AI Competitor Mention Map Template

AI Competitor Mention Map (fill-in)

When to use: One row per buyer query, with named alternatives, source pages, fit context and an action, plus owner fields and a completion check.

AI competitor mention map

One row per buyer query. Write down which alternatives the AI answer names, the pages it draws on, the fit it gives each one, and what you will do about it.

Map owner: [name] - Your brand: [brand + domain] - Engines checked: [e.g. ChatGPT, Perplexity, Gemini, Google AI Overviews] - Map date: [YYYY-MM-DD]

Per-row detail (fill for every query)

• Query type: [best X / X for Y / X vs Y / alternatives to X / how to choose X]
• Named alternatives: [brand 1], [brand 2], [brand 3] - your position: [named first / named later / not named]
• Framing per brand: [recommended / listed / cautioned against]
• Source pages: [URL] - type [rival site / review site / list article / forum / docs / your site] - supports [which brand]
• Fit context: [the buyer, size, budget, region or use case the answer ties to each brand]
• Fit check: [is your fit described correctly, missing, or wrong]
• Evidence: [screenshot + raw answer text file, named by row ID]

Next action

• Map owner: [name] - task: fill every row for the current query set by [date]
• Content owner: [name] - task: fix or build the page for each fix/build row by [date]
• Outreach owner: [name] - task: contact the source pages on each earn row by [date]
• Re-check: [name] - task: re-run the same queries on [date] and update the map

Completion check

• Every row has the exact query, the engine and the date.
• Named alternatives are listed in answer order, and the row says if you were named.
• At least one source page is written in full for each named alternative, or marked none shown.
• Fit context is quoted or closely noted from the answer, not your own opinion.
• Every row has one action and one owner with a date.

Named Alternatives and Fit Context Guide

When to use: How to read one AI answer and code the brands, their order, framing and the fit each is given, the same way every time.

Named alternatives and fit context: how to read one answer

Read the answer slowly once. Then fill the row in this order, so two people code the same answer the same way.

Step by step

• List every brand named, in the order the answer gives them. Include brands named only in a caveat.
• Mark your own brand: named, named with a link, or not named.
• Tag the framing for each brand: recommended, listed, or cautioned against.
• Copy the fit words. "Best for small teams" or "good if you need offline mode" is the fit context. Quote it short.
• Note what decided the fit: price level, company size, region, feature, integration or support.

Fit context codes

A NONE row matters. It means the engine knows the brand but has no reason to pick it. It may be a page or source problem rather than a product one.

To see recorded AI answers that name a rival, enter its domain in the AI Visibility Checker, open a prompt and read the full answer and its cited sources.

Source Page Tracer

When to use: Trace which cited pages stand behind each named alternative, and whether you can get onto them.

Source page tracer

AI answers lean on pages. This sheet traces which pages stand behind each named alternative, so you know where the recommendation comes from.

How to read the pattern

• Rival's own pages cited: their page answers the query better than yours. Compare page to page.
• List articles and review sites cited: third parties carry the recommendation. You need to be on those pages.
• Forums cited: real users mention the rival. Check whether your users talk about you there.
• Your page cited, but a rival named: your page may describe the category without making your own fit clear.
• No sources shown: write none shown. Do not guess the source.

Open a rival's domain in the Backlink Analyzer to list the pages linking to it, then check which of those also appear as AI sources.

Action Picker (fix, build, earn, watch)

When to use: Turn each row into one action with an owner, based on what the answer and its sources show.

Action picker

Every row ends in one action. Pick it from what the row shows, not from the brand you worry about most.

Before you act

• Re-run the query at least once on another day. One answer is a sample, not a pattern.
• Check the fit claim against facts. Never copy a rival's claim into your page without checking it.
• One row, one action, one owner. Split a row if it needs two actions.

For a build row, plan the page with the Content Brief generator. For a gap that also shows in Google, compare domains in Competitor Analysis.

Publishing a comparison page from a row? Log every rival claim first in the Competitor Comparison Evidence Matrix.

Spreadsheet Header (CSV)

When to use: A paste-ready header row and allowed values to run the map in Google Sheets or Excel.

Paste-ready spreadsheet header

Paste this row into cell A1 of Google Sheets or Excel, then split text to columns by comma. Use one row per query, per engine, per date.

row_id,map_date,query,query_type,engine,model,location,your_brand_named,your_position,alt_1,alt_1_framing,alt_1_fit,alt_2,alt_2_framing,alt_2_fit,alt_3,alt_3_framing,alt_3_fit,source_urls,source_types,fit_check,action,action_owner,due_date,status,evidence_file,notes

Allowed values

• query_type: best_for, alternatives_to, versus, how_to_choose, category
• your_brand_named: named_linked, named_no_link, not_named
• alt_framing: recommended, listed, cautioned
• alt_fit: SIZE, BUDGET, USE, FEAT, REGION, NONE
• fit_check: correct, missing, wrong
• action: fix, build, earn, watch, rerun
• status: open, in_progress, done, rechecked

Sheet tips

• Add a pivot of alternatives by query. It shows which rival owns which question.
• Highlight rows where your_brand_named is not_named and action is empty.
• Keep a second tab that lists every source URL once, with the brands it supports.

Need a clean query set first? Build it in the AI Buyer-Query Library Template, then map each query here.

Worked Example (fictional Keelwren marina software)

When to use: Three filled rows for an invented marina software brand: one earn, one fix and one build action.

Worked example: Keelwren marina software

Example only. Keelwren, Tidebrisk, Harbellion and Berthollow are invented brands. The queries, answers, URLs, names and dates are made up to show a filled-in map.

Map owner: Dana (marketing) - Your brand: Keelwren, keelwren.example - Engines checked: ChatGPT and Perplexity - Map date: 2026-11-10

How the rows were filled

• Row 1: the only source was a third-party list that skips Keelwren. That makes it an earn row, not a content row.
• Row 2: Keelwren was named, but the fit was wrong. The cited features page never said who the product is for.
• Row 3: the rival's own comparison page was the only source, and it says Keelwren has no offline mode. The team checked: in this example the claim is out of date, so they build a page with dated facts.

Why this works

• Each action follows from the source pages, not from a guess.
• A wrong fit is treated as seriously as a missing name.
• Every row has one owner and a date, and all three queries are re-run on 2026-12-10.

Want help turning a mention map into a GEO plan? See our SEO and GEO services, or book a free consultation with Anatolii Ulitovskyi.

Monthly Map Review

When to use: A monthly routine to spot new rivals, fit drift, shared source pages and stalled actions.

Monthly map review

Once a month, the map owner reads the map as a whole. Rows show answers. The review shows patterns.

Build the summary grid

Checks

• New names: any brand you have never tracked that now shows up.
• Fit drift: any brand whose fit changed between maps, including yours.
• Shared sources: one page cited for several rivals is a priority target.
• Open actions: rows with no progress get a new date or a new owner.

Review fields

• Review date: [YYYY-MM-DD] - Reviewer: [name]
• Rival named most often: [brand]
• Query where you are missing that matters most: [query]
• Top three actions for next month: [row IDs + owners]

Answers move between runs. Log repeat runs of your most important queries in the AI Visibility Repeated-Run Log before you call a change a trend.