v2.0

GoRules Version 2 is here - redesigned, now with managed cloud.GoRules Version 2 is here!

Watch the launch videoWatch

Credit risk software for lenders,
owned by your risk team.

Cutoffs, exposure limits, override rules, and post-booking review triggers in visual decision tables your credit risk team edits directly, versioned and approved like code. This is risk decisioning for a lending book, not accounts-receivable credit control.

From auto-approved to risk review,
start to finish.

Watch a morning on a portfolio risk desk: subprime exposure past the board cap, a one-cell ceiling change, a replay over 5,200 limit decisions, a signed release, and $3.2M of committed exposure pulled back the same day.

scroll to play
Tuesday, 09:12

Subprime exposure crosses the cap.

The committed book opens at $41.8M, $1.8M past the board cap. 148 limit increases went out this week and 76 percent of them were decided without a human.

The hot cell

Already at 0.82. Line raised anyway.

27 of those increases went to subprime accounts averaging 0.82 utilization today. The policy reads utilization after the increase, so $1,800 more line lands them at 0.75, just under the 0.80 ceiling.

One cell

The risk officer tightens the ceiling.

The rule that fired opens the live policy. The subprime ceiling drops from 0.80 to 0.65 - on a branch, prime and near-prime bands untouched.

Proof before production

5,200 limit decisions re-run.

Three months of limit decisions replay against the draft while you watch: 340 flip to refer, prime approvals do not move, and subprime exposure lands under the $40M cap.

Sign-off

The CRO signs. v2.9.0 ships.

Jane Cooper approves the release and it promotes through staging to production with the full trail recorded. Rollback stays one click away.

Same day

The curve bends the same day.

Auto-approvals fall from 76 to 58 percent, 47 increases route to risk review, and committed subprime exposure lands at $38.6M - $3.2M pulled back, prime and near-prime untouched.

Your credit policy is a document.
Your credit system is not.

Search for credit risk software and half of what comes back is built for a different risk. Accounts-receivable and order-to-cash suites manage trade credit for your own customers: credit applications, customer credit limits, blocked-order release, DSO and collections dashboards. Useful if you invoice on terms, irrelevant if you lend money. Lenders are asking a narrower question, and it is rarely about analytics: where does our credit risk policy actually run, and who is allowed to change it?

On paper the policy is governed. The credit committee approved score and PD cutoffs, exposure caps per borrower, product, and sector, collateral and DSCR floors, override authority by approval level, and the review cadence each risk grade triggers after booking. In production that document is scattered: a few conditions hard-coded in the origination service, a cutoff configured inside a scorecard vendor's console, an overlay spreadsheet the credit team maintains by hand, and an exception register somebody rebuilds before every board pack.

The market answers with size. Credit risk management software suites bundle spreading, rating models, portfolio analytics, and a decisioning module you configure through the vendor's professional services queue, so moving one threshold becomes a change request with a delivery date. Scoring vendors sell the model and stop at the score. The rules in between, the ones that actually approve, decline, refer, cap, and escalate, end up owned by whoever last touched the code.

The gap shows up under examination. SR 11-7 and OCC Bulletin 2011-12 put real change control around the model, yet the policy layer consuming its output often has weaker governance than the model itself. The OCC expects the board credit committee to monitor loan policy exceptions, and persistent underwriting exceptions read as a fair lending flag as well as a credit one. If nobody can say which version of a cutoff was live in March, who approved it, and how many exceptions it produced, the answer is a folder of emails.

Risk policy above
scores and models.

Your scorecards and PD models keep doing their job. GoRules owns the layer above them: the cutoffs, limits, overlays, and review triggers your credit committee actually signs off on.

01

The policy your committee approved, in tables it can read

Credit policy is already a matrix: risk grade by exposure by product, resolving to a decision, a cap, and a review cadence. GoRules decision tables are that matrix, edited spreadsheet-style by credit risk analysts and rendered in natural language so the table and the policy document say the same thing. Hard knockouts, sector caps, and segment overlays are modeled as Policies, ordered blocks evaluated in sequence, so an exposure breach or an active default fires before any score is consulted.

02

Every threshold change has an author and an approver

Branch the live policy, change the grade C cap, and let the committee read the diff in plain language before anything ships. Git-like version control records who changed which threshold and when, approval workflows gate promotion through dev, staging, and production, and one click rolls a change back if the portfolio disagrees. GoRules AI can draft the edit, run your test cases, and explain the resulting decisions while you iterate; releasing stays a human action with a name attached to it.

03

One policy from application to portfolio review

The rules that decide the application also decide what happens to the exposure afterwards. Grade migration, DSCR compression, utilization creep, late borrower reporting, and bureau triggers become monitoring rules in the same project, evaluated on an event in real time or in a nightly pass over the book, returning a watchlist tier and a review date instead of another dashboard. Application-time cutoffs and pricing sit in the same project; see credit decisioning for the origination side.

