Artificial Intelligence
From AI strategy to working systems, delivered in the Kingdom — We help you decide where AI pays off, then build, govern and run it on your data, with a team that stays from the first assessment to daily operation
Why Tarmeez
15+ years, 250+ projects
Delivery for government entities and large enterprises across the Kingdom
Teams that work on site
The people who advise you are the people who build and run what they recommend
Your data stays in the Kingdom
Private AI on your own premises or a Saudi cloud, never sent abroad
AI Strategy & Governance
Decide where AI pays off, set it up safely, and prepare your people.
AI Readiness Assessment
A scored view of readiness on the CST guide's five dimensions
AI Strategy & Roadmap
Prioritised use cases, a business case and a phased roadmap, with a working pilot
AI Office Setup
The operating model, roles, processes and tools for an AI office
AI Governance & Compliance
Risk levels, policies and approval steps, aligned with PDPL and the SDAIA ethics principles
AI Literacy & Training
Programmes for leaders, all staff and technical teams
Private AI & Custom Solutions
AI built on your data, hosted in the Kingdom, with no lock-in to one vendor.
Private AI Platforms
Language models on your own servers or a Saudi cloud
Knowledge Assistants
Answers from your documents, with sources shown
Model Customisation
Models trained on your own knowledge and terminology
Custom AI Applications
AI features built and integrated with existing systems
Workflow Automation
Automated processes on open tools you can own
AI Agents & Customer Experience
Agents that answer and act for your staff and your customers, on the channels they already use.
Employee Copilots
Internal assistants for support, HR, finance and IT requests
Chat & Messaging Agents
Sales and support agents on WhatsApp, Instagram, Messenger, web chat and e-commerce
Voice Agents
Inbound and outbound phone agents in Arabic and English, with handover to a person
Agent Integration
Connecting agents to CRM, ERP, ticketing and booking systems, with logs and controls
AI Engineering Enablement
Help your engineers ship faster with AI, safely and measurably.
AI Coding Rollout
Choosing, setting up and adopting AI coding tools across a team
Secure AI Development Lifecycle
Rules for review, testing, code ownership and data exposure
Developer Training & Productivity
Hands-on training, measured before and after
AI Security & Operations
Test AI systems before launch, keep them running, and keep their cost under control.
Vendor Independence Review
Architecture and contract review so models stay swappable
Managed AI Operations
Monitoring, updates, incident response and cost control
AI Red-Teaming
Testing your AI systems against attack and misuse before real attackers try, delivered with specialist partners
Pre-launch AI Assurance
A go-live check for agents, generative AI and AI-written code: quality, safety and compliance evidence
Ways to start
Each one is a defined piece of work with a clear end, so you can see results before committing to more.
AI Readiness Sprint
A fixed-scope assessment that tells you where you stand and where to begin
Strategy + Pilot in 90 Days
For government entities: a strategy and a working pilot within one quarter
Private AI Pilot
A knowledge assistant on your own documents, hosted in the Kingdom
AI Coding Rollout
AI coding tools set up and adopted across one engineering team
Agent Launch
One chat or voice agent, live on one channel
Embedded AI Team
A Tarmeez team on site, working alongside yours
How we work: forward-deployed teams
Our engineers work inside your organisation, alongside your people and on your systems. The same team stays from the first assessment to daily operation.
- 01
Assess
We measure readiness, map your data and systems, and find where AI can make a real difference
- 02
Prioritise
We rank use cases by value and feasibility and agree a business case for the first ones
- 03
Pilot
We build a working pilot on your data and measure it against the agreed goals
- 04
Govern and launch
We put policies, approvals and controls in place, then take the pilot into production
- 05
Run, improve and hand over
We operate and improve the system, and transfer the knowledge your team needs to own it
Responsible AI
Personal data
Every system is designed to comply with the Personal Data Protection Law (PDPL): what data is used, why, and who can see it.
Ethics
Our governance work follows the SDAIA AI ethics principles, with risk levels and approval steps set before anything goes live.
Security
Deployments are built against the NCA cybersecurity controls that apply to your organisation.
Adoption
Readiness and strategy work is structured around the CST AI adoption guide, so the results map onto what regulators expect.
Hosting
Your data stays in the Kingdom, on your own premises or a Saudi cloud. Where a model runs is agreed with you, in writing.
Technology
We are vendor-neutral: we pick the tools that suit your data, your hosting and your budget, and design so they can be swapped later.
Open-source language models
- Open-weight models
- Arabic-capable models
- Self-hosted inference
Knowledge retrieval
- Document ingestion
- Vector search
- Answers with sources
Workflow automation
- Open-source orchestration
- System integrations
- Human approval steps
Voice and messaging channels
- WhatsApp and social messaging
- Web chat
- Arabic and English voice
In-Kingdom cloud
- On-premises servers
- Saudi cloud regions
- Private networking
Delivery track record
All case studies-
Ministry of Energy Project: CCE
A knowledge hub for the national circular carbon economy programme.
-
Innovation Platform Services at King Abdulaziz City for Science and Technology
Over 200 services and process improvements.
-
Employee Training Platform (Gamification)
We connect the bank's systems and employees with their management.
Questions
Where is our data hosted?
In the Kingdom: on your own servers or a Saudi cloud region. Nothing is sent to a provider abroad unless you decide it should be.
Which models do you use?
Whichever suits the task, your hosting and your budget, often open-weight models you can run yourself. We design so a model can be swapped without rebuilding the system.
How long does a pilot take?
Usually a few weeks for a focused pilot, and up to 90 days when it comes with a full strategy.
What does our team need to provide?
A business owner for the use case, access to the relevant data and systems, and a few hours a week from the people who will use the result.
How is success measured?
Against goals agreed before the pilot starts, such as time saved, requests resolved or accuracy, measured before and after.