Competitors

The first time you open Competitors, expect to see at least one name you didn't put there. That's the point of the page, and it's the part most people are unprepared for.

In SEO, you already know your competitors. You picked them. You watch their rankings for the same keywords you care about, and the field is more or less stable. AI search doesn't work like that. When a buyer asks ChatGPT or Perplexity to recommend a tool, the model answers with whatever set of brands it thinks fits the question. Sometimes that's the three companies you'd name in a board deck. Sometimes it's a YC startup you've never heard of, a five-year-old open-source project, or a generalist tool you don't consider in your category at all. Either way, that's who AI just listed next to you.

So the Competitors page isn't a "your rivals" page. It's a record of who AI puts you next to when someone is choosing what to use. The list is built from real AI answers, not a wishlist.

Competitors in AI search aren't the competitors you'd pick

The shift from SEO is small to describe and large to absorb.

The practical consequence: the Competitors leaderboard is a downstream view of your prompts. Change the questions you track, and the competitor set changes with it. Add a question about a use case you don't usually own, and you'll likely see new names. Remove a generic category prompt, and a generalist tool may drop out. The list is honest about who you're actually fighting for, which sometimes means it's honest about things you don't want to hear.

If a competitor you obsess over is missing, that's a finding, not a bug. AI isn't lining you up against them on the questions you're asking. Either they're not in the answer, or the prompt isn't pulling them in.

How competitors get on the list

There is no "add competitor" button, and that's deliberate. Prompts are the only input Trakkr takes from you. Everything else, including who counts as a competitor, comes from the AI answers your prompts produce.

The discovery loop is short:

A brand has to appear in more than a single one-off response before it lands on the leaderboard. That filter is what keeps the list useful. Without it, every passing mention of an irrelevant company would clutter the view, and your rankings would drift around noise.

Two more things shape what you see:

• Name normalization. AI doesn't always write the same brand the same way. "Brooks", "Brooks Running", and "Brooks Ghost" are the same company; left alone, they'd show as three competitors with split scores. Trakkr collapses obvious variants and surfaces the rest in the Manage Brands drawer as suggested groupings you can accept, edit, or ignore.
• Hiding. A genuinely irrelevant brand (a different industry, a personal name, a product Trakkr couldn't classify) can be hidden from your view without affecting raw data. Hidden competitors live in Manage Brands and can be restored any time.

Why you can't manually add one

A manually added competitor would have no data behind it. There would be no responses they appeared in, no prompts to compare on, no head-to-head record, no trend. The leaderboard line would be empty.

If you want a specific competitor to show up, the right lever is the prompts. Add the questions where you'd expect them to come up. Comparison prompts ("project management tools for remote engineering teams"), category prompts a buyer in their wheelhouse would ask, prompts that target a use case they own. If they're a real competitor on those questions, they'll appear in the next run. If they don't, AI doesn't see them as part of your competitive set, which is itself a fact worth knowing.

What gets compared

Once the list exists, the page lines you up against everyone on it across two views.

Competitors view (the leaderboard)

The default view is a ranked list, you and every tracked competitor in one column, sorted by visibility on your prompts. Each row carries the numbers that let you read where you stand without clicking in:

Click any row and it expands in place. The expansion is the head-to-head: their trend against yours over time, the specific prompts they win on, the prompts you win on, the gap on each one, and (where available) the citation sources behind their wins.

At the top of the page, two charts toggle alongside the leaderboard:

• Trend plots everyone's visibility over time, so you can see who's accelerating and who's coasting.
• Share breaks the same data into a share-of-voice view, useful when you care about who owns the most mention volume in your category.

Prompts view (the battleground)

Flip the toggle and the page reorients. Now each row is one of your tracked prompts and the columns ask: who's winning this one? You see the leader, your position, the gap, the trend, and (on expand) the actual AI response so you can read what won.

This is the view you reach for when the question isn't "how am I doing overall" but "which questions am I losing, and to whom." It's also where you go to chase a specific competitor: filter prompts by leader and you get a focused list of every question they're beating you on.

Model breakdown

Both views can be sliced by AI model. A competitor that buries you on ChatGPT may be invisible on Claude. Filter to a single model and the leaderboard reranks for that model alone, so model-specific dynamics aren't averaged into mush.

