Where do TzamunAI models run?

TzamunAI models can run on locally hosted infrastructure in Saudi Arabia or on your own on-premise servers. The deployment choice should be agreed with your team based on your infrastructure and business requirements.

How do TzamunERP and TzamunAI work together securely?

TzamunERP connects to TzamunAI through Tzamun’s MCP integration. The AI agent operates within the ERP user’s permissions and cannot perform tasks beyond that user’s authority. Actions that require approval remain subject to approval. TzamunAI models can run on Saudi cloud infrastructure or on-premise servers.

Where is my data stored and who can access it?

Your data stays in the Kingdom — either fully on-premises within your own infrastructure, or in our Saudi-based cloud. With on-premises deployment, your data never leaves your network. We enforce end-to-end encryption, role-based access control, and full audit trails.

What do we need to run ILM on-premises?

ILM can be deployed entirely within your own data center, even fully air-gapped. Our team assesses your hardware and sizes the GPU servers to your workload. We handle installation, integration, and training — and ILM:Coder keeps the environment self-healing afterward.

How good is ILM's Arabic compared to global models?

Arabic is a first-class language for ILM, not an afterthought. ILM:Flash is a dedicated Arabic model with native fluency and Gulf dialect understanding, trained on the Saudi work environment — including local terminology, regulations, and business etiquette.

Can ILM integrate with our existing ERP and systems?

Yes. ILM connects to any enterprise system through our REST API, webhooks, and SDK, and comes pre-integrated with TzamunERP. ILM:Coder also handles custom connectors for legacy or in-house applications.

How long does deployment take?

Cloud deployment can be live almost instantly. On-premises timelines depend on your infrastructure and integration scope — typically a few weeks from assessment to production, with our team supporting every step.

What kind of support do you provide?

A Saudi-based support team that understands local business needs, available around the clock. You get a dedicated point of contact, plus ongoing model updates and optimization.

First Saudi Enterprise AI Solution

Enterprise AI Agents Built for Saudi Arabia

TzamunAI delivers intelligent AI agents that seamlessly integrate with any enterprise system. Choose between on-premises hosting or cloud deployment on Saudi servers.

TzamunAI · Smart PipelineDemo

4 ILM models · One pipeline · One answer

Specialized ILM Models

99.9%

Uptime SLA

KSA

Saudi-Based Servers

From capability to application

See ILM in your everyday work

Choose a use case to explore an illustrative example in the chat preview.

Our ILM Models

Meet the ILM Family

A family of specialized ILM models, each trained on Saudi Arabia's work environment and regulations, working together behind one smart pipeline to transform your business operations.

Why the name "ILM"?

عِلْم

Knowledge — in Arabic

Intelligent Language Model

What the letters stand for

ILM carries a dual meaning: in Arabic, ʿIlm (عِلْم) means Knowledge — and it also stands for Intelligent Language Model. One name, two promises.

ILM:Pro

Our 70B-parameter flagship for general intelligence and agentic work. Plans, reasons, and executes multi-step tasks across your enterprise systems.

Agentic WorkflowsAdvanced ReasoningMulti-Step ExecutionEnterprise Knowledge

70B

Parameters

24/7

Available

10x

Productivity

Click to see example

ILM:Pro

Example

Prepare our Q2 board pack: pull sales figures from the ERP, compare against targets, flag the top risks, and draft an executive summary for Sunday's leadership meeting.

✅ **Board pack ready — here's the workflow I executed:** 🔄 **Step 1: Data Collection (ERP)** Pulled Q2 actuals: Revenue SAR 12.4M (+8.2% QoQ), Gross margin 34%, DSO 52 days. 📊 **Step 2: Target Comparison** • Revenue: 12.4M vs 11.8M target → +5.1% ✅ • New clients: 14 vs 20 target → -30% ⚠️ • Gross margin: 34% vs 36% target → -2pts ⚠️ ⚠️ **Step 3: Top Risks Flagged** 1. Client concentration — top 3 clients = 47% of revenue 2. DSO rising from 41 → 52 days (cash-flow pressure) 3. New client acquisition 30% below plan 📝 **Step 4: Executive Summary Drafted** 2-page summary saved to /Board/Q2-Pack. Calendar invite sent for Sunday 10:00 AM with the pack attached. Want me to add a sensitivity analysis on the Q3 forecast?

Click to go back

ILM:Coder

A 32B-parameter specialist in coding and tool calling — and a self-healer for your IT environment that detects, diagnoses, and fixes issues automatically.

