Competitor analysis for SEO and AI: read both battlegrounds
Your competitors already ran the experiment. The pages outranking you and the brands AI names instead of you are a ranked to-do list. Here's how to read the gap on both surfaces.
Your competitors are a free roadmap
You don't have to guess what works in your market — your competitors have already run the experiment for you. Every page that outranks you tested a title, an angle, and a depth of coverage that Google decided to reward. Every brand an AI engine names instead of you passed a trust check you haven't yet. Read together, those signals are a ranked list of exactly what to close. Competitor analysis is the discipline of turning that list into a plan you can ship against this quarter.
The mistake most teams make is treating it as a one-time audit — a slide deck built once, admired once, then forgotten. The competitors who beat you don't stop optimizing the day you finish your audit. So the real job isn't a snapshot; it's a standing benchmark you check the way you check rankings.
Two battlegrounds now, not one
Classic competitive analysis looked only at the SERP — ten blue links and who sat where. That surface still matters, but it is no longer the whole game. A growing share of buyer questions never reach a SERP at all: they're answered inside ChatGPT, Perplexity, Gemini, and Google's AI Overviews, where one or two brands get named and the rest are invisible. You now have to read both surfaces, because winning one and losing the other still loses the buyer.
- Search — who ranks above you, how deep their content goes, and the keywords they rank for that you don't (the keyword gap). This is demand you can see and quantify.
- AI answers — when a buyer asks an engine "what's the best X for Y," which competitor gets *named and cited*, and how it's described next to you. This is trust, and it compounds.
The two are linked. AI engines lean heavily on pages that already rank and earn citations, so SERP strength feeds AI presence — but not automatically. A competitor can own the AI answer for a query while sitting at position 6 organically, simply because their page is structured to be quoted. That disconnect is exactly the gap you want to find.
A repeatable workflow
You don't need a research department. You need a short loop you can run monthly.
1. Pick the right competitors
Pick by SERP overlap, not by company size or revenue. The site beating you on *your* terms is a more useful competitor than the famous brand that doesn't compete for them. Pull the domains that show up repeatedly across your priority keywords; three to five real rivals beat a list of twenty aspirational ones.
| Step | What you do | Output |
|---|---|---|
| Pick rivals | Find domains that recur across your target SERPs | 3–5 true competitors |
| Keyword gap | List their ranking terms minus yours | A demand backlog |
| Content audit | Read their top pages for the terms you both want | Depth + format notes |
| AI bench | Ask the engines your buyer's questions | Share-of-voice picture |
| Prioritize | Score by volume × winnability × intent | This quarter's targets |
2. Run the keyword gap
A keyword gap is the set of terms your competitors rank for and you don't. It's the cleanest map of demand you're leaving on the table. Sort it by intent before volume: a 200-search "X vs Y" comparison query from someone choosing a tool is worth more than a 5,000-search informational term from someone who will never buy.
Say you're a project-management SaaS. The gap might surface that two rivals both rank for "gantt chart for small teams" and "client portal software" — terms you have no page for. That's not a content idea; that's a validated content idea, because the demand and the willingness of Google to rank a page already exist.
3. Read the gap, don't copy the page
The point is never to clone a competitor. It's to find the *intent* you're under-serving and serve it more completely. Copying a page that already exists gives Google no reason to prefer yours. Instead, ask where the ranking page is thin — a missing comparison table, no pricing, stale screenshots, no answer to the obvious follow-up question — and win on that margin.
| Signal | Question it answers | Where you act |
|---|---|---|
| Keyword gap | What demand do they capture that I don't? | New pages |
| Content depth | Where do they answer more completely? | Existing pages |
| AI share of voice | Who do engines trust on this topic? | Brand + citations |
| Citations | Whose pages get used as the source? | Linkable assets |
4. Benchmark the AI surface
Ask each engine the questions your buyers actually type — "best [category] for [use case]," "[your brand] vs [rival]," "is [your brand] good for [job]." Record which brands get named, in what order, and which sources get cited. Run the same prompts across engines, because they disagree: a competitor can dominate Perplexity and barely register in Gemini.
A realistic first read for a mid-market brand often looks like this — a clear leader, you somewhere in the pack, and a long tail of names you didn't expect:
That 14% isn't a verdict; it's a baseline. The brands above you are getting named because their pages answer the question cleanly and get cited as the source. The fix is rarely "more content" — it's clearer, more quotable content on the exact questions buyers ask, plus the third-party mentions engines treat as corroboration.
Turn the audit into a backlog
Score each opportunity the same way so the list ranks itself: search volume, how winnable the term is given your current authority, and how close the intent sits to a purchase. High-intent, winnable terms where a competitor is thin go to the top. Informational terms with no path to revenue go to the bottom or get cut.
Pick competitors by SERP overlap, not by company size — the site beating you on your terms matters more than the famous brand that doesn't.
Benchmark it continuously
The reason to make this a loop rather than a one-off is simple: both surfaces move. Rankings shift weekly, AI answers can change with a single model update, and a competitor's new comparison page can take an answer slot overnight. A benchmark you run once tells you where you stood; a benchmark you run on a schedule tells you whether you're gaining.
Lume benchmarks you against competitors on both surfaces — a keyword gap on the SERP side and a multi-engine AI share-of-voice bench across Gemini, ChatGPT, and Perplexity — so you always know the next term to take. Rank tracking runs daily on every plan, so the SERP half of the picture stays current without you refreshing it. Run a free scan and see where you stand on both battlegrounds.
