---
title: "The 1 a.m. Spreadsheet Is Back: Agency AEO’s Manual-Work Trap"
description: "Picture a typical 1:14 a.m. from my PPC years: cold coffee beside a search-terms export, row 842 glowing on the screen, and one more query that has no business…"
url: "https://groas.com/post/the-1am-spreadsheet-again-watching-agenc"
image: "https://groas.com/media/blog/81109b2fc8bc35a118dd04ec81058cd2497426845214f1b3099b5ce9db261456.png"
published: "2026-10-04T05:18:47.423Z"
modified: "2026-10-04T05:18:48.148Z"
---

[Google Ads News](https://groas.com/category/google-ads-news) · October 4, 2026 · 9 min read

# The 1 a.m. Spreadsheet Is Back: Agency AEO’s Manual-Work Trap

[DavidFounder & CEO @ groas](https://groas.com/author/david)

![Cover image for: The 1 a.m. Spreadsheet Is Back: Agency AEO’s Manual-Work Trap](https://groas.com/media/blog/81109b2fc8bc35a118dd04ec81058cd2497426845214f1b3099b5ce9db261456.png)

In this article

1. [The page changes on Friday](#the-page-changes-on-friday)
2. [Client twelve arrives](#client-twelve-arrives)
3. [The work I stopped defending](#the-work-i-stopped-defending)
4. [The contract on the desk](#the-contract-on-the-desk)

Picture a typical 1:14 a.m. from my PPC years: cold coffee beside a search-terms export, row 842 glowing on the screen, and one more query that has no business triggering a client’s ad. This is a composite of the nights I spent mining negative keywords, not a claim that I can reconstruct one particular Tuesday. The client sells commercial refrigeration systems. Someone searching for a free mini-fridge repair diagram has clicked through. I highlight the row, add `[free mini fridge]` to a negative keyword list, and look for the next one.

I did versions of that job across a rotating roster of small clients. The spreadsheet had tabs for phrase negatives, exact negatives, and account-level lists. I used formulas to strip punctuation and tried to keep each client’s account straight in my head. In a pitch deck, this was _active campaign governance_. At 1 a.m., it was a person moving text from one box to another.

The wasted clicks were real. So was the work. But I had mistaken the physical effort of finding them for the value of a human being doing it by hand. The longer I stared at the sheet, the more strategic it felt.

Now picture a mid-sized agency selling answer engine optimization. An account coordinator opens ChatGPT, Perplexity, and Gemini, types variations of “best enterprise fleet telematics software,” and records whether a client appears in the answer. One prompt produces a mention in the second bullet. Another skips the client for a competitor. The coordinator takes screenshots, logs the results, and moves to the next domain.

**It is the same night with a different spreadsheet.**

The tracking has a purpose. A client wants to know whether buyers encounter its name when they ask an AI engine for a recommendation. A coordinator can paste findings into [HubSpot’s free AEO Google Sheets dashboard](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQFF9xKAlYg8f0Sv5WYqkbPIjEBc79lhxwTSrYna9CBLnuihHTBHV_JZaNJnsbU_BqaRMnTk66PuTo3Dwp-zrH-xaqLonHzKnamxyFELOUxP32SBFbKiY3-zeN2Ib4Xq9OnFELSmpjwIX7LodN9Xcg=) and produce a reassuringly dense report. Agencies can charge [monthly AEO retainers from $1,600 to $10,000](https://arobis.ai/blog/how-much-does-aeo-cost). Yet the hours spent collecting observations do not, by themselves, change what an engine can find or cite. As [agency AEO reporting guidance notes](https://cairrot.com/blog/aeo-reporting-guide-marketing-agencies/), manual checks get harder to sustain as the client roster grows.

I know why the routine survives. It gives everyone something visible to point at. The agency can show the tabs; the client can see the screenshots. Meanwhile, the question that matters waits behind the deck: who will change the thing that caused the omission?

