The black box for AI decisions

When an AI decision gets challenged, can you prove what actually happened?

AIDAL creates cryptographically verifiable evidence for AI decisions — with privacy-preserving ingestion and offline verification.

Investigate a decisionRead the evidence specification→
<0sSeal time
0Jurisdictions
0mTo integrate
Live Audit Chain Demo
Audit chainLive
a7f3c9b2e4d81f06c3a9...
3e8b2a91c7f04d22b185...
f2c7a3e9b1d58031e742...
Built for
EU AI Act·MAS FEAT·OJK AI Governance (2025)·DFSA UAE·FCA UK·CFPB US·APRA AU
Independently verifiable, not just logged
7 jurisdiction modules live
30-minute integration
AI explanation per decision
No-cost sandbox tier
Regulation text mapped article by article
01

AI makes
a decision.

→
02

Someone asks
“Why?”

→
03

Nobody
knows.

→
04

AIDAL already
has the answer.

01The problem

AI makes millions of decisions.
None of them are provable.

Gap 1

No explanation

When an AI denies a loan or flags a transaction, neither the company nor the customer can explain why — in a way that satisfies a regulator or stands up in court.

Gap 2

No audit trail

AI decisions live in databases that can be edited. There is no independently verifiable record of what was decided, when, by which model, and with what confidence.

Gap 3

Regulators are arriving

The EU AI Act, Singapore MAS FEAT, Indonesia OJK, and UAE DFSA all point the same direction: explainable, auditable AI decisions. MAS FEAT applies today; OJK has published AI governance guidance for banks; DFSA's DIFC is consulting on AI rules through mid-2026. The EU AI Act's high-risk deadline was pushed to December 2027, but enterprise procurement and audit-trail buildout take months, so it's closer than it looks. Most companies have nothing.

Solution

AIDAL solves all three

Every decision gets a record you didn't have to build yourself: independently verifiable, plain-English explained, and checked against your regulator's actual requirements. One request logs it — the hard part was never the integration.

02How it works

One line of code.
The tape starts here.

1

Get your API key instantly

Sign up with your company name and email. Your unique aidal_live_xxxx key is generated in seconds.

2

Send decisions after your model runs

After your AI makes a decision, send it to AIDAL — a single request, any language, any model, any cloud provider. Choose client-side hashing to send only a digest and keep your raw data off our servers entirely, or send the full decision if you want AIDAL's fairness/bias reporting.

3

We seal it and check compliance

You get back an independently verifiable record, a plain-English explanation, and a check against your regulator's actual requirements. Under the hood: cryptographically hashed and chained to every decision before it — the mechanism, not the headline.

4

Audit anytime. Download the PDF.

Open your dashboard. Download a regulator-ready compliance PDF in one click. Hand it to MAS, OJK, or the EU auditor. Done.

Python · One API call
# After your model runs
requests.post(
"https://aidal-production.up.railway.app/decision",
headers={"Authorization": "Bearer aidal_live_xxxx"},
json={
"decision_type": "loan_approval",
"model_used": "xgboost-v2",
"input_features": applicant_data,
"output": model_result,
"jurisdiction": "SG"
})
 
# Response in <1 second
{
"audit_id": "aud_a7f3c9b2...",
"hash": "sha256:3e8b2a91...",
"explanation": "Loan approved...",
"rules_check": {"status": "PASSED"}
}
03Jurisdiction modules

Built for your
specific regulator.

Pass your jurisdiction code and AIDAL automatically checks compliance, flags missing fields, and tells you exactly what your regulator requires.

