The signal, minus the AI noise.
Three threads we actually follow: the Gulf, the models, and what awarding bodies are deciding about both.
GCC news30 June 2026
In-region hosting arrives for Gulf centres.
Learner data staying inside the Gulf is no longer a roadmap item. What sovereign deployment means in practice for centres marking Arabic and English cohorts.
Read the thread →
AI, education & qualifications2 July 2026
The five questions regulators now ask about AI marking.
Determinism, evidence, human oversight, data boundaries and change control — the due-diligence pattern hardening across awarding bodies.
Read the thread →
GCC news.
Deployments, partnerships and regulation across the Gulf — where in-region hosting and Arabic-capable extraction matter most.
30 June 2026
In-region hosting arrives for Gulf centres.
Learner data staying inside the Gulf is no longer a roadmap item. What sovereign deployment means in practice for centres marking Arabic and English cohorts.
12 June 2026
Marking Arabic submissions without losing the evidence trail.
Verbatim quoting is harder when the quote is right-to-left inside a left-to-right marksheet. Notes from making the evidence chain hold in both scripts.
20 May 2026
The GCC qualifications landscape, briefly mapped.
Which professional bodies are growing fastest in the Gulf, and why distributed assessment makes identity assurance the region's first question.
AI news.
What's changing in machine learning that actually affects regulated assessment — separated from what's just noise.
25 June 2026
Why 'temperature zero' still isn't determinism.
Sampling settings are not an architecture. What it actually takes to re-mark a script months later and get the same grade, byte for byte.
3 June 2026
Dozens of scripts, not thousands.
The industry assumption is that useful marking models need enormous labelled corpora. Schema-driven extraction changes the arithmetic.
8 May 2026
LLM judges are easy. Auditable judges are not.
A judging model is a weekend project; a judging system an EQA can interrogate is the actual work. Where the gap lives.
AI, education & qualifications.
Awarding bodies, centres and regulators working out what AI marking should be allowed to do — and what it must prove first.
2 July 2026
The five questions regulators now ask about AI marking.
Determinism, evidence, human oversight, data boundaries and change control — the due-diligence pattern hardening across awarding bodies.
18 June 2026
Assessors aren't the bottleneck. Re-reading is.
Centres that keep assessors reviewing evidence — instead of reading scripts cold — report the same oversight in a fraction of the hours.
27 May 2026
Resubmissions deserve first-class marking too.
Referred learners are the highest-stakes cohort in the building. Why attempt-keyed records and additions-checking matter more the second time round.
Get the quarterly note.
One email a quarter: what changed in AI marking, what regulators asked, and what we shipped. No drip campaigns.