Skip to content

VERIFICATION INFRASTRUCTURE FOR REGULATED AI

Prove what
your AI did.

AI now writes, decides and advises inside regulated firms. When a regulator asks what happened, most firms have nothing to show. Velkron is the audit layer: grounded, cited rule checks now, cryptographic proof of inference next.

console.velkron.xyzILLUSTRATIVE PREVIEW
Review queueTODAY · 24 ITEMS
promotion_draft_v3.pdfVERIFIED
crypto_promo_v1.htmlFAILED
pension_transfer_advert.docxREFUSED
isa_landing_page_v2.htmlVERIFIED
q3_fund_factsheet.pdfRUNNING
1,204 runs this month · 3.1% refused to humans · 0 guesses

SELECTED · LIVE REPLAY

EVIDENCE PACK · FIN-PROM-4471RUNNING
artefactpromotion_draft_v3.pdf
rule setFCA COBS 4 · FSMA s21
check 1/3risk warning present · COBS 4.2.1R
check 2/3balanced view of benefits · COBS 4.2.1R
LIVE REPLAY · 3 CASES

Product preview with sample data. The live console is at console.velkron.xyz.

ONE ENGINE · ANY RULEBOOKUK · FCALIVEADGM · FSRANEXTDIFC · DFSANEXTEU · AI ACTNEXT+ 10 MORE ON THE ROADMAP
FCA COBS 4.2.1R ·EU AI ACT ART. 12 ·ADGM FSRA COBS ·FCA FG24/1 ·DFSA COB MODULE ·EU AI ACT ART. 19 ·FCA COBS 4.12A.34R ·MAS FAA-G10 ·FCA SYSC 3.2.20R ·SEC RULE 206(4)-1 ·FCA PRIN 2A.5 ·EU AI ACT ART. 72 ·

THREE VERDICTS, NEVER A FOURTH

No black-box confidence score.

PASS

Every claim grounded, nothing flagged.

check 3/3no implied guarantee · COBS 4.5.2R ✓

SAMPLE FINDING FROM AN EVIDENCE PACK

FAIL

The exact rule and the exact words that broke it, cited.

breachincentive to invest found · COBS 4.12A.34R ✗

SAMPLE FINDING FROM AN EVIDENCE PACK

REFUSED

The rulebook doesn't clearly cover this case, so it says so instead of guessing.

actionescalate to compliance officer · no rule matched ⚠

SAMPLE FINDING FROM AN EVIDENCE PACK

Deterministic grounding

Every citation is checked in code against the rulebook and your artefact, not just prompted for. An invented rule or a quote that isn't verbatim is dropped before you see it.

Learn more

Sealed receipts

Every run is hashed and chained. Months later, in an audit, you can prove exactly what ran and what it found.

Learn more

Proof of inference

Zero-knowledge proofs that a specific model made a specific decision, without exposing weights or data. In research.

Learn more
01THE PROBLEM

The AI stack has no audit layer.

0promotions the FCA forced firms to amend or withdraw in 2024, up 97.5% year-on-year.

SOURCE: FCA.ORG.UK/DATA/FINANCIAL-PROMOTIONS-DATA-2024

1.1Regulators ask for evidence

From the EU AI Act to the FCA and the Gulf regulators, firms must show why AI output was compliant. A chat log is not evidence.

1.2Generation got the money

Billions went into making models produce more. Almost nothing went into checking what they produce.

1.3Unverifiable means unusable

In regulated work, an answer nobody can verify is a liability. The firms that deploy AI fastest will be the ones that can prove what it did.

§THE REQUIREMENT ALREADY EXISTS

We didn't invent the demand. Regulators on three continents wrote it down.

A firm must ensure that a communication or a financial promotion is fair, clear and not misleading.
FCA HANDBOOK · COBS 4.2.1R
High-risk AI systems shall technically allow for the automatic recording of events (logs) over the lifetime of the system.
EU AI ACT · ARTICLE 12(1)
An investment adviser may not include any untrue statement of a material fact in an advertisement.
US SEC · MARKETING RULE 206(4)-1 (SUMMARISED)

The FCA and SEC obligations above are live today. The EU AI Act Article 12 logging duty is enacted law with its high-risk compliance date deferred to December 2027 under the AI Act Digital Omnibus, coming but not yet in force. Velkron is the infrastructure that turns requirements like these, on whatever date they bind, into checks, records and proofs.

02HOW IT WORKS

Four steps. No black box.

01

Submit

A promotion, a decision, a document. Dashboard or API.

02

Check

Model-assisted read of the rulebook, grounding enforced in code. No invented citations.

03

Verdict

PASS, FAIL with the breach named, or REFUSED to a human.

04

Receipt

The run sealed in a hash chain. Audit-proof, months later.

FULL PRODUCT TOUR →

0

PROMOTIONS AMENDED OR WITHDRAWN, FCA 2024 (+97.5% YoY)

3

VERDICTS · PASS / FAIL / REFUSED

0

GUESSES · REFUSAL IS A FEATURE

SHA-256

EVERY RUN SEALED AND CHAINED

4

RULEBOOKS · FCA LIVE, GULF + EU NEXT

03THE API (PLANNED)

One call. Verdict, citations, receipt.