IndonesiaActive

OJK Indonesia

  • OJK AI governance guidance (2025)
  • Explainability recommended
  • 5-year record retention
  • Risk classification logging
5 yr retention
SingaporeActive

MAS FEAT Singapore

  • Fairness checks per decision
  • Ethics documentation
  • Accountability trail
  • Transparency reporting
5 yr retention
European UnionActive

EU AI Act

  • Art 12 automatic logging
  • Art 13 plain-language explanation
  • Human oversight records
  • High-risk system classification
10 yr retention
UAEActive

DFSA UAE

  • DIFC AI & data protection consultation (2026)
  • Autonomous Systems Officer duties proposed
  • AI governance documentation
  • Not VARA — DFSA covers lending/AI, not crypto
7 yr retention
United KingdomActive

FCA United Kingdom

  • SM&CR senior manager accountability
  • No new AI rules — existing frameworks apply
  • Consumer Duty (PS22/9) outcomes testing
  • SYSC governance & controls requirements
7 yr retention
United StatesActive

CFPB United States

  • ECOA Reg B adverse action notice
  • CFPB Circular 2023-03 explainability
  • Disparate impact testing
  • Fair Credit Reporting Act (FCRA)
5 yr retention
AustraliaActive

APRA Australia

  • CPS 234 AI model inventory
  • CPS 230 audit trail for critical ops
  • Board-level AI governance
  • Post-deployment monitoring
7 yr retention
04Why AIDAL

Cheaper, faster,
already built.

AIDALBuild in-house
Integration time30 minutes6–12 monthsEngineering + legal + QA
CostFrom $0/mo$50k–$200k+Initial build, ongoing maintenance
Proof typeCryptographic hashDatabase recordEditable, not independently verifiable
ExplainabilityAI-generated, per decisionManual write-upPer incident, not per decision
Jurisdiction compliance7 modules liveHire a lawyer per regionMAS FEAT ≠ OJK ≠ EU AI Act
Regulator PDFOne-click downloadCustom builtMonths of compliance ops work
05Who uses AIDAL

Built for anyone
using AI to make decisions.

Fintech and lending

Credit decisions, loan approvals, risk scoring. Every decision logged, every rejection explainable. Mapped to OJK, MAS FEAT, and EU AI Act requirements.

Insurance

Claims approval, underwriting, fraud flagging. Full audit trail that satisfies regulators and silences complaints.

HR and hiring

AI screening decisions. The EU AI Act classifies these as high-risk. AIDAL makes every hiring decision contestable.

Dec
2027.
EU AI Act — high-risk AI deadline

The EU's Digital Omnibus pushed the high-risk deadline for credit, insurance, and employment AI systems from Aug 2026 to Dec 2, 2027 — but audit-trail buildout, model documentation, and procurement take 12–18 months, so 2027 is closer than it looks. Non-compliance: fines up to €15M or 3% of global annual turnover. Meanwhile Singapore's MAS FEAT already applies today, Indonesia's OJK has published AI governance guidance for banks (non-binding, but built on binding IT-governance rules), and UAE's DFSA is consulting on AI rules through mid-2026.

Get your API key now
06Independent verification

Your database is testimony.
This is proof.

A regulator doesn't trust what you control. AIDAL's record is verifiable by anyone, independent of AIDAL's own servers — a SHA-256 hash chain underneath, mathematically checkable, not just policy-promised. That is the difference between passing an audit and failing one — a black box recorder anyone can pull the tape from, not just AIDAL.

Verify a record →View anchor log →
How it works

Company gives regulator an audit ID. Regulator pastes it into the verify page. Green = untampered. No sensitive data revealed.

Daily anchor log

Every 24 hours AIDAL publishes a master hash to GitHub — permanent and verifiable even if AIDAL goes offline.

Why this matters

AWS and Azure cannot do this. AIDAL's proof exists independently of AIDAL. That is a fundamentally different level of trust.

Regulatory research

Built from the
actual regulation text.

Every jurisdiction module was built by reading the regulation documents directly — OJK's 2025 AI governance guidance, MAS FEAT Principles, EU AI Act Articles 9–17, DFSA's DIFC AI consultation, FCA SM&CR/Consumer Duty, CFPB Circular 2023-03 & ECOA Reg B, and APRA CPS 234/230 — across all seven jurisdictions we support.

