One decision to make in this room: the register the new site speaks in. Everything else is already settled — and independently peer-reviewed.
Four tiers, July 2026 — global consultancies, the Saudi peer set, GCC integrators, platform implementers. Each is built as a list of what they sell. None is built around the decision the buyer is trying to make.
Alliance-tied. Craft and research, but generic across sectors and never Arabic-native.
Each answers to an owner — a bank, a ministry, a telco — or is narrowed to one identity.
Execution-led, platform-aligned, English-first. Breadth without a promise.
Locked to one vendor. Every recommendation resolves to the platform they sell.
No site is structured around the buyer's decision. Judgment as the architecture is unclaimed.
Owned by a bank, a ministry, a telco — or locked to one platform. Genuine independence is unoccupied ground.
No one presents platforms as a governed delivery capability rather than a reseller badge.
And never native Arabic. One peer sets the Arabic bar; the globals fall short of it.
Only the product implementers run interactivity. An instrument that gives value before it asks is unclaimed.
None of it is sector-operational, Arabic-native, and structured to be cited by AI assistants.
Two independent teams derived these separately from the same sources and agreed. Not a menu — all five apply at once.
The register: typography, structure, and the voice of the page. What a senior buyer reads in the first second, before a word is processed.
Our verb spine — the device the strongest firms in this category each own. It carries the page in every register below.
Warm paper, large quiet serif, hairline rules. An imageless, type-led hero. Case studies read like a deal list.
The bravest hero in the market is the one with nothing to hide behind.
A visible grid, mono labels, the mark's angle as the only geometry. The six units drawn as a schematic.
The site itself is evidence of engineering discipline.
The working build. Serif display on a white canvas with a single accent. Proven strategy, but a widely-used look.
Already built — you will see it running.
Arabic-first, sand-tinted, architectural photography of real projects. Vision 2030 as the centrepiece.
A Saudi institution — not a translated one.
All four carry the same strategy. All four survive the logo redesign. They differ in what the buyer feels before they read.
Six independent AI systems from six different families — three of them running locally, offline, on our own hardware. Each received identical information, argued its position, then was shown every rival argument anonymously and forced to defend or concede. Two rounds. Disagreements were resolved on evidence, never on majority.
Each system independently ruled on register, meeting sequence, the readiness check, and the single biggest blind spot in the plan.
Every dissent was put back to the group. Two systems changed position — with reasons on record. One abandoned its own proposal.
Different model families fail differently. When independent systems converge, the agreement is real. When they split, the split is information.
Full record, including every dissent and the conditions that would reverse the verdict, is preserved in the project file.
It reads as a systems vendor — which is precisely the category we are leaving. That is a reversion, and it costs us the whole repositioning.
It risks sounding like every other consultancy. But that is a fixable execution problem, not a wrong position — and the grid fixes it.
No advisory firm in the market offers one. Only the platform sellers do — and theirs are lead-capture. The review split hard on this, and the split produced the rules.
The dissenting argument — that senior buyers want counsel, not quizzes — is correct the moment any line above is crossed. These are guardrails, not preferences.
A register is typography, structure, and voice. None of it depends on the mark. Waiting would freeze the content architecture and hold up every page behind a logo timeline — for no gain.
One reviewer proposed waiting. In round two, it withdrew its own proposal.
Arabic is an equal language of the grid — not an English layout mirrored afterwards. Two reviewers changed position on this. Our references are Western; without the discipline, Arabic becomes a translation and we lose the wedge hardest for anyone to copy.
Industry and business-unit pages carry real decision frameworks — criteria, trade-offs, the honest case against. If they collapse into service descriptions, the homepage is a veneer over a catalog.
Clients know the current mark. A register this different needs visual anchoring through the change — recognition is an asset we should not discard by accident.
Every comp and page survives a palette swap — a build rule from the first file, not a cleanup task later.
You will see the hero three ways next. Choose one. Our recommendation, unanimously peer-reviewed: the institution's voice on the engineered grid.
Where is the logo redesign? If a draft exists, the site's geometry should be derived from it — starting now, not later.
The homepage copy is written and under review. Confirm it, and the design locks around final words rather than drafts.
The industry and business-unit page template — the pages that must prove independence — is the next real piece of work. It matters more than the homepage now.
A full working homepage in one register, bilingual, with the readiness check running. It proves the strategy works. It is not the register decision.
No competitor in the Kingdom can write that sentence. The site exists to make it visible in the first second — before a word is read.