All case studies

Wishfy · GEO Engagement

LLM Citation Strategy — Getting Cited by ChatGPT, Claude, and Perplexity

How Wishfy engineered its own content and schema to earn citations from major AI engines within 30 days of launch.

2 engines in 30 days

LLM citations earned

15 across 4 languages

Tracked prompts

7 (Organization, Person, Service, FAQPage, Article, BreadcrumbList, ProfessionalService)

Schema types deployed

4 languages

llms.txt coverage

The challenge

Traditional SEO gets you into Google. But when a user asks ChatGPT "Who can help me with multilingual SEO in the Netherlands?" — Google rankings don't matter. The LLM decides who to cite based on entity authority, structured data clarity, and content that answers the question directly. Wishfy needed to be the answer.

The goal: earn citations from at least two major LLM engines (ChatGPT, Claude, Perplexity) within 30 days of launching wishfy.ai. Not by gaming the system — by building the kind of entity authority that LLMs reward.

Entity authority first

LLMs don't rank pages. They resolve entities. Before writing a single line of marketing copy, we defined Wishfy's entity graph: the Organization, the Person (founder Waseem Sabir), the Services offered, the geographic market served, and the relationships between them.

Every schema block on the site reinforces the same entity relationships. The Organization schema references the founder. The Person schema references the Organization. Service schemas reference their provider. This isn't decoration — it's how LLMs build confidence that an entity is real and authoritative.

Schema strategy

We deployed seven schema types across the site:

  • Organization — name, URL, founder, contact, service area
  • Person — Waseem Sabir with sameAs links to LinkedIn-profile and professional profiles
  • Service — one per service page, linked to the Organization
  • FAQPage — on every service detail page, giving LLMs ready-made Q&A pairs
  • Article — on blog posts and case studies, with author reference
  • BreadcrumbList — on every page, giving LLMs navigational context
  • ProfessionalService — on the homepage, tying the business to its category

Every schema block includes `inLanguage` matched to the page locale. LLMs use this to determine which language context an entity belongs to.

llms.txt as a first-class channel

Each language gets its own `/[lang]/llms.txt` endpoint — a plain-text file that describes Wishfy's services, team, and methodology in structured prose. Think of it as a robots.txt for AI: it tells crawlers exactly what the business does, who runs it, and what problems it solves.

The file is generated at build time from the same content collections that power the site. No manual sync, no drift.

Tracked prompts methodology

We defined 15 prompts across four languages — questions a potential client might ask an LLM. "Best SEO agency in the Netherlands." "Who does multilingual GEO?" "Astro website development company." We tracked these weekly across ChatGPT, Claude, and Perplexity, recording whether Wishfy was cited, mentioned, or absent.

This prompt tracking is not vanity — it's the measurement loop for GEO, the same way rank tracking is the measurement loop for SEO.

The outcome

Within 30 days of launch, Wishfy earned citations from two major LLM engines on tracked prompts. The combination of clean entity schema, llms.txt coverage, and direct-answer content made Wishfy resolvable as an entity — not just a website. That's the difference between being indexed and being cited.

Want results like these?

Let's talk about what we can build for you.

Contact us

Let's make your brand the obvious answer

Get a free audit. We'll show you how ChatGPT, Claude, Perplexity, and Google AI see your business — and where the wins are.