Citino teamTo see where competitors beat you in AI answers, collect every answer that names a competitor and not you, group the losses by question and engine, and read why each answer chose the competitor. The reasons the answers give and the pages they cite tell you whether each loss is something you can fix.
This is a diagnosis step in AI search analytics: it turns a share of voice number into a list of specific questions and causes.
Step 1: Find the answers you lose
Start from a fixed set of buyer questions, asked on each engine on a schedule. From all the answers, keep the ones that name at least one competitor and do not name you. That is the most useful filter in AI visibility data: every answer in it is a buyer who heard about someone else.
Then add the near misses: answers that name you, but well behind a competitor. Being fifth in a list of six is closer to a loss than a win.
Step 2: Group the losses
One lost answer tells you little; answers vary from run to run. Group losses so patterns show:
| Group by | What it shows |
|---|---|
| Question topic | Where you lose: billing questions, integrations, a use case |
| Engine | Whether the loss is everywhere or on one engine |
| Competitor | Who takes the answers, and on which topics |
| Question type | Category, problem, comparison or alternatives questions |
Rank the groups by how often you lose and how much the questions matter to buyers. A topic where you lose most answers on every engine is worth more attention than one lost answer on one engine. How to measure share of voice in AI answers covers how to count this with a margin of error.
Step 3: Read the reasons the answers give
AI answers do not just name brands; they explain the choice. “Northport is the stronger option for 3PLs because it bills each client natively.” Pull every reason given for the competitor out of the lost answers, and count how often each one appears across engines and runs.
The reasons that repeat are the ones the sources agree on. Write them down as plain claims, such as “native multi-client billing” or “NetSuite integration”. The brand names here are fictional.
Step 4: Read the pages the answers cite
Engines that search the web cite what they read. Google says AI Overviews and AI Mode may run multiple related searches across subtopics to build a response; Perplexity says every answer has numbered citations to its sources; and ChatGPT’s search answers carry citations to the pages used.
For each lost topic, list the pages cited and sort them by owner:
| Cited page | What it tells you |
|---|---|
| Your own page | You are read, but the page does not make the case |
| A third-party page that names the competitor and not you | A review site, directory or article you are missing from |
| The competitor’s own page | Why they win; nothing for you to change there |
Step 5: Diagnose each loss
Every loss comes down to one of three causes:
| Cause | Sign | Fix |
|---|---|---|
| Source gap | Cited third-party pages cover the competitor and leave you out | Get onto those pages: how to find and fix source gaps |
| Content gap | The reason given is true of you, but your pages never state it plainly | Write it where engines read: how to fix content gaps |
| Product difference | The reason given is not true of your product | Leave it, or change the product; do not write claims you cannot back |
A topic can have more than one cause. Fix the source gaps and content gaps first; they are the ones a page change can move.
Step 6: Fix, then check against the field
Record which questions each fix targets before it ships. Once engines have read the changed page, measure those questions again and compare your change with the median competitor’s over the same days. If everyone moved, the move was not yours. How to tell whether an AI search fix worked sets out the steps.
The answers you lose already explain why you lost them. Read the reasons and the citations before you write anything.
Where Citino helps
Citino is an AI search analytics tool. It asks ChatGPT, Google AI Overviews and Perplexity your buyers’ questions every day, with Gemini, Google AI Mode and Claude as add-ons, and keeps every answer with the brands it named and the pages it cited. You can filter for every answer that named a competitor and not you, see the sources behind it sorted by who owns them, and turn each loss into a work item with a verdict once the fix is live. See Answers and Gaps.