--- title: "AI Search Standup Template: A 30-Minute Weekly Agenda for Marketing Teams" description: "Run a 30-minute weekly AI search standup with a pre-read report template, a timed agenda, a roles table and an act, watch or ignore rule for each change." canonical: "https://zerply.ai/resources/blog/ai-search-standup-template" author: "Anshul Motwani" date: "2026-11-11T05:01:26+00:00" category: "AI Workflows" image: "https://storage.zerply.ai/teams/92/blogs/612/d2c3a024eb2a28a2-1793771631838-a91038d9-dad9-4783-8d34-adcfc552fc1a.png" --- # AI Search Standup Template: A 30-Minute Weekly Agenda for Marketing Teams > **The short answer:** An AI search standup is a 30-minute weekly meeting where a marketing team reviews AI visibility and search data against earlier weeks, assigns an owner to each change worth acting on, and checks last week's actions. A fixed agenda and fixed roles keep the time on decisions instead of screen-sharing dashboards. Put an AI visibility dashboard on a shared screen, and the meeting turns into a tour. Someone scrolls through mention rates by engine, someone asks why Gemini looks different from last week, and 30 minutes later nobody owns anything. The Gemini question has a dull answer. In a [SparkToro and Gumshoe study](https://sparktoro.com/blog/new-research-ais-are-highly-inconsistent-when-recommending-brands-or-products-marketers-should-take-care-when-tracking-ai-visibility/) of 2,961 prompt runs, ChatGPT and Google's AI returned the same list of brands less than once in 100 runs of the same prompt. The same study found that how often a brand appears across many runs is the steadier measure. A standup built on one week's screenshots tracks noise, so this format reviews trends across weeks and ends each item with a decision. ## What a weekly AI search standup covers A weekly AI search standup reviews four inputs, each compared against the rolling average of the previous four weeks: - AI visibility from your prompt tracking, meaning how often each tracked prompt mentions or cites your brand and which competitors appear in your place. - Citations by page, meaning which of your URLs AI answers cite and which ones lost citations. - Search performance from Google Search Console, meaning clicks, impressions and average position for the pages on the list. - AI referral sessions in GA4, meaning visits arriving from ChatGPT, Perplexity and other assistants, and what these visitors do. The fifth input is last week's action list. Every standup starts by checking it, because a meeting that only adds actions turns into a backlog nobody reads. Schedule the meeting for Wednesday or Thursday. Search Console data arrives [about 48 hours after it's collected](https://support.google.com/analytics/answer/13682862), so a Monday meeting reviews a week with two days missing. ## SEO weekly report template for the pre-read The data owner sends a one-page report the day before the meeting, so the standup spends its 30 minutes on decisions instead of reading numbers aloud. The report holds five fields: | Field | What goes in it | | --------------------- | ---------------------------------------------------------------------------------------------------------------------------------------- | | Prompt changes | Tracked prompts whose mention rate moved more than 10 percentage points against the four-week average, with the competitor now appearing | | Page citation changes | URLs cited in at least three fewer tracked prompts than their four-week average, with the URL now cited in their place | | Search changes | Pages on either list whose Search Console clicks are down more than 20% against the four-week average | | AI referral traffic | GA4 sessions from AI assistants by landing page, with key events, against the four-week average | | Open actions | Last week's actions with status: done, in progress or blocked | The thresholds are starting points. Raise them if the report runs past one page, and lower them if it comes back empty two weeks in a row. An AI visibility report longer than one page gets skimmed, and the item that needed a decision sits in the part nobody read. ## AI search standup agenda template The agenda splits 30 minutes into five blocks, and each block ends with a written output: | Time | Agenda item | Led by | Output | | ------------ | ------------------------------------------------- | ----------- | -------------------------------------------------------- | | 0 to 5 min | Last week's actions | Facilitator | Each action marked done, in progress, blocked or dropped | | 5 to 12 min | Prompt and citation changes from the pre-read | Data owner | Questions answered, data disputes settled | | 12 to 18 min | Search and AI referral changes for the same pages | Data owner | Context for each change | | 18 to 26 min | Act, watch or ignore for each change | Facilitator | Action list with one owner and a due date per action | | 26 to 30 min | Blockers that need someone outside the room | Anyone | Escalations with a named person to contact | Use the same agenda as your weekly standup meeting template every week. A fixed order lets people prepare, and it shows quickly when one block keeps running over, a sign the pre-read thresholds are too low. ![30-minute AI search standup agenda with five blocks, each showing who leads it and the written output it produces](https://storage.zerply.ai/teams/92/blogs/612/fe8044d670b2ca94-1793771675592-91ee4bb9-f92d-4cbd-a9e9-3998b351dfb5.png) ## Who runs