A vendor I never heard of beats us in AI answers - what now?

Verify first: re-run the exact question and its variants several times, because AI answers vary between runs. If the vendor appears consistently, diagnose which pages carry them and why, fix that specific gap, and re-scan the same questions after a few weeks to confirm the fix worked.

How do I respond to a new competitor in search?

Scope it before responding: one question or many, one surface or several. Then match the response to the diagnosis - a dedicated answer page if they win on directness, review-site presence if they win through third-party surfaces, citation work if a comparison omits you. Avoid a wholesale positioning rewrite over one rival.

Why does a smaller company outrank us?

Usually because specific pages carry them: a dedicated page that answers the question directly, a strong review-site profile, or a third-party comparison that includes them. Size matters less on answer surfaces than clarity, specificity and presence on the sources those engines cite - which is fixable once you identify the exact pages.

How long before a fix shows up in AI answers?

Typically weeks rather than days, and the exact lag is not knowable in advance - answer engines refresh their sources on their own schedules. That is why the playbook ends with re-measuring the same questions at a realistic interval instead of judging the fix within a few days.

What to do when an unknown vendor outranks you

Guide · AI Visibility · 5 min read · last verified 2026-07-27

Reviewed before publication Editorial board Independent commercial review

In shortA calm incident playbook for being outranked by an unknown vendor: verify with repeated evidence, read which pages carry them, close that specific gap, and re-measure on the same questions.

When a vendor you have never heard of outranks you - in search results or in AI answers - the correct response is a four-step incident procedure: verify the observation with evidence, read why they are being preferred, close the specific gap their advantage exposes, and re-measure on exactly the same questions after a sensible interval. What the moment does not call for is panic, a positioning rewrite, or a company-wide email. One surprising answer is a data point; the procedure turns it into a decision.

Treating it procedurally matters because these moments are becoming routine. Answer surfaces are volatile, new vendors invest early in them, and the first anyone internally hears of a rival is often a screenshot from a prospect or a colleague. A calm playbook beats an improvised reaction every time.

Step 1: verify before reacting

Start by establishing that the observation is real and repeatable, because single observations mislead. AI answers vary between runs, between phrasings and between users; search results vary by location and history. Re-run the exact question several times, and run the near-variants a buyer would plausibly ask. Record what you see each time.

Three outcomes are possible. The vendor appears once and never again: log it and move on. The vendor appears intermittently: worth watching, not worth mobilising. The vendor appears consistently across runs and phrasings: you have a genuine incident, and it deserves the rest of this playbook.

While verifying, scope it. Is this one question, a cluster of related questions, or your whole category? Is it one surface - a single AI assistant, one search engine - or several? A vendor consistently recommended across many questions and surfaces is a strategic matter; a vendor winning one long-tail question is a content task. The response should match the scope, which you cannot know until you measure it.

Step 2: read why they are preferred

Outranking is never magic; it has a citation surface. AI assistants and search engines prefer sources, and the vendor beating you is being carried by specific pages. Your job is to find them.

Look at which pages are cited or linked when the vendor appears: their own documentation or comparison pages, a third-party review site, a community thread, an industry publication. Then characterise the advantage. Common patterns include: they answer the question directly on a dedicated page while your equivalent content is buried in a feature tour; they are present on a review platform or directory that the answer engine leans on and you are not; a third party wrote a comparison that includes them and omits you; or their content simply matches the buyer's phrasing while yours matches your internal vocabulary.

Be honest about the least comfortable possibility: sometimes the smaller vendor is preferred because their material is genuinely clearer or more specific than yours. That is not an algorithm being unfair; it is feedback.

Write the diagnosis down in one sentence per question: who wins, on which surface, carried by which pages, for what apparent reason. This sentence dictates the fix.

Step 3: close the specific gap

The fix should be as narrow as the diagnosis. If they win because a dedicated page answers the question and yours does not, create or restructure your page so the direct answer appears at the top, in the buyer's words. If they win through third-party surfaces - review sites, directories, comparison articles - your move is presence on those surfaces: complete the profile, pursue reviews honestly, or offer the publisher genuinely better material. If they win because a comparison omits you, the gap is relationship and citation work, not on-site content at all.

Prioritise by journey stage. Losing an awareness question is a slow leak; losing a comparison or pricing-adjacent question means buyers meet the rival at the moment of shortlisting, and that gap goes first. If several questions are affected, fix the late-stage ones before the early-stage ones, regardless of which is easiest.

Equally important is what not to do. Do not rewrite your entire positioning in response to one rival's presence. Do not disparage the vendor - to buyers, or in your content; if they are winning answers, they are legitimate in the buyer's eyes and disparagement reads as fear. And do not launch a broad content programme when the diagnosis names three specific pages; volume is not the remedy for a targeted gap.

Step 4: re-measure on the same questions

A fix you have not re-measured is a hope. After shipping the changes, re-run the same questions on the same surfaces - the comparison only means something if the questions are held constant. Allow a realistic interval before judging: answer engines and indexes take time to reflect new and changed content, typically weeks rather than days, and the exact lag is not knowable in advance.

This is where a locked baseline earns its keep. Magrios handles the pattern by keeping the question set fixed and re-scanning against it, so the before-and-after is genuinely like-for-like; a spreadsheet with dated runs of the same questions achieves the same discipline manually. Either way, record three things: whether your presence improved on the target questions, whether the rival's presence changed, and whether anything moved on the questions you did not touch - your control group.

If the gap closes, fold the vendor into your routine tracking and stand down the incident. If it does not, the diagnosis was wrong or incomplete: go back to step 2 with the new evidence, and look harder at surfaces you may have discounted.

Keep the playbook, lose the alarm

The first unknown-vendor incident feels like an emergency; by the third it should feel like weather. Ending each incident with a short written record - question, diagnosis, fix, result - builds the pattern library that makes the next one faster, and re-running discovery on a regular cadence means the next unfamiliar name appears in your scan before it appears in a prospect's screenshot.