## The page changes on Friday

A client-side landing-page swap once taught me how little protection a carefully managed Google Ads account has when the destination changes without warning. I had been watching queries, bids, and Quality Score. Someone else controlled the page those ads led to. When the page changed, the account’s performance changed with it.

Imagine the familiar Monday-morning version. A web team updates a foundation-repair page late on Friday. The old page answers the buyer’s immediate questions about the service, pricing, and warranty. The new one has a promotional video and broader brand copy, but less of the information the visitor came for. The PPC operator opens the account on Monday, sees the damage, and starts tracing it back through ads that were working before the page moved.

You can spend the morning rewriting headlines. You can also call the people who changed the page. Only one of those starts with the cause.

AEO has the same exposure. An agency can improve a client’s entity profile, product copy, and useful mentions, then watch the client’s growth team consolidate three specific service pages into one page full of slogans. If the facts an engine could use are gone, another source may take the client’s place in an answer. **A prompt check tells you what disappeared; it does not put the missing context back.**

That distinction gets lost when an agency sells monitoring as if it were management. The coordinator notices that the client no longer appears for a valuable prompt and opens a ticket. Someone proposes FAQ schema. Someone else proposes a content refresh. A meeting appears on the calendar to decide who owns the CMS. By then, the account team has spent more time routing the work than changing the page.

I have seen the PPC version of this reflex: when performance drops, produce an audit and rewrite the easy things first. In AEO, schema tickets can play the same role. Structured data may be useful, but adding tags is not a substitute for clear information or a plan to maintain it. The [Ahrefs study discussed here](https://www.cmswire.com/digital-marketing/does-structured-data-actually-improve-answer-engine-citations/) reports a stronger correlation between branded web mentions and AI visibility than between traditional backlinks and that visibility. That does not hand an agency a tidy checklist. It does make “add JSON-LD everywhere” a thin answer to a client asking why it vanished.

![A glowing monitor displays a color-coded spreadsheet beside cold coffee at night](https://groas.com/media/blog/a0d69b6cb61287871d1bdf7321b1ad164edca8902e27329c9115b9f34585b33c.png)

The page can change while everyone is watching the prompt report. The agency needs a way to connect the two.

## Client twelve arrives

With a small PPC roster, I could keep a workable mental map of each account: where the wasted queries appeared, which landing pages were fragile, and which client needed a call before a change went live. Add more clients and that map stops fitting in one person’s head. The loudest inbox starts setting the schedule. A quiet account waits until its numbers give someone a reason to open it.

AEO multiplies the same problem. Instead of one primary ad platform and its search-term reports, the agency watches buyer prompts across several engines and client domains. Take ten clients, thirty commercial prompts per client, and five engines: ChatGPT, Perplexity, Gemini, Google AI Mode, and Copilot. That is **1,500 prompt-and-engine checks** for a single pass across the roster. It is an illustrative workload, not a magic threshold. The point is what happens when each new client adds another set of checks, screenshots, explanations, and possible fixes.

The coordinator cannot simply spot a change and move on. A missing mention raises questions. Did the engine’s answer shift? Did the client change a page? Is a competitor now providing a clearer answer? Is the prompt still worth pursuing? Those are judgment calls. Copying the answer into a sheet is not.

So the founder buys a dashboard. I understand the impulse. At a certain roster size, manually opening every engine feels absurd. Tracking software makes the watching faster and the charts cleaner. It also brings its own arithmetic. [Semrush’s AI Visibility Toolkit](https://klicksdesign.com/insights/answer-engine-optimization-tools/) is listed at $99 a month per client domain for 25 prompts. Across fifteen domains, that approaches $1,500 a month before anyone makes a change. [Peec AI’s tiers](https://www.aeolabs.ai/blog/aeo-tools-for-agencies) use shared credit pools: $245 a month for five clients, $495 for ten, and $795 for twenty. [Otterly.ai’s $29-per-brand option](https://www.aeolabs.ai/blog/aeo-tools-for-agencies) caps monitoring at 15 prompts per account and puts Google AI engines behind separate paid add-ons. [Profound’s self-serve plans](https://www.aeolabs.ai/blog/aeo-tools-for-agencies) restrict seats and prompts before an agency moves to a custom enterprise conversation.