the standup: roles and owners The standup needs four roles. On a small team one person can hold two, but every role needs a name: | Role | Typical owner | Before the meeting | During the meeting | After the meeting | | --------------- | --------------------------------- | -------------------------------------------------- | ---------------------------------------------- | -------------------------------------------- | | Facilitator | Head of content or marketing lead | Checks the action list is current | Keeps time and forces a decision on every item | Posts the action list the same day | | Data owner | SEO or analytics lead | Sends the one-page pre-read a day ahead | Presents changes and answers data questions | Reviews thresholds once a month | | Content lead | Content manager or senior editor | Reads the flagged pages | Takes content actions and estimates effort | Briefs writers on new actions | | Technical owner | Technical SEO or web developer | Checks flagged pages for crawl and indexing issues | Takes technical actions | Files tickets for anything needing a release | Product marketing joins when a flagged page makes claims about pricing, features or positioning. Keep the room to five people or fewer. Each extra person adds questions to the meeting without adding owners for the work. ## How to decide what to act on: act, watch or ignore Every change in the pre-read leaves the meeting with one of three labels: - **Act** when the change passes its threshold for a second week in a row, or when it hits a prompt or page tied to revenue, such as a pricing page or a "best tool for" prompt. - **Watch** when the change passes its threshold for the first time on a prompt or page with no direct revenue tie, so the team checks it again next week before spending time on it. - **Ignore** when the change sits under its threshold, or when it affects prompts the team has already decided not to compete for. Ignore is the hardest label to give, and it keeps the meeting useful. Cap new actions at five a week. A list of fifteen never gets finished, and the unfinished items crowd out next week's review. ![Act, watch or ignore decision rule for AI visibility changes, with the condition and an illustrative example for each label](https://storage.zerply.ai/teams/92/blogs/612/6e056dc4442b2486-1793771683467-bc3cd200-7585-4387-a347-72f4618f82ad.png) ## Example: one week's standup notes The notes below show one meeting's output. The prompts, pages and numbers are illustrative: | Change | Evidence | Decision | Owner | Due | | ------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------- | ----------------- | ----------------- | | "Best payroll software for small business" prompt group | Mention rate 28% against a four-week average of 42%, second week below threshold, a competitor's comparison page now cited | Act: refresh the comparison page | Content lead | Next Friday | | /blog/payroll-compliance-checklist | Cited in three fewer prompts than its four-week average, first week | Watch | Data owner | Recheck next week | | Gemini mentions across all prompts | Down 6 points, under the 10-point threshold | Ignore | None | None | | /pricing | ChatGPT referral sessions up 35% against the four-week average | Act: confirm every price on the page is current | Product marketing | Wednesday | | Last week's schema fix on /integrations | Blocked on a release | Escalate to the web team lead | Technical owner | This week | The pricing row is the easiest to misread. A rise in AI referral traffic looks like good news to file away, but it means an assistant is sending buyers to the page, so every price on it has to be right this week. ## How often should you hold an AI search standup? Hold it weekly if your team publishes or updates pages every week, or tracks enough prompts for a weekly report to fill one page. Move to every two weeks if the pre-read comes back with fewer than three changes for three weeks running, because at that point weekly variation in AI answers produces more noise than decisions. Never run it daily. The SparkToro and Gumshoe data shows single AI answers vary from run to run, so a daily review reacts to that variation instead of to anything your team changed. ## Running the AI search standup in Zerply Zerply's AI Visibility Tracking supplies the prompt section of the pre-read, with a mention rate for every tracked prompt and the competitors appearing where you don't. [Citation Decay](https://zerply.ai/resources/blog/introducing-citation-decay) supplies the page section. Each cited page gets a status of Rising, At peak, Decaying, Past half-life or Low volume, and the Who took the slot panel names the URL now cited in your place. In alternate weeks, [Signals](https://zerply.ai/resources/blog/introducing-signals) can replace the pre-read. It arrives every two weeks with your costliest gaps, a numbered plan and a record of which past recommendations were resolved or progressed, and the Execute with Zerply agent button opens a chat already briefed on any item the standup assigns. For the search and traffic blocks, Zerply Agent pulls your AI visibility data, Search Console and GA4 into one chat. A data owner can ask "Which tracked prompts lost mention rate this week, and did the matching pages also lose Search Console clicks?" and bring the answer to the meeting. Conversations stay saved in the sidebar, so the next week's standup can pick up the same thread. For more on the GA4 side, see [how to measure AI