothing about quality.
How does a small agency roll out Profound across clients?
Jordan Digital Marketing began with a single-client pilot to check that Profound fitted its workflow, then deployed it across its portfolio once the first AI visibility report came back. The agency uses it to see which blog posts, product pages and buying guides earn AI citations for each client, and to double down on those formats.
For a nursery furniture client, the team used Profound's Prompt Volumes to find which questions expecting parents ask chatbots, then wrote short-form content around them. For a cybersecurity client, it used AI crawler analytics to restructure pages into machine-readable formats. Profound describes its product as measuring AI citations, showing what people ask answer engines, and tracking AI bot traffic.
Jordan reports, in Profound's case study, a 34% revenue increase, roughly double the profit, and a client's visibility in one prompt rising from 0% to 80%. These are the agency's own figures, tied to its own business and one client; they are not a benchmark and do not isolate the tool's effect. Its stated plan is to add Profound Agents next.
Gr0 is also listed by Profound at OBSERVED level. Profound's customer story for it describes campaign managers tracking prompts like keywords, turning the most-cited URLs into an outreach list, drafting content briefs with Profound Agents and watching brand sentiment. Treat that as a vendor-published description of one agency, with a reported rise from about $1,000 to over $100,000 monthly LLM-attributed revenue for one client.
How do agencies use Scrunch to audit, plan and pitch?
Proper Propaganda, a tech PR and GEO agency, ran its own program first. It chose nearly 700 prompts across ChatGPT, Perplexity, Gemini and Google AI Overviews, benchmarked itself against competitors, rebuilt its pages, and used Agent Traffic, citation analytics and Site Maps to decide what to fix. It also used citation data to target third-party placements. Scrunch's case study reports it as the most visible agency across 695 tracked prompts as of May 2026 and a 5x rise in leads; these are vendor-published claims.
AlchemyLeads, in a guest post by its founder on Scrunch's site, benchmarks each client against competitors, reviews topics and subtopics, reads prompt reports for questions keyword research missed, and uses the citations report to find sources models trust. It also runs a Scrunch audit on prospects before the first call. The author attributes three enterprise contracts in Q4 2025 partly to this; that is a self-reported claim.
Clapping Dog Media, a Scrunch Agency Partner, sells AI plus SEO audits as a service. Its founder used Scrunch's Site Audit tool on client websites; in the vendor-published spotlight, a healthcare provider listed accepted insurance plans as icons that AI could not read, and replacing them with text improved that client's visibility. That is the audit job in its simplest form.
What do the Peec AI and Evertune lines tell you?
Peec AI is shown by the vendor with Seer Interactive, Omniscient, Mindshare, Eskimoz, Peak Ace and Amsive. Those lines show adoption only. Peec's own site describes tracking visibility, position and sentiment, source citations, strategy recommendations, a Looker Studio connector, CSV export, an API and MCP access. A quote from Amsive's VP of SEO Strategy on that site concerns tracking brand mentions and sentiment, but no agency workflow is described.
Evertune's only Agency & Studio line is WPP, at OBSERVED level. Evertune describes a consumer prompt panel, statistical sampling of prompts, an Insights Agent and ads in AI conversations, but nothing in the record shows what WPP does with it. Do not read that logo as a workflow.
Which named US and UK agencies are linked to which tool?
Jordan Digital Marketing (US, Profound) is the one CONFIRMED peer with a described workflow. Gr0 (US, Profound) is OBSERVED, with a vendor customer story. Stratabeat and Big Leap (US, Scrunch), Seer Interactive, Omniscient and Mindshare (US, Peec AI) and WPP (GB, Evertune) are OBSERVED: the vendor shows them, and their lines name no task.
For a peer reader, the lesson is narrow: only a few named agencies publish how they work. WPP and Mindshare appear only as vendor-listed names, with no task described. That does not show a preferred tool in this market.
What should you check before copying one of these workflows?
Start where Jordan started: one client, one fixed prompt set, a first report, then a decision. That is an editorial suggestion drawn from its pilot, not a vendor feature. Check whether the tool covers the AI platforms your clients care about, whether it separates client workspaces, and what reports you can hand to clients.
Profound and Scrunch each list a free trial; Profound lists SOC 2 Type II. Prices for any of the four are not established in this record, so confirm them yourself. The reported gains came from specific clients, prompt sets and content work, and may not transfer to yours.
Method and limits
2026-11-07
Scope: 20 companies classified in Agency & Studio on usedby with a public line on a AI search visibility (GEO) tool. Every count on this page covers this sector only, not the tools' whole customer base. Company size is not recorded and is not claimed.
This dated snapshot counts distinct companies in public customer evidence. CONFIRMED means a case study, named testimonial or press source; DECLARED means the company itself told usedby; OBSERVED means a vendor listing; SIGNAL means a job post. All four count, but their strength differs. Removed and unlevelled lines are excluded. Examples are a capped selection verified within 90 days. Vendor facts are sourced records checked within 60 days. Coverage is uneven and missing evidence is unknown. These figures measure neither market share nor product performance; capture dates are not adoption dates.