Skip to content
Get Ranked On TopIllustrative case studyContent Engine plan

A fictional business. The figures are a modelled scenario, not a client result.

Wardan Marble Worksوردان للرخام

Marble and granite plant, 220 staff · Cairo

How a Cairo marble plant became the source AI assistants quote, in Arabic and English

+140% organic export enquiries

18 months · Middle East & Africa · Content Engine plan · Topic clusters · Answer-ready passages · Arabic–English pages · Spec library & schema

One block, sawn in sequence; every second slab turned over like a page.
Organic export enquiries / month
35 → 84
AI answers citing Wardan (of 120 test prompts)
3 → 64
Arabic–English page pairs
106
Organic clicks / month
2,300 → 11,600

IThe challenge

The plant knew the stone. Someone else gave the answers.

Architects asking AI assistants about Egyptian marble got distributors' answers: the plant's know-how sat in English brochure PDFs, with no Arabic.

Wardan quarries beige marble on the Galala plateau and in Sinai and cuts it in Cairo for buyers in the Gulf, Italy and East Asia. Twenty years of test data, finishing know-how and fixing advice lived in 38 English brochure PDFs. Search engines indexed them poorly and AI assistants barely at all, and there was not one page in Arabic, the language half the region's architects and contractors search in. So when a buyer asked an assistant to compare two Egyptian marbles, it quoted a distributor's blog instead of the plant that cuts the stone.

38

brochure PDFs held the plant's technical knowledge

0

pages in Arabic

Who the assistants cited · 120 buyer prompts, before

IIDiagnosis

Core samples from the old site.

Locked in PDFs

92%

38 brochure PDFs held 92% of the technical content: water absorption, flexural strength, finishes, slab sizes. Most were scanned images, so neither search engines nor assistants could read the numbers inside.

Sample 1 of 4

The Arabic gap

46%

Not one page in Arabic. Yet in our keyword model, about 46% of regional demand for Egyptian marble and granite was searched in Arabic, and almost all of it went to importers.

Sample 2 of 4

AI answers

3 / 120

We put 120 real buyer questions to four AI assistants. Wardan was cited in 3 answers; distributors in 71, often quoting figures that came from Wardan's own brochures.

Sample 3 of 4

Thin stone pages

60 words

Fourteen stone pages averaged 60 words each: a photo, a name, a contact form. No comparisons, no applications, nothing an assistant could quote.

Sample 4 of 4

Where the technical knowledge lived

Before

Month 18

  • PDF only
  • English HTML only
  • Arabic + English HTML

IIIThe strategy

Cut the block along the grain of the question.

Built on our Content Engine plan: content and AI-search work for businesses whose expertise is real but invisible to the engines that now answer buyers.

  1. I

    Topic clusters

    Nine clusters mapped from the questions buyers actually ask: stone against stone, use by application, finishes, sizes and logistics. Each has a pillar page and 8–14 narrower pages that link up to it and across to each other.

  2. II

    Answer-ready passages

    Every page opens with a 40–60 word passage that answers the question in its first sentence, names the stone and gives the number. Those are the passages AI overviews lift and cite.

  3. III

    Arabic–English pages

    Written in pairs by Cairo-based technical writers, never machine-translated: Arabic for the Gulf and the Egyptian market, English for Europe and Asia, linked with hreflang so each buyer lands in their own language.

  4. IV

    Spec library & schema

    The 38 PDFs rebuilt as 140 HTML spec sheets: each test result on its own row with its standard and unit, marked up with Product schema; the PDFs stay as downloads.

IVAnswer slab

Pick a buyer question. Watch one block become four answers.

Each question is a block of raw knowledge. We cut it into a pillar page, a comparison table, a spec sheet and an FAQ, every one in Arabic and English, and each opens with a passage an AI overview can quote whole.

Buyer questions

Cut from this block

14 page pairs in this cluster

English page

Egyptian beige marble: Galala, Sinai Pearl and how to choose

Two stones from one country, quarried 300 km apart: colour, density, finishes and where each performs best.

wardanmarble.example/en/egyptian-beige-marble

Arabic page

الرخام المصري البيج: جلالة وسيناء بيرل وكيف تختار بينهما

حجران من بلد واحد، تفصل بين محجريهما 300 كم: اللون والكثافة والتشطيبات وأين يتفوق كل منهما.

wardanmarble.example/ar/egyptian-beige-marble

Why this piece existsThe hub of the topic cluster: answers the broad question in full and links down to every narrower page.

What an AI assistant answers

Before

The assistant summarised a distributor's blog post and suggested buying through it. Wardan, which quarries both stones, wasn't mentioned.

Cited source

stonetrade-hub.example

Cited Wardan in 1 of 30 test prompts on this topic

After · Quoted passage

Galala and Sinai Pearl are both Egyptian beige marbles, but they behave differently. Galala is denser and more even, with water absorption around 0.6%, which makes it the safer choice for large bookmatched walls and busy floors. Sinai Pearl is creamier, with fossil flecks and absorption near 1%; it is best honed and kept dry.

Cited source

wardanmarble.example/en/egyptian-beige-marble

Cited Wardan in 19 of 30 test prompts on this topic

VMonth by month

18 months, slab by slab.

  1. Months 1–2

    Audit

    Text extracted from 38 PDFs, a 120-prompt AI baseline, an Arabic and English keyword model for six export markets.

  2. Months 3–4

    Cluster map

    Nine topic clusters, 106 page pairs planned, a passage template and a style guide for technical writing in Arabic.

  3. Months 5–7

    Spec library

    140 spec sheets and 22 comparison tables rebuilt in HTML from the lab reports, with Product schema and downloadable PDFs.

  4. Months 8–13

    Bilingual build

    Pillars first, then cluster pages, about 12 pairs a month, each one checked by the plant's quality engineer.

  5. Months 14–18

    Cite and refresh

    The 120 prompts re-run every quarter; passages that weren't quoted rewritten; new pages for questions buyers started asking.

VIResults

Now the plant answers for its own stone.

Organic export enquiries / month

35 → 84

+140%

AI answers citing Wardan (of 120 test prompts)

3 → 64

Measured each quarter in four assistants

Arabic–English page pairs

106

212 pages written in pairs, never machine-translated

Organic clicks / month

2,300 → 11,600

Organic clicks a month, all languages

Modelled scenario for a fictional business — not a client result. Rankings and traffic are never guaranteed; see our Terms.

Organic export enquiries per month35 → 84 a month
0306090Bilingual pages live3584M1M6M12M18
AI answers citing Wardan · of 120 test prompts
04080120distributors, before: 7134M39M621M938M1252M1564M18

Sample rankings, as searched

QueryBeforeMonth 18
galala marble supplier#18#3
sinai pearl marble#24#4
egyptian marble manufacturer#31#6
bookmatched marble slabs egypt#47#5
رخام جلالةnot in top 100#2
أفضل رخام مصري للأرضياتnot in top 100#1

Before the work, Arabic queries had no Wardan page to rank.

Get a plan like this

Expertise buried in PDFs, or customers searching in a language your site doesn't speak? Start with the Content Engine plan, or tell us what you make.

← All case studies