referral traffic](https://zerply.ai/resources/blog/how-to-measure-and-act-on-ai-referral-traffic). Signals needs enough tracked data to produce its first report, and Citation Decay needs several weeks of citations before it charts a page. Standard plans track a fixed four engines, ChatGPT, Google AI Overviews, Google AI Mode and Gemini, with Perplexity, Claude and Copilot on Enterprise. ## Conclusion Book 30 minutes on Wednesday, name the four role owners, and send the first pre-read the day before. Run the agenda unchanged for four weeks. Then compare the action list against what changed. If most actions came from one section of the pre-read, give that section more time and tighten the thresholds on the rest. ## Frequently asked questions ### What is an AI search standup? An AI search standup is a 30-minute weekly meeting where a marketing team reviews AI visibility, Search Console and GA4 data against the previous four weeks. Each change leaves the meeting labeled act, watch or ignore, and every action gets one owner and a due date. ### What should be on an AI search standup agenda? A 30-minute agenda covers last week's actions, prompt and citation changes, search and AI referral changes for the same pages, a decision on each change, and blockers. Each block ends with a written output, such as an updated action list. ### Who should attend a weekly AI search standup? Four roles cover the meeting: a facilitator, a data owner, a content lead and a technical owner. Product marketing joins when a flagged page makes pricing or feature claims, and the room works best at five people or fewer. ### What should an SEO weekly report include? A one-page weekly report should list tracked prompts whose mention rate moved more than 10 percentage points, pages cited in fewer prompts, pages losing Search Console clicks, AI referral sessions in GA4, and the status of last week's actions. Each figure is compared against the previous four-week average. ### How often should marketing teams review AI visibility? Weekly reviews suit teams that publish or update pages every week. Teams with few tracked prompts can review every two weeks, and daily reviews react to run-to-run variation in AI answers instead of real change. ### Why do AI visibility numbers change from week to week? AI answers vary between runs of the same prompt. A SparkToro and Gumshoe study of 2,961 prompt runs found ChatGPT and Google's AI returned the same brand list less than once in 100 runs, so trends across weeks are more reliable than single results. ```json {"@context":"https://schema.org","@type":"Article","headline":"AI Search Standup Template: A 30-Minute Weekly Agenda for Marketing Teams","description":"Run a 30-minute weekly AI search standup with a pre-read report template, a timed agenda, a roles table and an act, watch or ignore rule for each change.","url":"https://zerply.ai/resources/blog/ai-search-standup-template","image":"https://storage.zerply.ai/teams/92/blogs/612/d2c3a024eb2a28a2-1793771631838-a91038d9-dad9-4783-8d34-adcfc552fc1a.png","datePublished":"2026-11-11T05:01:26+00:00","dateModified":"2026-11-11T05:01:26+00:00","author":{"@type":"Person","name":"Anshul Motwani"}} {"@context":"https://schema.org","@type":"FAQPage","mainEntity":[{"@type":"Question","name":"What is an AI search standup?","acceptedAnswer":{"@type":"Answer","text":"An AI search standup is a 30-minute weekly meeting where a marketing team reviews AI visibility, Search Console and GA4 data against the previous four weeks. Each change leaves the meeting labeled act, watch or ignore, and every action gets one owner and a due date."}},{"@type":"Question","name":"What should be on an AI search standup agenda?","acceptedAnswer":{"@type":"Answer","text":"A 30-minute agenda covers last week's actions, prompt and citation changes, search and AI referral changes for the same pages, a decision on each change, and blockers. Each block ends with a written output, such as an updated action list."}},{"@type":"Question","name":"Who should attend a weekly AI search standup?","acceptedAnswer":{"@type":"Answer","text":"Four roles cover the meeting: a facilitator, a data owner, a content lead and a technical owner. Product marketing joins when a flagged page makes pricing or feature claims, and the room works best at five people or fewer."}},{"@type":"Question","name":"What should an SEO weekly report include?","acceptedAnswer":{"@type":"Answer","text":"A one-page weekly report should list tracked prompts whose mention rate moved more than 10 percentage points, pages cited in fewer prompts, pages losing Search Console clicks, AI referral sessions in GA4, and the status of last week's actions. Each figure is compared against the previous four-week average."}},{"@type":"Question","name":"How often should marketing teams review AI visibility?","acceptedAnswer":{"@type":"Answer","text":"Weekly reviews suit teams that publish or update pages every week. Teams with few tracked prompts can review every two weeks, and daily reviews react to run-to-run variation in AI answers instead of real change."}},{"@type":"Question","name":"Why do AI visibility numbers change from week to week?","acceptedAnswer":{"@type":"Answer","text":"AI answers vary between runs of the same prompt. A SparkToro and Gumshoe study of 2,961 prompt runs found ChatGPT and Google's AI returned the same brand list less than once in 100 runs, so trends across weeks are more reliable than single results."}}]} ```