AI visibility work: the screenshots and the numbers

This is where NeuralGen shows the work: before-and-after screenshots of real AI answers, and the share-of-voice numbers that moved. We anonymise clients but never the method — you'll see exactly what we changed and what happened across ChatGPT, Perplexity, Claude and Google AI. We launched with two pilot engagements and add new studies as results mature. If you want the same picture for your own brand, the free AI Visibility Scorecard is the fastest way to see where you stand today.

How to read these case studies

Every result here is measured the same way we measure yours: share of voice — presence, position and sentiment — across five AI engines, tracked from a dated baseline. Screenshots are dated, and clients are anonymised by sector rather than named, at their request.

Pilot 1 — Professional-services firm

The situation

The client had no presence in AI-generated answers for its core buying questions, while two named competitors appeared consistently.

What we changed

Entity and schema fixes, a citation-source push into sector-relevant directories and press, and monthly measurement across five engines.

The result

Pilot 2 — [Sector]

The situation

What we changed

The result

Want this for your brand?

Get your free AI Visibility Scorecard

— or read how the full service works on ai-seo-agency.

Frequently asked questions

Why are your case studies anonymised?

AI visibility is competitive, so most clients prefer we don't name them while work is live. We anonymise the brand by sector but keep the method and the numbers fully visible — you see exactly what changed and how share of voice moved.

What results can I realistically expect?

It varies by market and starting point, which is why we measure first. Typically brands see movement in AI answers within 6–12 weeks. We don't publish guaranteed figures because no honest provider can — the case studies show real ranges, not promises.

How do you prove a result?

With dated before-and-after screenshots of the actual AI answers, plus the share-of-voice score across five engines at each point in time. It's the same evidence you'd get in your own monthly report — verifiable, not a testimonial quote we wrote ourselves.