How do I monitor where my brand is mentioned in AI-generated search results?

Brand monitoring in AI search borrows from social listening: you watch a stream, define what counts, and review it on a schedule. The difference is that the stream is generated. The same question can produce different answers on different days, so the monitoring setup matters more than any single result.

This is how we set it up for clients.

Start with the definition

Before the first prompt runs, we write down what counts as a mention.

Brand name variants, product names and common misspellings.
Whether a mention in a long list counts the same as a top recommendation.
Which competitors we compare against.
Which countries and languages matter.

In Promptwatch, the Brand Book holds the analysis instructions that decide what counts as a mention, and brand aliases and brand merging handle name variants. We do this first because changing the definition later breaks comparisons with earlier data.

Build the panel

We use 30 to 50 prompts per client to start: buying questions, comparison questions, "alternatives to" questions and a small number of direct brand questions. They are tagged by topic, intent and prompt type (organic, brand, competitor comparison), and personas are stacked on top where buyers differ. The Prompt Explorer suggests prompts from Search Console, ranking keywords, People Also Ask or any page, and Prompt Volumes shows search volume and difficulty.

Monitors then run daily per model, country and language, reporting Visibility Score, Share of Voice, Sentiment and Position in Answer.

Watch the sources

This is the difference between counting mentions and monitoring them. A mention is only actionable if you know where it came from.

Citation analytics gives page and domain level citations with URL positions.
Offsite mentions show where the brand is named in third-party pages without a link.
Reddit, YouTube and LinkedIn reports show the platforms engines draw on, with a My channels filter for the client's own presence.
Page Tracker follows specific URLs, including competitors'.
Citation Trends classifies citations by content type and source type.

When a client says "we are not mentioned on this topic", we open the sources behind the answers that name competitors. Usually there is a page, a thread or a roundup that we can influence or replace.

The weekly review

Scan the headline numbers. Visibility Score and Share of Voice against the competitor set.
Open the heatmap. Models against topics, or countries against topics, to find where the brand is absent.
Read the sentiment summary. Each response has a score from 1 to 100, and a weekly AI summary describes what changed.
Check sources for any change. New pages, new threads, lost citations.
Check crawler logs. If our own page lost a citation, were AI bots still fetching it?
Set actions. The Action Board holds a prioritized list, and a Discovery agent suggests up to five evidence-backed actions each week.

The A/B period comparison is how we show a client what changed since the last review.

What gets reported to clients

We keep the report short: share of voice against named competitors, the three biggest gains and losses, the sources behind them, and the action list. Promptwatch generates PDF reports of any dashboard and scheduled PDF reports, and agency plans include custom reports and the Looker Studio connector. Scheduled Chats can deliver a summary to the client in the app or in Slack.

What monitoring cannot do

It cannot make an engine mention a brand, and a mention is not a sale. We connect mentions to outcomes through visitor analytics, which shows AI-referred traffic and conversions via a lightweight script, the GTM template or the WordPress plugin. Where there is a gap in mentions, we pair the monitor with content work: content gap analysis with coverage scores, then new or refreshed pages.

For clients on Business and Enterprise, the Content Agent can plan, write and publish against gaps on a schedule. We use the review inbox. Publishing goes to WordPress, Webflow and Framer.

Plans we use

The free Explore plan is 10 prompts on ChatGPT only. Essential at $95 a month includes all models with 50 prompts. Professional at $245 adds crawler logs and custom reports. Business at $579 adds state and city tracking, Ads Radar and the automated Content Agent. Agency plans, Kick-off at $199, Growth at $399 and Scale at $799, include unlimited projects and prompts.

Our advice

Define the mention, fix the panel and give it two weeks. Then spend your time on the sources, because that is where the work is. We use Promptwatch because the panel, the sources, the crawler data and the traffic sit together, so a missing mention leads straight to a next step.