AI Search
& GEO.
Watch how the four answer engines actually describe your category. Real prompts, real APIs, verbatim answers. Every score in Ampli ladders up to a piece of evidence you can put in front of a sceptical client.
Real prompts. Real engines. No simulation.
You approve a prompt set, Ampli fires it against the live ChatGPT, Gemini, Claude and Perplexity APIs. We capture the verbatim answers, the prominence, the cited domains and the competitors named alongside you.
- Sentiment, prominence and recommendation strength prompt by prompt
- Verbatim AI answers stored as evidence behind every score
- Branded vs niche prompt cohorts tracked separately
- Diff every run vs the previous one, never lose a regression
| Prompt | Engine | Sentiment | Mention | Prominence | Cited | Competitors |
|---|---|---|---|---|---|---|
| Best cold brew coffee subscription for at-home brewing | ChatGPT · GPT-4o | positive | yes | top | yes | Trade, Atlas Coffee |
| How does Lumen Field Coffee compare to Trade Coffee | Perplexity · Sonar Pro | positive | yes | top | yes | Trade |
| Single origin coffee subscription with sustainable sourcing | Gemini · 2.5 Pro | neutral | yes | middle | — | Counter Culture, Blue Bottle, Atlas Coffee |
| Best monthly coffee club for espresso drinkers | Claude · Sonnet 4.5 | neutral | — | none | — | Trade, Blue Bottle, Stumptown |
| Where can I get freshly roasted whole bean coffee delivered | ChatGPT · GPT-4o | positive | yes | middle | yes | Trade, Atlas Coffee |
| Lumen Field Coffee subscription pricing | Gemini · 2.5 Pro | positive | yes | top | yes | — |
| Affordable specialty coffee delivery under $20 a month | Perplexity · Sonar Pro | neutral | — | none | — | Trade, Driftaway, Atlas Coffee |
| Coffee subscription with flexible skip and pause | Claude · Sonnet 4.5 | neutral | yes | middle | — | Trade, Mistobox |
reddit.com | sprudge.com | wirecutter.com | lumenfield.coffee | bonappetit.com | |
|---|---|---|---|---|---|
| Best cold brew subscription | |||||
| Sustainable single origin | |||||
| Espresso coffee club | |||||
| Whole bean delivery | |||||
| Affordable subscription | |||||
| Flexible skip & pause |
The new backlink graph.
When AI assistants answer, they cite specific sources. The citation set is the new search index. Ampli ranks the domains earning that credit for your category, calculates your share of voice and surfaces the gaps you can pitch.
- Cited domain leaderboard with deltas vs last audit
- Heatmap that shows which prompts each domain earns
- Reddit, YouTube and editorial coverage rolled up cleanly
- Outreach targets exported as a working list for the PR team
Map every prompt to the page that should win it.
Ampli scores how lift-ready each page is against the prompt that should claim it. Match under fifty? You have a structure problem. Match zero? The page does not exist yet — and you have your next brief.
| Prompt | Best matching page | Match | Structure | Schema |
|---|---|---|---|---|
| Best cold brew coffee subscription | /subscriptions/cold-brew | 88 | Comparison | Article + FAQ |
| How to brew cold brew at home | /brew-guides/cold-brew | 76 | How-to | HowTo |
| Single origin coffee subscription | /subscriptions/single-origin | 64 | Comparison | Article |
| Best coffee club for espresso drinkers | — gap — | gap | — | — |
| Sustainable coffee brands | /about/sourcing | 41 | Editorial | Article |
| Lumen Field Coffee subscription pricing | /pricing | 92 | Pricing table | Product |
Built by people who run audits, not by a model alone.
Every prompt is approved by a human. Ampli refuses to fire a prompt your team has not signed off on. No surprise spend, no garbage prompts.
Citations are normalised against the brand's own domains and tracked competitors before any leaderboard is computed. No double counting, no inflated share-of-voice.
Every score in the report carries an evidence-volume axis and a recency axis. You see at a glance what is rock solid, what is provisional, what just changed.
Each prompt is fired with model temperature pinned and prompts versioned. We capture the seed, the response and the evaluator. Re-run the same set in three months and compare cleanly.
Common questions about the AI search surface.
Generative Engine Optimisation is the practice of measuring and improving how AI assistants like ChatGPT, Gemini, Claude and Perplexity talk about your brand. As AI search replaces traditional click-through behavior, the brands that get cited are the brands that get chosen.
Pillar /02 — Content & Research.
Once you know which prompts you do not show up for, the next move is the content that fixes it. Read on.