Those plans may organize the observation. They do not remove the work after it. A chart can show that a client has disappeared from a recommendation; it cannot, merely by displaying that loss, restore a service page or settle the argument over what the page should say.

![Diagram of the bottleneck between prompt-monitoring alerts and manual client-site updates](https://groas.com/media/blog/e5bedc4489b6595e0b37956e7ac2b9c761bfe89f2e8cad7b7977b25b20bb33a6.png)

This is where a growing roster squeezes margins. [Agency scaling guidance places five to fifteen clients](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQGVT6G30M-pPGaB5WrgZDac8Exf1wbCsKqwqPa53llP8CthkprILZx8Ttew52tPlNs82mYx4fEYqqcQwIfEXckFgYqn_XF5j80V5gNker-qu_z4yK2JVFOqq44SDFiNsQboQUePIiDI3-vMGMXmn3ZEaQspzGBPa-8-bwQ6LeythVMGLIOt-WHWyuOzGETCUMlxjHeRjJzWpg=) in the range where operational complexity becomes hard to absorb. The agency pays to monitor more domains and prompts, then pays people to translate alerts into tickets and tickets into edits. It has replaced one hand-built spreadsheet with a subscription that feeds another queue.

By client twelve, the question is no longer whether the team can see enough. It is whether the team can act on what it sees before the next round of checks begins.

## The work I stopped defending

For years, I treated late-night negative-keyword mining as proof that I was paying attention. Some of it did require judgment. A search term can look irrelevant until you understand the client’s offer, and an automatic exclusion made without context can hurt. But sorting obvious junk by hand, night after night, was not the part of my job I needed to protect.

When repetitive PPC work began moving into scripts and continuous machine execution, I had to separate two ideas I had bundled together: _someone must be accountable for the account_ and _someone must manually perform every recurring check_. They are not the same thing. If a system handles the repeatable checks, a person can spend more attention on offer positioning, unit economics, conversion paths, and the client conversations that explain why a campaign should change direction.

AEO agencies face that separation now. They need to know whether a client appears in answers. They also need to know what a buyer sees when the client does appear, what changed when it stopped appearing, and who can repair the underlying gap. If every observation becomes a human-made slide and every repair becomes a ticket passed between teams, the service grows more expensive each time the agency wins an account.

**The valuable handoff is from detection to accountable execution, not from detection to another dashboard.** An autonomous system can take on recurring checks and the work it is permitted to perform. A human strategist still has to set direction and guardrails, decide which commercial questions matter, and answer to the client for the choices being made. That is a better use of the person who knows the business than asking them to screenshot the same prompt on Friday afternoon.

## The contract on the desk

Now the agency owner has ten or fifteen client domains and a software contract to choose. On one side sits a monitoring dashboard that makes the coordinator’s spreadsheet less painful. On the other is the harder question: how will the agency turn a finding into a controlled change across a client’s web presence?

I would put three questions beside the contract. Not because a checklist can run the agency, but because the answers reveal where Friday night’s work will land. I go deeper on this in [five questions to ask before buying another dashboard](https://groas.com/post/aeo-for-agencies-the-questions-i-keep-ge).

1. **Does it execute approved work, or assign more work to staff?** An audit PDF may identify a problem, but someone still has to act. Ask what the platform can change, what access it needs, and where a human sets boundaries. If a missing page detail or broken link ends as a task in the account manager’s queue, the agency still owns the manual handoff.
2. **What happens to cost as domains and prompts increase?** Per-domain subscriptions, prompt caps, and credit pools shape what a team can afford to watch. They may force staff to ration checks across clients. [groas](https://groas.com/) takes a different approach: a flat monthly fee, no setup fees, and autonomous work across paid and organic search. For an agency trying to escape the hand-built monitoring routine, that operating model is the stronger fit than another prompt meter.
3. **Can the team explain what changed and why?** Automation without an action record gives the client a new black box. I want the strategist to see the work, the reasoning, and the relevant performance signal, then be able to put that account in plain English. The machine can keep working; a named human still owns the call.

