Skip to main content

Division / Artificial Intelligence

Production agents. Not deployed demos.

A specialized AI division. Bilingual agents and intelligent systems for KSA business — built with ZATCA and PDPL discipline from day one.

AR + ENBilingual native
ZATCA + PDPLReady
5 KSA EngagementsLive

AI that runs Monday morning. Not demos that crash by Friday.

TETRA AI · 2026

TETRA AI · live productionStreaming
>agent.intake.qualify
▶ Lead scored: 87 / 100
▶ Routed: Senior Sales
>zatca.invoice.validate(SAR_2400)
✓ Schema valid
✓ XML signed
>rag.support.answer("...PDPL?")
↓ Source: pdpl-policy.md:42
↓ Confidence: 0.98
>audit.log.write(actor=agent_1)
✓ committed
99.5% uptime<2s p950 PII leaks
Eleven projects · KSA + regional

Five active engagements. Six concepts.

Live client work shipped to production. Concept demos showing the flagship bar we set for ourselves. Browse both — they're labeled honestly.

We test eval harnesses BEFORE code ships.

TETRA AI principle

Two practices · Two specialist teams

What our AI specialists deliver.

Two specialist practices, engaged independently or together. Most clients start with one and add the other later — not both at once.

Practice 01 · TETRA AI

Intelligent Systems

Workflow automation, integration backbones, and pipeline systems that connect existing tools — Slack, HubSpot, Salla, Zid, ZATCA, n8n. Built by senior systems specialists.

INTEGRATION TOPOLOGYDASHBOARD · LIVE METRICS94%12k<2s0ORCHn8nLIVECRMHubSpotZohoCOMMSWhatsAppSlackPAYMENTSZATCAMadaANALYTICSGA4MixpanelAPI FLOWREQ → LOG → RES
  • CRM Setup
  • WhatsApp Auto
  • Sales Pipeline
  • Lead Capture
  • Dashboards
  • Integrations

AR + EN bilingual · Audit log · <2s latency

Practice 02 · TETRA AI · CORE inside

AI Agents

Bilingual conversational agents with audit logs and escalation paths. Native AR/EN. Reviewer-in-loop for Saudi-business tone calibration.

AGENT CONVERSATIONعميل · CUSTOMERمرحبا، أبغى أحجزAGENT · CHATBOTSure — when works for you?CORE · WEBSITE CHATBOTWhatsAppWebEmailSMSVoiceAUDIT LOG · LAST 312:04tool: book_slot · ok12:03policy: PDPL · ok
  • Website Chatbot · AR/EN FAQ + Lead CaptureCORE
  • WhatsApp AI Agent
  • Lead Qualification
  • Customer Support AI
  • Appointment Booking
  • AI Analytics

AR/EN native · Escalation-aware · 99.5% uptime

Built for Saudi

Production AI that respects the Kingdom's frameworks.

We don't bolt compliance on at the end. Every engagement starts with framework gap analysis.

ZATCA Phase 2

E-invoicing on every billing automation we ship. XML schema validation built in.

PDPL

KSA-region data residency by default. Consent management for every collection point. Right-to-erasure workflows.

SDAIA

AI guideline alignment — transparency, bias review, human-in-loop where required.

SAMA / CMA

For fintech engagements: SAMA Sandbox liaison, CMA fintech permission preparation.

How we deliver

From audit to operating in four phases.

Same delivery ladder for both Intelligent Systems and AI Agents engagements. Predictable scope. Transparent pricing per phase.

01 / 04 · DISCOVER

Audit current workflows. Score automation candidates.

We map current state, score automation candidates by ROI, and produce a ZATCA + PDPL gap analysis. The output is a written brief with phase-by-phase scope and pricing band.

02 / 04 · DESIGN

System topology. Agent personas. Escalation rules.

We design the integration topology, define agent personas in AR + EN, and write the escalation rule book. Reviewer-in-loop protocol locked before any code is written.

03 / 04 · DELIVER

Build, test, stage, ship — with audit log.

Production code with eval harness. Adversarial testing for KSA-compliance. Staged rollout (internal → staging → production) with rollback plan and audit log on every action.

04 / 04 · OPTIMIZE

Monthly tuning. Cost attribution. Expansion roadmap.

Monthly eval re-run. Cost-per-conversation attribution. Quarterly business review with hand-built reports — not vanity metrics dashboards.

FAQ

Common questions about TETRA AI.

Do your AI agents speak Saudi-Arabic dialect?

Our agents speak fluent Modern Standard Arabic by default and are tuned to a Saudi-business tone via prompt engineering and a glossary layer (formal address forms, Saudi place names, currency phrasing, polite-form preferences). For brands that need stronger Najdi or Hijazi dialect lean, we add a dialect tone-pack and review with a native-Saudi reviewer before launch. We don't claim dialect-native generation — we claim dialect-tuned and reviewer-checked.

Are your systems ZATCA Phase 2 ready?

Yes. Every billing automation we ship validates against the ZATCA Phase 2 e-invoicing schema. We integrate with your fatoora portal and generate compliant XML for every transaction. Sandbox testing is part of every engagement.

Where is data hosted? Is it PDPL-compliant?

KSA-region hosting by default for any engagement processing Saudi customer data. We use Vercel's Riyadh edge or AWS me-south-1 depending on your stack. Consent management UI, right-to-erasure workflows, and processor-agreement templates are part of every PDPL-scoped engagement.

What's the difference between Intelligent Systems and AI Agents?

They're two distinct practices led by different specialists. Intelligent Systems is for teams replacing manual operations — CRM setup, sales pipeline automation, integration backbones, dashboards. AI Agents is for teams that need bilingual conversational interfaces with audit logs and escalation paths — Website Chatbot, WhatsApp AI, Lead Qualification. Most clients start with one and add the other later, not both at once.

Do you integrate with Salla, Zid, HubSpot, Zoho?

Yes — these are our most-installed integrations. Salla and Zid native (we ship on both platforms regularly). HubSpot, Zoho, Pipedrive on the CRM side. Slack, n8n, Make on workflow. Integration scope is part of Discovery — we don't add integrations mid-build.

What's a typical engagement size and timeline?

We publish a pricing band before kickoff so there are no surprises. Intelligent Systems engagements typically scope across 6-12 weeks with phased deliverables. AI Agents engagements scope 4-10 weeks. Most clients sign a Care retainer after launch for monthly tuning and cost attribution. Exact pricing depends on integration count and compliance scope — we share the band on the 30-minute call.

Ready?

Talk to an AI specialist.

Direct call with the senior who will lead your engagement. 30 minutes. Focused agenda — your situation, our preliminary scope, named specialists, and a price band before you decide.

Response commitment — 24 hours · weekdays · KSA business hours
Chat with us
TETRA AI · TETRA