Basirah · Pitch Deck
Banking & Fintech · RISK & DATA SCIENCE
Basirahبصيرة
Risk, AML, and credit models that actually read Arabic transaction memos.
neuriya.ai — Sovereign, Arabic-first enterprise AI
Problem
Saudi banks run global risk models on data those models can't actually read
- Most vendor credit and AML models are trained on English-language transaction data and simply drop or ignore Arabic free-text memos and merchant descriptors — the exact field where structuring and layering patterns hide.
- Annual credit-segmentation reviews mean a customer's risk tier can be six to twelve months stale by the time it's revisited, even as their transaction behavior changes weekly.
- AML analysts spend most of their time triaging alerts from static, rules-based thresholds tuned years ago, with no explainability trail a SAMA examiner or internal auditor can walk through in one screen.
- Model risk teams have no continuous view of data drift or demographic bias in production — problems surface only at the next scheduled model validation, often a year later.
Solution
Risk models trained on real Saudi transaction data, in the language it's actually written in
Product
Basirah in action
How it works
From first touch to outcome
Connect
Plug into core banking, card, and payment-rail feeds without duplicating your existing data warehouse.
Train
Models are trained and validated on your own transaction history, including Arabic memo and merchant text.
Decide
Credit, AML, and segmentation scores flow into underwriter and compliance workflows with full explainability.
Monitor
Drift, bias, and performance are tracked continuously, with recalibration proposed before accuracy slips.
Market
Sizing the opportunity
- TAM — $3.2B. GCC banking AI, analytics & RegTech spend
- SAM — $340M. Saudi credit-risk, AML & segmentation software market
- SOM — $22M. 3-year obtainable share across licensed Saudi banks and finance companies
SAMA's tightening AML/CTF supervision and Open Banking rollout are pushing more than 30 licensed Saudi banks and finance companies to replace batch, rules-based risk systems, and Vision 2030's Financial Sector Development Program treats this modernization as a national priority, not a discretionary IT upgrade.
Business model
How Basirah makes money
Annual platform license
Priced per institution and tiered by asset size and transaction volume, covering credit, AML, and segmentation modules.
Usage-based scoring API
Metered pricing per transaction or per model call for institutions that want to start with a single use case before expanding.
Implementation & model-tuning fee
One-time fee to train and validate models on the institution's own transaction history, including Arabic memo and merchant text.
Revenue streams
Where the revenue actually comes from
Competition
Where we win
Our edge
- Only Arabic-native transaction and memo model built and trained specifically on Saudi banking data, not a translation layer.
- Explainability is native to every score, not a bolt-on report — underwriters and examiners see the same evidence.
- Sovereign, PDPL-aligned, NDMO-classified model hosting — training data and weights never leave Kingdom infrastructure.
- Founding team combines Saudi enterprise delivery experience with hands-on applied-AI engineering, not a research demo looking for a customer.
Traction
Early validation, built for pilot readiness
- Full risk platform designed and built end-to-end: portfolio dashboard, AML alert queue, and credit decision detail with live explainability.
- Architecture pre-aligned to SAMA CSF, PDPL, NDMO Classification, and ISO 27001 from day one, not retrofitted for compliance.
- Actively selecting design-partner banks and finance companies for a first pilot cohort.
Go-to-market
Design-partner banks first, then sandbox-backed expansion
- 1Land with 1-2 design-partner banks or finance companies willing to validate models against their own portfolio data.
- 2Prove credit, AML, and segmentation outcomes over a defined pilot period against the institution's existing systems.
- 3Expand through the SAMA regulatory sandbox and industry risk-management networks to reach the broader licensed-institution market.
Team
Built by operators, not just builders
The ask
Raising a seed round to reach first design-partner banks
We're raising to fund pilot deployments with design-partner banks, model validation and SAMA CSF alignment work, and the core data-science and engineering team through first-revenue.