Code GenerationTool CallingSelf-Healing ITDevOps Automation

32B

Parameters

Auto

Self-Healing

100+

Tools

Click to see example

ILM:Coder

Example

Our invoicing API on PRD-02 has been throwing 502 errors since 02:00 AM and the on-call engineer is unreachable. Diagnose and fix it.

🔧 **Issue resolved — service restored at 02:47 AM. Self-healing report:** 🔍 **Diagnosis** • Checked nginx logs: upstream timeouts to invoicing-svc:8080 • Container memory at 98% — OOM killer terminated 3 workers • Root cause: memory leak in PDF generation introduced in v2.4.1 ⚡ **Automatic Remediation (executed)** 1. Rolling restart of invoicing-svc (zero downtime) 2. Raised container memory limit 2G → 3G as temporary mitigation 3. Routed traffic to healthy replicas during restart 🩹 **Permanent Fix (proposed)** invoice_pdf.py — close the canvas after rendering: - pdf = canvas.Canvas(buffer) + with closing(canvas.Canvas(buffer)) as pdf: 📈 **Prevention** • Added memory alert at 80% threshold • Scheduled a leak regression test in CI Shall I open the pull request and notify the on-call channel?

Click to go back

ILM:Vision

The multimodal model that sees, speaks, and creates. Understands images and documents, generates natural voice with XTTS, and creates images with NAMAA.

Image UnderstandingXTTS VoiceNAMAA Image GenerationDocument Analysis

3-in-1

Multimodal

XTTS

Voice

NAMAA

Images

Click to see example

ILM:Vision

Example

This morning's shipment arrived damaged — here's the photo. Document the damage, generate a claim-ready report image, and record an Arabic voice summary for the warehouse team.

👁️ **Image Analysis Complete** **Damage detected in photo:** • 4 of 12 cartons show crush damage (left pallet side) • Water staining on bottom row — likely transit exposure • Shrink wrap torn; pallet label intact: PO-88231, Supplier: Gulf Pack Co. 🖼️ **NAMAA — Claim Report Image Generated** Created an annotated damage-report image with bounding boxes on each damaged carton, severity color-coding, and your company letterhead. Saved to /Claims/PO-88231/. 🔊 **XTTS — Arabic Voice Note Recorded** 38-second Arabic summary generated for the warehouse team: "وصلت شحنة أمر الشراء 88231 صباح اليوم متضررة..." 📋 **Next Steps** • Insurance claim draft prepared (est. value SAR 8,400) • Supplier notification email ready to send Send the claim and notify Gulf Pack Co.?

Click to go back

ILM:Flash

A fast 10B-parameter Arabic model specialized in general tasks. Native Arabic fluency with Gulf dialect understanding at lightning speed.

Native ArabicLightning FastGulf DialectsEveryday Tasks

10B

Parameters

<1s

Response

Arabic-First

Click to see example

ILM:Flash

Example

Draft a formal Arabic reply to a customer whose order is 3 days late — apologize, explain the delay, and offer a 10% discount on their next purchase.

⚡ Generated in 0.4 seconds: عميلنا العزيز، نشكر لكم تواصلكم، ونعتذر بصدق عن التأخير في تسليم طلبكم لمدة ثلاثة أيام، والذي نتج عن ضغط استثنائي في عمليات الشحن. تقديراً لثقتكم الغالية، يسعدنا أن نقدم لكم خصماً بنسبة 10% على طلبكم القادم، وقد تم تفعيله تلقائياً على حسابكم. مع خالص التحية والتقدير، فريق خدمة العملاء --- 💡 **Why ILM:Flash?** Native Arabic fluency • Formal business tone • <1s response time Want a friendlier Gulf-dialect version or an English translation?

Click to go back

One Smart Pipeline. Many Minds. One Answer.

All ILM models run behind a single smart pipeline. Every request is routed automatically and may be processed by one model — or several working together — yet the response reaches you as if a single agent did all the work.

Illustrative example"Analyze this contract photo and send me an Arabic voice summary"

Your Request

Smart Pipeline

One Unified Answer

Request received and routed by the smart pipeline

Flexible Deployment

Choose Your Deployment Model

Deploy AI agents your way. Whether you need complete data sovereignty with on-premises hosting or the flexibility of our Saudi-based cloud.

On-Premises Hosting

Complete control and data sovereignty. Deploy TzamunAI within your own infrastructure for maximum security and compliance.