Drop verification into an existing approval flow, a CMS or an agent pipeline. If your AI publishes, the call goes in front of it. The shape below is the target design, not a live contract; the working product today is the web checker at console.velkron.xyz.

REQUEST

POST /v1/checks
Authorization: Bearer vk_live_...

{
  "artefact_url": "https://firm.com/promo.pdf",
  "rulebook": "uk-fca",
  "modules": ["COBS4", "CONC3"],
  "mode": "pre_publication"
}

RESPONSE · 200

{
  "verdict": "fail",
  "breach": {
    "rule": "COBS 4.12A.34R",
    "reason": "incentive to invest"
  },
  "citations": ["COBS 4.12A.34R"],
  "grounding": "sufficient",
  "receipt": {
    "sha256": "c81f30aa…77c2",
    "chained": true
  }
}

GROUNDING IS DETERMINISTIC · CLASSIFICATION IS MODEL-ASSISTED, NOT ENCODED LOGIC · EVERY VERDICT CARRIES ITS RECEIPT

04THE PRODUCT

One verification stack. Wedge first.

Others add AI to compliance. Velkron adds proof to AI, starting where the regulatory pain is sharpest. The same layer then extends to every regulated vertical.

Promotions checkerPILOT OPEN

Checks financial promotions against a regulator's rulebook before publication. Returns pass or fail with the exact rules cited, plus an evidence pack a compliance officer can file. When the rules don't support a verdict it refuses. It never guesses. First rulebook live: UK FCA. The engine is rulebook-agnostic; ADGM FSRA and DIFC DFSA are next. Try it at console.velkron.xyz →

Decision receiptsIN DEVELOPMENT

A tamper-evident record of every AI decision: model, input, output, sealed in a hash chain. Edit the history and the chain breaks in plain sight.

Proof of inferenceRESEARCH

Zero-knowledge proof that a specific model produced a specific decision, without exposing the model weights or the customer data.

05COVERAGE

One engine. Every regulated market.

A rulebook is content, not code. The verification engine stays the same; each jurisdiction is an encoded rule module with its own citations. Statuses below are honest: one live today, the rest in the order the market is pulling us.

EUROPE

United Kingdom

FCA

LIVE

European Union

AI ACT · ESMA

NEXT

Switzerland

FINMA

HORIZON

MIDDLE EAST

Abu Dhabi · ADGM

FSRA

NEXT

Dubai · DIFC

DFSA

NEXT

Saudi Arabia

CMA

HORIZON

AMERICAS

United States

SEC · FINRA

HORIZON

Canada

CSA

HORIZON

Brazil

CVM

HORIZON

ASIA-PACIFIC

Singapore

MAS

HORIZON

Hong Kong

SFC

HORIZON

Australia

ASIC

HORIZON

India

SEBI

HORIZON

LIVE = IN PRODUCTION · NEXT = IN BUILD ORDER · HORIZON = ROADMAP · THE WAITLIST SETS THE ORDER →

THE WAITLIST SETS THE BUILD ORDER

Your jurisdiction, sooner.

Rulebooks are built in order of demand. Join with your work email and jurisdiction: every signup moves that rulebook up the queue, and you get first access when it goes live.

THE POSITION

Trust is an assertion.
Proof is a receipt.

Every AI vendor asks regulated firms for trust. Velkron is built so nobody has to give it. Check the output, cite the rule, seal the record. That is the whole company.

06QUESTIONS

Asked and answered.

Q1Is this another AI compliance copilot?+

No. Copilots generate suggestions with a model and ask you to trust them. Velkron's model proposes findings; a deterministic layer then enforces every citation in code, dropping any rule id or quoted excerpt it can't verify against the rulebook and your artefact. The classification is model-assisted, not encoded logic; the grounding underneath it is.

Q2What happens when the system is not sure?+

It says so. Below a grounding threshold, Velkron returns REFUSED and escalates the case to a human, with its reasoning attached. The refusal rate is reported to you, not hidden. We consider this the most important feature in the product.

Q3Which rulebooks are supported today?+

UK FCA financial promotions (COBS 4, including the COBS 4.12A crypto regime, plus FSMA s21 and the FG24/1 standalone social-media guidance) is live. ADGM FSRA and DIFC DFSA are next on the roadmap, followed by EU AI Act conformity evidence. The engine is rulebook-agnostic; a new jurisdiction is content work, not a rebuild.

Q4What does a pilot involve?+

You send a set of recent promotions, we return full evidence packs, and your compliance team judges the output against their own review. Pilot terms are agreed per firm. If the packs are not useful, you will know within a week.

Q5Is our data used to train models?+

No. Your artefacts are processed for verification only, never for training. Receipts are hashes; they do not contain your content and cannot be reversed into it.

Q6Who is behind Velkron?+

A solo technical founder with a public build record in AI verification: open-source libraries, a clinical speech model, a formally tested blockchain and a RegTech retrieval engine. The full log is on the company page.

07CONTACT

Taking on a small group of design partners in regulated finance.

Wherever you're regulated: if your firm publishes financial promotions, or is putting AI anywhere a regulator can see, write to us. We reply to every request within two working days.

NOT READY FOR A PILOT? JOIN THE WAITLIST