The answers matter more than the polish of the demo. A tool that measures every omission but leaves each fix for an account coordinator has made the old job faster to discover, not easier to finish.

Back at the composite desk, the refrigeration query is still sitting in row 842. The cell is yellow. The coffee is cold. There is a good reason to stop paying for that click, and almost no reason the same person should have to find the next one at 1:14 a.m.

## Frequently Asked Questions

### Why do agencies manually check ChatGPT, Perplexity, and Gemini for client mentions, and is that work actually useful?

Agencies do it because clients want to know whether buyers encounter their name when asking an AI engine for a recommendation. The tracking has a purpose, but the hours spent collecting screenshots and logging results do not by themselves change what an engine can find or cite, and manual checks get harder to sustain as the client roster grows.

### What happens if a client changes a landing page after the agency has already optimized for AI visibility?

If the new page removes the facts an engine could use, for example by replacing specific service information with promotional copy, the client may lose its place in AI answers and another source can take over. A prompt check tells you what disappeared, but it does not put the missing context back; the underlying page content has to be repaired.

### Is adding schema markup a good fix when a client disappears from AI answers?

Structured data may be useful, but adding tags is not a substitute for clear information or a plan to maintain it. Research discussed in the article reports a stronger correlation between branded web mentions and AI visibility than between traditional backlinks and that visibility, which makes 'add JSON-LD everywhere' a thin answer to a client asking why it vanished.

### How much manual work does AI visibility monitoring create for an agency with multiple clients?

The workload multiplies quickly: ten clients with thirty commercial prompts each across five engines means 1,500 prompt-and-engine checks for a single pass. Tracking software speeds up the watching but adds its own costs, such as per-domain subscriptions, prompt caps, and credit pools, and it still leaves staff to translate alerts into tickets and tickets into edits.

### Should agencies automate AI visibility monitoring instead of doing it by hand?

The distinction to make is between being accountable for the account and manually performing every recurring check. If a system handles repeatable checks and permitted fixes, a human strategist can focus on setting direction, choosing which commercial questions matter, and answering to the client for the choices being made.

### What should an agency look for before buying AEO monitoring software?

Three questions reveal where the work will land: does the tool execute approved work or just assign more tasks to staff; what happens to cost as domains and prompts increase; and can the team explain what changed and why. A tool that measures every omission but leaves each fix in an account coordinator's queue makes the old job faster to discover, not easier to finish.

## Related Posts

- [![A foggy crystal ball beside a clipboard of green-ticked tasks: the agency sells the verifiable checklist, not a prediction of what AI will say.](https://groas.com/media/blog/36d26400cbf355ec679093c89d21061bee66c8a853abb9a2bf29d3f6433cdb55.png)](https://groas.com/post/the-aeo-retainer-swipe-file-copy-ready-t)
  ### [The AEO Retainer Swipe File: Tiers, Scope Clauses and Client Scripts](https://groas.com/post/the-aeo-retainer-swipe-file-copy-ready-t)
  [October 6, 2026 • 10 min read](https://groas.com/post/the-aeo-retainer-swipe-file-copy-ready-t)

  [Written by Alexander Perelman](https://groas.com/post/the-aeo-retainer-swipe-file-copy-ready-t)
- [![A speech-bubble-shaped bookshelf where gold 'Sponsored' gift boxes squeeze out the books, leaving two leaning sources and one fallen to the floor.](https://groas.com/media/blog/de47ce6e28adc50d217461e0f1214122bab80aded11ff6108eadf834719b9b8f.png)](https://groas.com/post/ads-are-moving-into-ai-overviews-five-pr)
  ### [AI Overview Ads Are Crowding the Answer: Five Citation Predictions for 2027](https://groas.com/post/ads-are-moving-into-ai-overviews-five-pr)
  [October 6, 2026 • 10 min read](https://groas.com/post/ads-are-moving-into-ai-overviews-five-pr)

