Every ministry, bank and operator asks the same question: Ask your enterprise data in Saudi Arabic, in-country, with a number your auditor will accept
Sawal, Bilytica's Arabic-first NL2SQL platform, is engineered in Riyadh for Saudi regulators, Saudi dialects and Saudi procurement. Hosted in-country on HUMAIN, aligned to SDAIA, PDPL, SAMA, CST and NCA ECC-2, with a published Arabic NL2SQL evaluation harness and a Vision 2030 KPI template library.
- Arabic NL2SQL accuracy
- 94.6%Arabic NL2SQL accuracy
- Saudi enterprise scenarios
- 1,200+Saudi enterprise scenarios
- enterprise security stack
- 7 layersenterprise security stack
- free PoC on your data
- 30-dayfree PoC on your data
In Saudi Arabia
What a cross-sector Saudi reviewer checks before an NL2SQL bot goes live
Every question, Arabic or English, runs through the same governed semantic layer. This is the evidence the sector authority expects before the bot touches production numbers.
SDAIA, NDMO
Metric definitions used by the bot match the definitions القطاعات السعودية reporting already uses.
Semantic layer built from the existing governed metric catalogue, not invented per question.
Metric-to-source mapping document reviewed by the reporting owner.
SDAIA, PDPL
Row and column-level entitlements applied before an Arabic or English question can return a number.
Identity resolved and entitlements applied at the semantic layer, ahead of query generation.
Entitlement test log showing the same question answered differently per role.
NCA ECC
Read-only execution path, with SQL validated and cost-checked before it runs.
Query validator gate ahead of execution, no write path exposed to the bot.
SQL audit trail attached to every answer, reviewable line by line.
Where it runs
The warehouse connection and the semantic layer run inside Saudi Arabia; the bot never ships raw data outside that boundary to generate an answer.
Arabic, as it is spoken here
Arabic business questions, including mixed Arabic-English phrasing common in Saudi teams, resolve against the same catalogue without a translation step.
Riyadh-based team, Arabic support in Riyadh business hours, WhatsApp channel
A proof of concept on YOUR sanitized data, you own the accuracy score
Published Arabic NL2SQL evaluation harness, audit the numbers, don't trust marketing
Named references in KSA banking, telecom, government and manufacturing (under NDA)
Questions your team actually asks, answered with governed SQL
In Arabic, English or a mix, the bot converts each question into SQL over a unified semantic layer.
ما نسبة السعودة في فروعنا بالرياض وجدة الشهر الماضي، مقارنة بالربع السابق؟
Which programs in the Vision 2030 portfolio are behind on Q3 milestones, ranked by budget impact?
أظهر إجمالي الإنفاق الحكومي على المشاريع الرقمية لعام 2026، مقسّماً حسب الجهة والمنطقة
Top 20 vendors by SAR spend, filtered by MOI compliance status, this fiscal year.
أعطني ملخص لتنفيذ مبادرات رؤية 2030 بجهتنا مع الجداول والتأخيرات
قارن أداء الفروع في الرياض بالفروع في المنطقة الشرقية على مدى ثلاثة أشهر
Where teams use NL2SQL in nl2sql for saudi enterprises
Ministries & PIF program offices
Program offices ask Vision 2030 milestone, budget, vendor and risk questions in Arabic, every answer respects entity-level RBAC and cites source ministries. Ready-made templates for MOF, MISA, MEWA, MOH, MOE, PIF program units.
Regulated sectors, banks, telcos, insurers
SAMA-, CST-, CMA-aligned NL2SQL bots with dialect-aware Arabic, in-country hosting, and immutable audit that passes external regulator review, not just internal audit.
Enterprise ERP + CRM unified Q&A
Business users ask across SAP S/4HANA, Oracle Fusion, Microsoft D365, Salesforce and Odoo, the bot joins them behind one semantic layer with reconciled ZATCA, IFRS, Saudization and Nitaqat rules.
Certified semantic layer + complete enterprise security layer
We never wire the bot straight to your warehouse, every question runs through a governed semantic layer and a security stack that respects every row and column.
Certified semantic layer
- One Arabic + English glossary per KPI, never two departments arguing over the same number.
- Semantic layer over dbt / Cube / LookML with pre-built Vision 2030 + ZATCA + Saudization templates.
- Public Arabic NL2SQL evaluation harness, accuracy scores your data team can independently audit.
- Graceful failure by default, the bot asks clarifying questions instead of guessing a wrong number.
Complete enterprise security layer
- In-country deployment on HUMAIN, Oracle OCI Jeddah, STC Cloud, or your own private tenant, never SaaS.
- SDAIA / PDPL aligned classification, retention and residency, cross-border data transfer: zero.
- Saudi SSO (NAFATH, Azure AD, Okta, PingID) with entity-, role- and time-window-level RLS.
- Prompt-injection defenses tuned for Arabic, Arabizi and code-switched jailbreak attempts.
- Immutable audit trail (WORM) suitable for NCA ECC-2, SAMA IT Governance Framework, CMA and CST reviews.
- ISO 27001 in scope; independent penetration test report shared under NDA before signature.
Questions Saudi leaders ask before rollout
Is Bilytica actually a Saudi company, or a Western vendor with an Arabic wrapper?
Riyadh-headquartered, Saudi ownership, Saudi delivery team, Arabic support in Riyadh business hours, WhatsApp channel for named accounts. Sawal was engineered in-country from day one, not translated from an English product.
Is the Arabic really built for Saudi dialects, not just MSA?
Yes. The bot handles Saudi MSA, Najdi, Hejazi and Eastern-region variants, plus Arabizi (Latin-script Arabic) and code-switched Arabic/English in the same sentence. Every dialect is measured, not claimed, in the published NL2SQL evaluation harness.
Our procurement is 9–18 months. What does the first month look like?
Week 1: signed NDA + sanitized data share. Weeks 2–3: bot connected, accuracy measured on your questions, PDPL/SAMA architecture reviewed with your risk team. Week 4: findings and fixed-scope pilot proposal. You spend zero on the PoC and walk away with an eval report you own, regardless of whether you buy.
How does hosting on HUMAIN help with sovereignty?
HUMAIN keeps model weights, prompts, embeddings and SQL logs inside Saudi Arabia, aligned to SDAIA guidance and PDPL cross-border rules. No customer question or answer leaves Saudi Arabia. We deliver the architecture diagram your NCA and SAMA reviewers accept.
What if the accuracy doesn't hit our threshold on our data?
You don't pay for the pilot. We agree the threshold in writing before the PoC starts. If we don't hit it in 30 days, the engagement ends, you keep the eval report, we keep the lesson. This is why we can afford to publish the harness: we bet on it.
See your first Arabic question answered on YOUR data, 20-minute working session with our Riyadh team. No PoC fees, no procurement paperwork.
No fees, no procurement commitment, no data leaving Saudi Arabia. You walk out of the session with an accuracy evaluation report your team owns, whether you buy or not.