Arqis
Solutions ▾
Neoclouds & GPU cloudsGPU supply and customer offtake Data-centre & AI infrastructure operatorsLeases, power and tenant concentration OEM & hardware suppliersDelivery, allocation and warranty exposure Banks, funds & private creditGPU-backed financing and covenants Infrastructure funds & investorsCross-border AI infrastructure capital Insurers & reinsurersStructured exposure and accumulation AI labs & compute buyersFrontier training and inference deals
How Arqis is used Risk coverage Pricing Extensions Methodology Security
Client login
Extensions

Two things your risk record should already be able to produce.

Both run on the same graph, the same methodology and the same audit trail as the three modules covered today. Neither is a separate engagement with a separate team, and neither requires you to hand over anything you haven't already scoped.

01 · Compliance manual review and updateYour manual, tested against the rules and against the contracts you actually signed 02 · Insurer-ready and lender-ready risk packagesThe same risk record, compiled into the shape an underwriter or a credit committee accepts
Extension 01 · Compliance manual review and update

Your manual was written for the business you were three years ago.

Manuals, sanctions and export-control policies, delegation matrices and obligation registers get written once, approved, and then drift from two things at the same time: the rules, which move without telling you, and the contracts you have signed since, which create obligations nobody walked back into the manual. The gap usually surfaces in an audit, a lender's diligence questionnaire, or after something has already gone wrong.

What Arqis does
Reads the manual against the regulatory registerEvery procedure that cites a superseded rule is flagged with the rule, its version and its effective date, rather than a general note that the section looks dated.
Reads it against your signed bookObligations you carry contractually with no procedure behind them, and procedures maintained for obligations you no longer have. Both are findings.
Returns it redlined, with a gap registerEach gap carries the clause or rule it comes from, an owner, a deadline and a residual-risk target, so it can be worked rather than filed.
Maps every control to what it servesControl → obligation → rule, so the next review re-tests the manual instead of re-reading it, and so an auditor can follow the same path you did.
Reopens on changeWhen a rule moves or a new contract lands, the affected sections reopen automatically. The manual stops being an annual event.
What you get
  • Redlined manual and a change log
  • Gap register with owners and deadlines
  • Control-to-obligation-to-rule map
  • Board and audit-ready summary
  • Refresh on every rule change
Measured on
  • Gaps closed against gaps found
  • Time from rule change to manual updated
  • Findings raised against the manual in audit or diligence
Runs onRegulatory Dependency Graph (04) and Contract Risk Engine (03)
What we do, and what stays with you
The engine does the readingFrontier models read the manual and the contracts; the deterministic layer does the mapping, so every finding names the rule, the version and the clause it came from rather than a model's impression of the section.
A named reviewer walks the exceptionsContested and fact-dependent findings are escalated, not guessed. That review is scoped in the fee and named at engagement, and it is where the hours actually go.
Sign-off stays with youWe do not approve, certify or attest your manual. The redline and the gap register go to your compliance officer or counsel, who owns the decision and the version they publish.
Typical trigger: an audit, a new jurisdiction, a lender or insurer questionnaire, or a rule change nobody has walked into the manual yet
Extension 02 · Insurer-ready and lender-ready risk packages

The insurer cannot price what it cannot see. Neither can the credit committee.

Today the submission is assembled by hand: a spreadsheet, a memo and a folder of PDFs, in a different shape for every counterparty. Every question that comes back costs another week, and the answer is rarely evidenced in a form the other side can rely on. Arqis compiles the risk record you already hold into the shape underwriters and credit committees accept.

What the package contains
Exposure by risk, in dollars and MWWith the method shown rather than asserted, so the number can be interrogated instead of taken on trust.
Mitigations, each citedEvery mitigation traced to the clause, security, guarantee or insurance that actually carries it, not to a policy statement that it should be carried.
Residual exposure after mitigationWhat remains open, who owns it and by when, so the other side is reading a position rather than a snapshot.
Concentration and accumulationThe same tenant, supplier, site or grid across the book, which is the exposure an underwriter is actually pricing.
Covenants, closing requirements and obligationsTracked on one data layer rather than across a spreadsheet and a folder of PDFs, so the committee is reading the same position the servicing team is.
Monitoring evidenceWhat is watched, how often, what fired, what was escalated and what was done about it, with dates. Breaches surface before they become defaults.
The audit trail behind every numberSo a question comes back with an answer already attached to it. The evidence an examiner or a credit committee asks for is produced by the record, not assembled for the review.
What you get
  • An underwriting submission pack
  • A lender and credit-committee pack
  • A covenant and closing-requirements register
  • A concentration and accumulation appendix
  • Refresh on demand, or when the facts change
Measured on
  • Time to produce a submission
  • Questions returned per submission
  • Covenant breaches caught before they become defaults
  • Premium or spread improvement on evidenced risk
Runs onCounterparty Risk (02), Contract Risk Engine (03) and Regulatory Dependency Graph (04)
Formats it is produced in

The pack is generated to a named target format, agreed at scoping, rather than to a generic template. The four below are the formats it is produced in.

Broker market submissionThe pack a broker takes to market: exposure by peril, mitigations, loss history and accumulation.
Credit-committee paperThe appendix set behind a committee memo: exposure, security, covenants and residual, each cited.
Borrowing-base and covenant certificateThe evidence pack that sits behind a certificate, refreshed on the reporting cycle, with every figure reproducible from the audit trail.
Annual renewal packThe same record refreshed for renewal, with what changed since the last submission called out.
This is also how Underwriting Engine (07) and Continuous Credit Monitoring (06) reach the market before they are live as modules
How they're priced

Both are extensions of a tier, not a separate engagement.

Neither extension is sold on its own: each one reads the record the platform is already maintaining, so there is a Review, Monitor or Enterprise tier underneath it. Enterprise includes an agreed volume of both.

Manual review
$20–40K
One-time, per manual or programme. Scoped on page count, jurisdictions and how much of the signed book it has to be read against.
Manual kept current
$8–15K
Per year. Affected sections reopen on every rule change, with a redline and a change log each time.
Risk package
$10–20K
Per package, per counterparty. Covers the first refresh; later refreshes are generated from the same record.
Packages, unlimited
$25–45K
Per year across the book, for customers running renewals, syndications or a placement programme on a cycle.
See the tiers these attach to → Talk to us about an extension