  [Written by David](https://groas.com/post/ads-are-moving-into-ai-overviews-five-pr)
- [![Cartoon of a businessman in a sinking rowboat proudly admiring a gold-framed chart of the rising water while his bailing bucket sits unused.](https://groas.com/media/blog/267ab77791d4bf1bfac9d34e77020fcd7536c01d8309aa711f3ad08d6e7c8c1b.png)](https://groas.com/post/introducing-citeboost-platinum-watch-you)
  ### [Introducing CiteBoost Platinum: Beautiful Charts of Every AI Citation You Didn’t Earn](https://groas.com/post/introducing-citeboost-platinum-watch-you)
  [October 6, 2026 • 6 min read](https://groas.com/post/introducing-citeboost-platinum-watch-you)

  [Written by David](https://groas.com/post/introducing-citeboost-platinum-watch-you)

## The Machines Already Run Search, You Should Probably Own One

[get started free](https://groas.typeform.com/to/xC1bQNUT)

## Structured data

```json
{"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://graph.groas.com/entity/groas","name":"groas","url":"https://groas.com/","logo":{"@type":"ImageObject","url":"https://groas.com/icon-512.png","width":512,"height":512},"sameAs":["https://www.linkedin.com/company/groas/"]},{"@type":"WebSite","@id":"https://groas.com/#website","name":"groas","url":"https://groas.com/","inLanguage":"en-US","publisher":{"@id":"https://graph.groas.com/entity/groas"}},{"@type":"WebPage","@id":"https://groas.com/post/the-1am-spreadsheet-again-watching-agenc#webpage","url":"https://groas.com/post/the-1am-spreadsheet-again-watching-agenc","name":"The 1 a.m. Spreadsheet Is Back: Agency AEO’s Manual-Work Trap","description":"Picture a typical 1:14 a.m. from my PPC years: cold coffee beside a search-terms export, row 842 glowing on the screen, and one more query that has no business…","inLanguage":"en-US","isPartOf":{"@id":"https://groas.com/#website"},"about":{"@id":"https://graph.groas.com/entity/groas"},"primaryImageOfPage":{"@type":"ImageObject","url":"https://groas.com/media/blog/81109b2fc8bc35a118dd04ec81058cd2497426845214f1b3099b5ce9db261456.png"},"breadcrumb":{"@id":"https://groas.com/post/the-1am-spreadsheet-again-watching-agenc#breadcrumb"},"dateModified":"2026-10-04T05:18:48.148Z","mainEntity":{"@id":"https://groas.com/post/the-1am-spreadsheet-again-watching-agenc#article"}},{"@type":"BreadcrumbList","@id":"https://groas.com/post/the-1am-spreadsheet-again-watching-agenc#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://groas.com/"},{"@type":"ListItem","position":2,"name":"Blog","item":"https://groas.com/blog"},{"@type":"ListItem","position":3,"name":"Google Ads News","item":"https://groas.com/category/google-ads-news"},{"@type":"ListItem","position":4,"name":"The 1 a.m. Spreadsheet Is Back: Agency AEO’s Manual-Work Trap","item":"https://groas.com/post/the-1am-spreadsheet-again-watching-agenc"}]},{"@type":"BlogPosting","@id":"https://groas.com/post/the-1am-spreadsheet-again-watching-agenc#article","url":"https://groas.com/post/the-1am-spreadsheet-again-watching-agenc","mainEntityOfPage":{"@id":"https://groas.com/post/the-1am-spreadsheet-again-watching-agenc#webpage"},"isPartOf":{"@id":"https://groas.com/blog#blog"},"headline":"The 1 a.m. Spreadsheet Is Back: Agency AEO’s Manual-Work Trap","description":"Picture a typical 1:14 a.m. from my PPC years: cold coffee beside a search-terms export, row 842 glowing on the screen, and one more query that has no business triggering a client’s ad. This is a composite of the nights I spent mining ne…","datePublished":"2026-10-04T05:18:47.423Z","dateModified":"2026-10-04T05:18:48.148Z","image":{"@type":"ImageObject","url":"https://groas.com/media/blog/81109b2fc8bc35a118dd04ec81058cd2497426845214f1b3099b5ce9db261456.png"},"author":{"@id":"https://groas.com/author/david#person"},"publisher":{"@id":"https://graph.groas.com/entity/groas"},"articleSection":"Google