You don't have to win every model. The right question is: where do my buyers actually search, and what does the picture look like there?

How to read the leaderboard

A few patterns come up enough to call out.

The unknown #2. A name you didn't expect, sitting near the top. Worth ten minutes of research before doing anything else. Sometimes it's a real competitor your team underweighted. Sometimes it's a content site or aggregator AI keeps recommending instead of a vendor. Sometimes it's a brand that fits a use case you don't think of as yours but should.

The leader who's slowing. A high score with a flat or falling trend. They're defending, not gaining. If your trend is up, you're closing the gap; if it's flat too, the market has stabilized.

You're #1 but barely. Top of the list with thin margins and rising challengers. The work is keeping the citations and coverage that put you there, not chasing the next prompt.

A flat field around you. Everyone hovering in a tight band. Usually means the prompts are generic enough that AI fans recommendations widely. Sharper, more specific prompts often clear up the picture.

A new entrant. A brand that wasn't on the list last month is suddenly mid-pack. That's a real signal in AI search, training data updates and the surface of mentions shifts. Worth checking what they're doing, especially in citations.

The leaderboard tells you what's happening. The expanded row tells you why and where. Both are read top to bottom.

What to do when you're losing

Losing is information. It's almost always one of three things, and each has a different response.

You're missing from key citations

If a competitor wins a prompt because they're cited on the sources AI trusts and you aren't, the fix is in Citations. Expand the competitor's row, find the sources behind their wins, and work down the list. PR, partnerships, contributing to industry listicles, getting on the right comparison pages, this is where citation work pays.

The prompt favors them by design

Some prompts are structurally tilted. A question about open-source alternatives won't list a closed-source product no matter how much you push. A question about enterprise pricing will favor the brands AI associates with enterprise. If a competitor is winning a prompt that's genuinely about a use case they own, the right move isn't to fight that prompt, it's to make sure you own your own prompts.

Filter to your own wins and ties, and make sure your strongest territories are defended.

Your positioning isn't in the model's training data

If you're a newer brand or you recently repositioned, AI may not have caught up. The fix is slower: getting written about, getting listed, getting reviewed, all of which feed the next round of model updates. Research helps here by surfacing where you're invisible vs where you're recommended-but-buried, so you can prioritize.

Whichever bucket a loss falls into, the next-step pattern is the same:

1. Open the prompt in the expanded row.
2. Read the actual AI response to see what won.
3. If a competitor is cited, check which sources cite them.
4. If you have an obvious gap, send it to Actions so the work doesn't get lost.

The page is designed for that loop. Status filters (threats, rising, trailing) let you triage the leaderboard. The prompts view lets you triage the questions themselves. Either is a defensible starting point.

Plan limits

Free accounts see a teaser. Paid plans get the full feature, with depth scaling up:

Common questions

Why is a brand I've never heard of on my list?

Because AI listed them in response to one of your prompts, more than once. That's the criterion. If the brand is genuinely irrelevant (different industry, personal name, mis-extracted text), hide it from Manage Brands and the leaderboard cleans up. If it's relevant but surprising, that's usually the most valuable thing the page has told you, take a look at the prompts they're winning.

A competitor I care about isn't showing up. How do I add them?

You can't add them directly, and that's the point. Their absence means AI isn't recommending them on the prompts you've added. Either add prompts they'd plausibly appear on (use cases they own, comparison questions, category questions tilted toward their strength) or accept that they aren't in this competitive set, which is itself useful intelligence.

Why do I have three rows for what looks like the same competitor?

Name variants. Open Manage Brands, go to the Competitors tab, and either accept the auto-detected grouping or create one yourself. Grouping consolidates the visibility and gives you a clean read.

Visibility on the leaderboard doesn't match my dashboard score. Why?

The dashboard score covers your full prompt set, all models, with the default window. The leaderboard recalculates for whatever date range, model filter, and tag scope you have active on the Competitors page. Reset the filters and the numbers line up.

How fresh is the data?

Competitor data is built from your daily report runs. New prompts and new mentions take a day to land. If you've just added prompts or made big changes, give it a day before reading too much into the leaderboard.

Does adding a competitor cost a prompt slot?

There's no such thing as adding a competitor, so no. The cost on this page is whatever prompts you've added; competitors come along for free.