Wise Pirates' proprietary engine for Generative Engine Optimization.
Own how AI answers about your brand.
WAVE measures how ChatGPT, Gemini, Claude and Perplexity see, cite and recommend your brand, then improves it, the way SEO once did for search. A proprietary Wise Pirates engine, grounded in our own GEO research and built to be compliant and non-adversarial.
Search is becoming an answer.
People are moving their questions from search boxes to AI assistants. In that world, winning is no longer ranking on a page of links. It is being the brand the model cites, trusts and recommends inside the answer itself.
Share of Model
The old metric was a ranking on a page. The new one is whether the model names you at all, how prominently, and how favourably, when a buyer asks.
Most brands have never measured it, and cannot see why a competitor gets recommended and they do not. WAVE makes that visible, and moveable.
Figures are industry estimates from market analysts and public research, offered as context. They describe the market, not WAVE's results.
WAVE is our engine for Generative Engine Optimization.
Wise AI Visibility Engine. It does two things most tools do not do together: it measures your visibility inside the models with rigour, and it moves it, by optimizing the content the models learn from. Proprietary to Wise Pirates, and grounded in our own GEO research.
A proprietary engine, not a dashboard
WAVE is our own system, grounded in our own GEO framework. It is not a reseller's citation tracker with a new label on it.
It optimizes, not just observes
Most tools tell you what already happened. WAVE finds what the models reward and rewrites your content toward it, before the next answer is generated.
Authority, not manipulation
No prompt injection, no hidden triggers. WAVE strengthens the structure, semantics and authority of your public content so models represent you accurately.
In one line: WAVE tells you what AI says about you, why, and what to change so it says something better. It is the AI-answer layer alongside our SEO & GEO and AEO & AI Mentions work.
Five families, and "visibility" stops being a guess.
WAVE does not report one vague score. It reads five families of metrics per brand, product or service, so you can see exactly where you stand against competitors and which lever to pull.
| Brand | Share of Model | Awareness | Sentiment | Authority | Differentiation |
|---|---|---|---|---|---|
| You | 0.34 | 0.71 | 0.78 | 0.42 | 0.55 |
| Competitor A | 0.58 | 0.80 | 0.74 | 0.76 | 0.72 |
| Competitor B | 0.41 | 0.63 | 0.66 | 0.60 | 0.58 |
| Category avg | 0.44 | 0.71 | 0.73 | 0.59 | 0.62 |
Share of Model
How often, and how prominently, a model names you versus competitors when a buyer asks.
Awareness
Whether the model actually knows you, spontaneously and accurately, without being led.
Sentiment
The tone the model uses about you: positive, neutral or negative.
Authority
Whether the model treats you as a trusted, cited source, and which sources it leans on.
Differentiation
Whether your unique value comes through, or you blur into the category.
A loop in four phases.
WAVE is a continuous feedback loop, not a one-off report. Each phase answers a plain question, and the output of one feeds the next.
Discovery
We map your public footprint the way models ingest it, so the audit is grounded in what the engines can actually read about you.
What can the models see?Intent modeling
We simulate the real prompts buyers ask, by market and language, instead of a few rigid keywords.
How will people ask?GEO audit
We run those prompts across the engines, store every answer over time, and score the five metric families.
What are they saying?Optimization
We turn the gaps into content changes on two tracks, then re-run the audit to prove the movement.
How do we improve?Store first, then score.
The audit is the diagnostic core. It submits your simulated prompts to a spectrum of engines, keeps a dated record of every response, and scores them. Because it stores before it scores, you can watch how the models drift, not just where they sit today.
An urgent snapshot on a selected panel of models. Low cost, quick to read.
A balanced panel of models and prompt samples. The everyday depth.
An integral read across every supported engine, for the moments that matter.
Model knowledge changes without notice. A dated history of answers lets you catch a drop the week it happens and separate a real shift in the market from a one-off in a single model. It also turns optimization from reactive to proactive: you fix before the next release re-reads your content.
Real intent, not rigid keywords.
People do not ask an AI the same question twice. WAVE translates a marketing goal into a testable search goal, then generates many realistic variants of it, so the read reflects how buyers actually talk.
From one search goal, WAVE fans out many prompt variants across engines, so a single answer never decides your score. The more entry points we test, the more solid the audit floor.
Two tracks, because not all models are the same kind of problem.
The audit finds the gap. Optimization closes it, and here WAVE splits in two, because a closed model and an open model can be improved in very different ways.