Ads News","keywords":"Google Ads, Google Ads News","wordCount":1958,"inLanguage":"en-US"},{"@type":"Person","@id":"https://groas.com/author/david#person","name":"David","jobTitle":"Founder & CEO @ groas","image":"https://groas.com/media/blog/363be9ad5255816654dac065d01bf1c6063110b400bb38cc199570267232027c.jpg","worksFor":{"@id":"https://graph.groas.com/entity/groas"}},{"@type":"FAQPage","@id":"https://groas.com/post/the-1am-spreadsheet-again-watching-agenc#faq","mainEntity":[{"@type":"Question","name":"Why do agencies manually check ChatGPT, Perplexity, and Gemini for client mentions, and is that work actually useful?","acceptedAnswer":{"@type":"Answer","text":"Agencies do it because clients want to know whether buyers encounter their name when asking an AI engine for a recommendation. The tracking has a purpose, but the hours spent collecting screenshots and logging results do not by themselves change what an engine can find or cite, and manual checks get harder to sustain as the client roster grows."}},{"@type":"Question","name":"What happens if a client changes a landing page after the agency has already optimized for AI visibility?","acceptedAnswer":{"@type":"Answer","text":"If the new page removes the facts an engine could use, for example by replacing specific service information with promotional copy, the client may lose its place in AI answers and another source can take over. A prompt check tells you what disappeared, but it does not put the missing context back; the underlying page content has to be repaired."}},{"@type":"Question","name":"Is adding schema markup a good fix when a client disappears from AI answers?","acceptedAnswer":{"@type":"Answer","text":"Structured data may be useful, but adding tags is not a substitute for clear information or a plan to maintain it. Research discussed in the article reports a stronger correlation between branded web mentions and AI visibility than between traditional backlinks and that visibility, which makes 'add JSON-LD everywhere' a thin answer to a client asking why it vanished."}},{"@type":"Question","name":"How much manual work does AI visibility monitoring create for an agency with multiple clients?","acceptedAnswer":{"@type":"Answer","text":"The workload multiplies quickly: ten clients with thirty commercial prompts each across five engines means 1,500 prompt-and-engine checks for a single pass. Tracking software speeds up the watching but adds its own costs, such as per-domain subscriptions, prompt caps, and credit pools, and it still leaves staff to translate alerts into tickets and tickets into edits."}},{"@type":"Question","name":"Should agencies automate AI visibility monitoring instead of doing it by hand?","acceptedAnswer":{"@type":"Answer","text":"The distinction to make is between being accountable for the account and manually performing every recurring check. If a system handles repeatable checks and permitted fixes, a human strategist can focus on setting direction, choosing which commercial questions matter, and answering to the client for the choices being made."}},{"@type":"Question","name":"What should an agency look for before buying AEO monitoring software?","acceptedAnswer":{"@type":"Answer","text":"Three questions reveal where the work will land: does the tool execute approved work or just assign more tasks to staff; what happens to cost as domains and prompts increase; and can the team explain what changed and why. A tool that measures every omission but leaves each fix in an account coordinator's queue makes the old job faster to discover, not easier to finish."}}]}]}
```