Computational linguistics · Kuwait · since 1998
We teach machines to read Arabic.
Most AI systems handle Modern Standard Arabic poorly and Gulf dialect not at all. We build the language models, analytics and digital services that work on the Arabic your citizens and customers actually write.
Off-the-shelf Arabic tools drop the two highlighted tokens — the words that carry the sentiment. That gap is the reason ministries and broadcasters come to us.
What we do
Five practices, one shared foundation
Everything we build rests on the same base: a research-grade understanding of how Arabic actually behaves — in formal documents, in dialect, and in the messy text people type online.
Arabic language processingالمعالجة الحاسوبية للغة العربية
Spell-checking and grammar APIs, morphological analysis, named-entity extraction, and sentiment scoring tuned for Modern Standard Arabic and Kuwaiti dialect. Deployable on-premise where data cannot leave the country.
Practice 02Generative AI & large language modelsالذكاء الاصطناعي التوليدي
LLM strategy, evaluation and deployment for Arabic-first organisations: retrieval over your own document archives, drafting assistants for legal and administrative teams, and governance frameworks that satisfy CITRA requirements.
Practice 03Data analytics & visualisationتحليل البيانات وعرضها
Text mining and audience intelligence at national scale, plus decision dashboards for executives and boards. Built with SAS, Klipfolio and Pulsar, extended with our own Arabic analysis layer.
Practice 04Digital transformation & e-learningالتحول الرقمي والتعليم عن بعد
Programme design and technical supervision for public-sector digital projects, from cloud migration to nationwide remote learning. Independent review of vendor deliverables when you need a second opinion.
Selected work
Where the systems went live
A sample of engagements delivered for Kuwaiti government entities, international vendors and universities.
Distance learning for a national school system
Moved a school system of 500,000+ students onto Microsoft Teams during the COVID-19 closure, training 109,000 teachers and staff. Documented by Microsoft as a published customer case study.
Reading public opinion in Kuwaiti dialect
Built an Arabic sentiment and content analysis layer on top of Pulsar so parliamentary staff could track how citizens were actually talking about legislation — not just how MSA tools guessed they were.
The national e-government portal
Contributed to the implementation of e.gov.kw and built the web platform used to submit and evaluate national IT projects, including entries for the Gulf E-Government Award.
Teaching a global platform to read Arabic
Worked directly with SAS R&D to extend SAS Text Miner’s Arabic processing, so an enterprise-standard analytics platform could handle regional text properly.
Independent review of the tax administration programme
Evaluated the technical outputs of the ITAS project and advised the steering committee on architecture and delivery risk, with written recommendations at each gate.
Digital transformation at a law school
Rebuilt core academic operations around digital workflows, including a computer-based aptitude examination system used for admissions.
The firm
Founded and led by a computational linguist
Information Age Consulting was built on a simple observation: the tools that let organisations understand their own text — search, analytics, automation, and now large language models — were designed for English and adapted for Arabic as an afterthought. In Kuwait, that afterthought is where the value is.
For more than two decades our founder has worked on both sides of that gap: teaching and researching Arabic language processing at Kuwait University, and advising the institutions that have to make it work in production. That combination is why our clients are typically the ones whose text cannot be sent abroad and whose decisions cannot be based on an approximate translation.
Public-sector advisory roles held to date:
Partners
Accredited to deliver, not just to resell
We hold formal partner status with the platforms we implement, and in several cases contribute engineering back to them.
Start here
Tell us what your text is doing wrong.
Send us a sample — a set of citizen comments, a document archive, a support inbox — and we will come back with a written read on what is achievable, what it would take, and whether you need us at all.