Competitive intelligence

Competitive intelligence compares your visibility with the brands competing for the same answers. Home's share of voice and Who AI names include every brand the engines named; the head-to-head standings and trend charts use only the competitors you track.

Adding competitors

Add competitors during onboarding, from site settings, or from the Competitive intelligence page. Each competitor needs its official public website URL, so Stride can attribute cited pages and compare the right brand. Your plan controls how many competitor slots are available. During onboarding Stride proposes up to ten competitors from your site; the ones you keep are saved in order until your plan's slots are full.

Choose the brands that actually appear in AI answers for your prompts — not necessarily the ones you consider competitors commercially. The two lists are often different, and the first is the one that matters here.

Who AI names

After a full scan, Who AI names lists every brand the engines named in your latest scan's measured answers: you, the competitors you track, and brands you don't track yet. Each row shows how many answers named the brand, across how many of your prompts, and which engines named it. A brand counts once per answer, however many times the answer repeats it.

Engines spell the same brand differently ("Maye the Label", "MAYE THE LABEL", "Mayé"). Stride combines spellings of one brand into a single row and lists the other spellings under it. It only combines names it can match with confidence, so a brand can occasionally still appear twice under two very different spellings.

Answers can also name retailers, publications, or adjacent products. Choose Not a competitor to hide a brand from this list, or choose Track, confirm the brand name, and enter its official website.

Untracked brands never enter the charts or history automatically. Stride applies your normal competitor slot limit and adds a tracked brand to comparisons after the next full scan.

If none of your answers named a brand, the section says so. Prompts that ask which brand or product to choose tend to draw named recommendations; advice questions often don't.

Reading the comparison

Competitor Comparison measures the fraction of successful answers naming you or each tracked competitor in the latest trusted full scan. Both sides use the same answers and configured brand identities; a brand can be named without its website being cited. The comparison includes ChatGPT, Perplexity, Gemini and Claude when measured. Engines with fewer than three successful answers are withheld rather than shown as a measured zero.

Home’s Share of Voice lists you, your tracked competitors and the brands AI named most, and divides each brand’s mentions by the mentions of every brand named in that scan, tracked or not. The percentages therefore differ from the comparison above even when the ranking agrees. A comparison may also differ when an engine lacks enough answers for a reliable per-engine rate.

Each tracked competitor gets its own card:

• Ahead, Close or Behind, with the gap in points. Within 5 points counts as Close.
• By engine · you % / them %: both rates on each engine with enough answers.
• Prompt gaps: prompts where the competitor was named and you weren't, with the engines it happened on. The card lists up to eight; the rest are in Actions.

Google AI Overviews isn't part of the competitor comparison.

Is the gap closing?

Is the gap closing? charts your rate against each tracked competitor across recent full scans, so you can see whether you're catching up or falling behind. Hover or focus a scan to see what changed since the previous one. It needs at least two comparable scans; daily priority checks don't count.

Turning gaps into work

Choose Close these gaps on a competitor card to open its action items in Actions. A prompt gap seen on at least half of the engines becomes an action to write about that question, ranked higher the more engines show it. An engine-level lead becomes an action to close the gap against that competitor. Look at what the engine cites when it names them; that's usually the content shape you're missing.