Complete Data Sovereignty

Air-Gapped Security

Custom Integration

Full IT Control

Ideal for: Large Companies, Enterprises, Finance, Healthcare

Popular

Cloud Hosting in Saudi Arabia

Leverage our secure cloud infrastructure hosted entirely within Saudi Arabia. Fast, scalable, and compliant with local regulations.

Saudi-Based Servers

Auto-Scaling

99.9% Uptime SLA

Instant Deployment

Ideal for: SMEs, Startups, Growing Enterprises

Why TzamunAI

Built for Saudi Enterprise Success

The first AI solution in Saudi Arabia specifically designed for enterprise companies, with deep understanding of local business culture and regulations.

Saudi-First Design

Trained on Saudi work environment, regulations, and business culture. Full Arabic language support with Gulf dialect understanding.

Universal Integration

Seamlessly connects with any enterprise system, ERP, CRM, or custom application through our comprehensive API and SDK.

Enterprise Security

Bank-grade security with end-to-end encryption, role-based access control, and full audit trails for compliance.

Regulatory Compliance

Pre-configured for Saudi regulations including ZATCA, labor laws, and data protection requirements.

Infinitely Scalable

From single department to enterprise-wide deployment. Scale your AI workforce as your business grows.

Local Support

Saudi-based support team with deep understanding of local business needs and technical requirements.

Why ILM

ILM vs. Generic Global AI

Global AI models weren't built for the Saudi enterprise. ILM is — from data sovereignty to native Arabic and local regulations.

TzamunAI · ILM

Generic Global LLMs

On-premises & full data sovereignty

Servers hosted inside Saudi Arabia

Native Arabic & Gulf dialect mastery

ZATCA, Saudization & local regulations built-in

Self-healing IT automation

Arabic voice (XTTS) & image generation (NAMAA)

Unified multi-model smart pipeline

Local Saudi support team

Included Limited Not available

Saudi Vision 2030

Built to Power Vision 2030

TzamunAI is aligned with the Kingdom's Vision 2030 — accelerating digital transformation with sovereign, locally-built AI.

Local Content & Talent

A Saudi-built AI platform that grows local technical capability and reduces dependence on foreign technology.

Data Sovereignty

Your data stays in the Kingdom — on your servers or our Saudi-based cloud, fully under local jurisdiction.

Digital Transformation

Automating enterprise operations to boost productivity and efficiency across both public and private sectors.

Seamless Integration

Works With Your Existing Systems

TzamunAI agents integrate with any application or come pre-integrated with TzamunERP for a complete enterprise solution.

TzamunERP Integration

Get the complete enterprise experience with TzamunERP. All AI agents pre-configured and ready to transform your business operations.

Visit TzamunERP →

Custom Integration

Integrate with any existing system through our REST API, webhooks, and comprehensive SDK. Support for all major enterprise platforms.

REST API

Webhooks

SDK Support

Custom Connectors

FAQ

Frequently Asked Questions

Everything you need to know about deploying ILM in your enterprise. Still have questions? Reach us on WhatsApp.

Ready to Transform Your Enterprise?

Join leading Saudi enterprises already using TzamunAI to drive innovation and efficiency.

Get in Touch

Contact Us

Have questions? Reach us instantly on WhatsApp or send us an email — we usually reply within minutes.

Chat with us on WhatsApp

The fastest way to reach our team. Ask about ILM models, deployment options, or pricing — we usually reply within minutes.

+966 54 964 7333

info@tzamun.ai

Phone

+966 9200 23452

Location

Saudi Arabia

Need immediate assistance?

Our support team is available 24/7 to help you with any questions or concerns.

Enterprise AI Agents Built for Saudi Arabia

See ILM in your everyday work

Reporting & sales

Document processing

IT operations

Customer support

Meet the ILM Family

Why the name "ILM"?

ILM:Pro

ILM:Coder

ILM:Vision

ILM:Flash

One Smart Pipeline. Many Minds. One Answer.

Choose Your Deployment Model

On-Premises Hosting

Cloud Hosting in Saudi Arabia

Built for Saudi Enterprise Success

Saudi-First Design

Universal Integration

Enterprise Security

Regulatory Compliance

Infinitely Scalable

Local Support

ILM vs. Generic Global AI

Built to Power Vision 2030

Local Content & Talent

Data Sovereignty

Digital Transformation

Works With Your Existing Systems

TzamunERP Integration

Custom Integration

Frequently Asked Questions

Ready to Transform Your Enterprise?

Contact Us

Chat with us on WhatsApp

Need immediate assistance?