Are you a compliance lawyer with MAS FEAT or OJK expertise? We'd like to speak with you →

OJK AI Governance — Indonesia

Guidance mapped (2025)

Aligns with OJK's 2025 AI governance guidance for banks (non-binding) and POJK No. 11/POJK.03/2022's binding IT-governance and retention rules.

MAS FEAT — Singapore

Principles mapped

Checks Fairness, Ethics, Accountability, and Transparency on every financial AI decision.

EU AI Act — Articles 9–17

Articles mapped

Article 12 automatic logging and Article 13 plain-language explanations on every decision.

DFSA — UAE

Consultation mapped

Tracks DIFC Consultation Paper No. 3 of 2026 on AI and data protection. Not VARA — VARA regulates virtual assets/crypto specifically, not general financial-services AI use.

FCA — United Kingdom

Framework mapped

SM&CR senior-manager accountability and Consumer Duty (PS22/9) implemented. The FCA has stated it does not plan AI-specific rules, relying on these existing frameworks instead.

CFPB — United States

Circulars mapped

ECOA Regulation B, CFPB Circular 2023-03, and Fair Credit Reporting Act adverse-action requirements implemented.

APRA — Australia

Standards mapped

CPS 234 information security, CPS 230 operational risk, and APRA's April 2026 AI Governance Letter requirements implemented.

07Compliance report

The document your regulator
actually wants.

One click from your dashboard. A professional PDF with an independently verifiable seal, AI explanations, and a compliance certificate — ready to hand to MAS, OJK, or an EU auditor.

✓
Independently verifiable certificate

Any third party can check it themselves, without trusting AIDAL's servers — a SHA-256 hash chain with a tamper-evident seal under the hood.

✓
Per-decision explanations

Every AI decision explained in plain language. Supports EU AI Act Art 13's explanation requirement, MAS FEAT, and OJK's 2025 AI governance guidance.

✓
Jurisdiction compliance mapping

Automatically checked against your regulator's exact requirements — not a generic template.

Download sample report →
aidal-production.up.railway.app/report/pdf
AIDAL. Compliance Audit ReportMay 2026
Acme Fintech Pte Ltd
Executive summary
Reporting periodQ1 2026
Decisions logged4,821
Rules-passed rate98.7%
JurisdictionSG — MAS FEAT
✓ Chain integrity verifiedCERT-2026-A7F3C9B2
08FAQ

Questions we get
before people sign up.

09Pricing

Serious about audit-readiness?
Start here.

Sandbox
$0forever · 1,000 decisions/month→ Integrate in 30 minutes
  • Full API access
  • AI-generated explanations
  • Independently verifiable audit log
  • 7 jurisdiction modules
  • Public verify page
Get your API key
Starter
$299per month · 50,000 decisions→ Integrate in 30 minutes
  • Everything in Sandbox
  • Compliance dashboard
  • Human oversight tracking
  • Model registry
  • Email support
Get started
Audit-Ready
$2,500per month · 500,000 decisions→ Integrate in 30 minutes
  • Everything in Starter
  • Regulator-ready PDF reports
  • Fairness detection
  • Incident reporting
  • Priority support + SLA
Get started
Enterprise
Customunlimited · white-label · SOC2→ Dedicated onboarding
  • Everything in Audit-Ready
  • White-label deployment
  • On-premise option
  • Custom jurisdiction modules
  • Dedicated team
Contact us

or email anthony@tryaidal.com

All plans include an independently verifiable audit trail (SHA-256 hash chain) and a public verify page. No credit card required for the Sandbox tier. Running at million-decision scale? Sandbox, Starter, and Audit-Ready cap at 1k–500k/month — talk to us about Enterprise (anthony@tryaidal.com) for unlimited volume.

10Get started free

Provable AI decisions.
Starting today.

No credit card. No manual approval. Start logging decisions immediately.

Your key is generated instantly. Start logging decisions in 30 minutes.

SHA-256 sealed· Instant key· No credit card