Keep the models.
Move the policy.

GoRules does not replace your scorecards, your core banking system, or your LOS. It is where the risk policy above them lives.

01

Connect

Your origination service, LOS, or portfolio system calls GoRules over REST, or embeds the MIT-licensed open-source ZEN engine through SDKs for Node.js, Python, Java, Go, C#, and Rust. Each request carries what the policy needs: bureau attributes, the PD or score your model emits, spreading ratios such as DSCR and LTV, and current exposure from your core.

02

Model the policy

Rebuild the credit policy as a decision graph: knockout Policies first, then the grade and limit matrix per product and segment, exposure and concentration checks, and a waterfall ending in approve, decline, refer, or approve with conditions. ZEN expressions handle derived values like DSCR, DTI, and LTV, and typed TypeScript function nodes cover custom work such as aggregating group exposure across related borrowers.

03

Test against your own book

Branch the change and replay historical applications and current exposures through both versions. See how many accounts move, which exceptions disappear, where the approval rate lands, and read the rule diff in natural language before the committee signs off.

04

Release and monitor

Promote through environments with recorded approvals, evaluate in production in under a millisecond, and keep every decision linked to the rules that fired. Exception and override counts become a query over decision logs instead of a spreadsheet assembled the week before a board pack.

Built for production,
not proof of concept.

  • Sub-millisecond evaluation with the open-source ZEN engine (Rust core), fast enough for real-time facility decisions and for replaying an entire book as a batch pass
  • Git-like version control on every threshold: branches, commit history, approval workflows, and one-click rollback when a change misses
  • Self-host with Docker or Kubernetes so borrower PII, bureau data, and exposure figures never leave your infrastructure, or run on GoRules Cloud
  • Complete audit trail of who changed which limit, when, and what fired on any decision, with SSO, role-based access control, and SOC 2 behind it
  • No lock-in: the engine is MIT-licensed open source with a Rust core, and 70+ industry templates, including loan approval and credit limit, give you a working starting point

Questions, answered.

What is credit risk software?

Credit risk software is the system a lender uses to measure and control the risk of not being repaid. It spans three layers: the models that estimate risk, such as scorecards and PD models; the policy that turns those estimates into actions, such as cutoffs, exposure limits, collateral floors, and override rules; and the monitoring that watches exposures after booking. Confusingly, the same phrase is also used for accounts-receivable tools that set trade credit limits on B2B customers and release or block sales orders. GoRules is the policy and decisioning layer for the lending sense of the term: the rules your risk team owns, versioned, approved, and evaluated in real time.

Is credit risk management software the same as accounts receivable credit management?

No, and the shared vocabulary wastes a lot of evaluation time. Order-to-cash and AR platforms manage trade credit extended to a company's own customers: online credit applications, customer credit limits, blocked-order release, DSO, and collections. Lender-side credit risk management software works on borrowers, facilities, risk grades, concentration limits, and regulatory reporting. A quick test on any vendor page: if it talks about blocked orders and DSO it is AR tooling; if it talks about borrowers, exposure caps, and exception reporting it is lending. GoRules serves the lending side.

What is risk decisioning?

Risk decisioning is the step between a risk estimate and an action. The scorecard or PD model answers how risky this borrower is; risk decisioning answers what you do about it given policy, product rules, affordability, fraud checks, and existing exposure: approve, decline, or refer, at what limit, at what price, and with what review cadence afterwards. Keeping that layer separate from the model means a cutoff move is a policy change made in an afternoon rather than a retraining and revalidation cycle. In GoRules it is decision tables and graphs your risk team edits directly.

Can our risk team change cutoffs and limits without an engineering release?

Yes, that is the point of the product. Decision tables are edited visually and rendered in natural language, so a credit analyst can move a grade C exposure cap or add a sector overlay without touching code, and GoRules AI can draft the edit, run the test cases, and explain the resulting decisions. Nothing reaches production on its own: changes travel through branches, approval workflows, and versioned promotion across dev, staging, and prod, with one-click rollback. Releasing stays a human action with an approver's name on it.

How does GoRules support credit policy exceptions, audit, and examiner questions?

Every evaluation returns the outcome plus the exact rules that fired, so an exception or a low-side override is recorded as a decision path rather than a note in a spreadsheet, and exception rates by product and segment become a query over decision logs. Every rule change carries an author, a timestamp, a diff, and an approver, so you can reconstruct which version of a cutoff was live on any given date. That is the record model risk management under SR 11-7, board credit committee reporting, and adverse action reasoning under ECOA all depend on. GoRules supplies the traceability; the obligations, the policy itself, and the validation work remain yours.

The same engine,
next door.

One decision layer serves the whole institution - these use cases run on the same tables, versioning, and audit trail.

Your limits. Your exceptions.
Your version history.

Start from the loan approval template, rebuild one cutoff and limit table with your current thresholds, and replay last quarter's applications through it this week.