Learn what the model prefers
For proprietary engines like ChatGPT and Gemini, the internals are hidden. So WAVE works empirically: it correlates what the models reward with your goals, infers the patterns behind a strong answer, and stores those preferences to reformulate your content toward them.
Optimize at the weight level
For open models like Llama and Gemma, we have the weights. That lets WAVE analyse the model's internals directly and shape content to align with how it actually attends to language, a far more precise, deterministic optimization than approximating from the outside.
Each read re-runs the same prompts, so the movement is measured, not asserted. Illustrative pattern; real gains depend on your starting point, category and how contested the prompts are.
Most of the market only approximates from the outside. Having the open-source track is what lets WAVE be exact where exactness is possible, and honest about where it is inference. Either way, the output is a content change your team approves and publishes.
The foundations under the engine.
You do not need this to read your audit. But if you want to know why the numbers are defensible, here is what sits underneath.
01 Built on real GEO research
WAVE formalizes our own research on Generative Engine Optimization, in the lineage of the field's founding academic work and its benchmarks. The five metric families and the four-phase loop come from that framework, not from a marketing deck.
02 An artefact and a search goal, defined
Every piece of your public content becomes an artefact the engine can reason over, and every test is a structured search goal, so the pipeline is reproducible rather than ad hoc.
A = { id, source_url, content_url, upload_time }
S = { goal_type, base_prompt, target_response, context, language, market }
03 Store-then-score, over time
Responses are captured and serialized with timestamps t0, t1, ... so analysis is longitudinal. That is what makes drift visible and optimization proactive instead of reactive.
04 Non-adversarial by design
The field has adversarial techniques that manipulate models with hidden triggers. WAVE deliberately does not use them. It improves genuine authority and structure, which is the only approach that survives the next model update and a client's own compliance review.
Where a model is open, the analysis is exact. Where it is closed, WAVE says so and reasons from evidence. Explainability is the point: a read you can defend is a read a serious team will act on.
If a buyer could ask an AI about you, this is for you.
Anywhere discovery is shifting into AI answers, being invisible or mis-described is a cost. Three profiles feel it first.
Consideration-driven
When customers research you through an assistant before they ever reach your site, the AI's answer is your new first impression. WAVE makes it accurate and favourable.
Recommendation-driven
Assistants increasingly compare and recommend products directly. Being the one that gets named, for the right reasons, is the whole game.
Managing many brands
One consistent way to measure and improve AI visibility across brands, markets and languages, instead of a different opinion in every room.
Start with a snapshot. Stay for the movement.
WAVE is a capability we run with you, not a platform to administer. Most brands start with the audit and grow into tracking and optimization as AI answers become a bigger share of how they are found.
GEO Audit
A dated snapshot of what ChatGPT, Gemini, Claude and Perplexity say about you today, versus your competitors, across the five metric families.
Format: a one-off diagnosticWAVE Tracking
Continuous monitoring of your Share of Model, sized to the number of prompts and products you track and how often you refresh.
Format: a monthly subscriptionContent Optimization
We execute the changes the audit surfaces, on both tracks, and prove the movement in the next read.
Format: an optimization retainerWAVE is a Wise Pirates proprietary capability, part of the same proprietary technology family as Wandrome and Wise Horizon. Run for clients and partners.
Improve your standing, honestly.
Non-adversarial
No prompt injection, no hidden triggers, no gaming. WAVE builds genuine authority and structure, the only kind of visibility that lasts past the next model update.
Explainable
Every score traces back to the dated responses behind it, and every recommendation names the metric and the reason. You can argue with it, which is the point.
Public content only
WAVE reads public-facing assets and simulates public prompts. It needs no access to your ad accounts and carries no end-customer personal data.
Human in the loop
WAVE recommends; your team decides what to publish. The optimization is a content change you own, not an automated action on your site.
The questions we hear most.
Last updated: September 2026. Illustrative figures on this page are examples, and market figures are industry estimates, not WAVE results.
What is WAVE?
What is Generative Engine Optimization (GEO)?
How is WAVE different from an SEO or AI-tracking tool?
Which AI engines does WAVE cover?
What does a WAVE audit measure?
Is WAVE compliant, or does it manipulate the models?
How fast is an audit, and how deep can it go?
How do we start with WAVE?
Find out what AI says about you, then change it.
Start with a GEO audit: a dated read of your Share of Model across the engines that are quietly replacing search. No guesswork, no manipulation, just what the models say and what to do about it.
Get your GEO audit →
