Enhanced decision nesting
Multiple nested conditions inside one rule block, for layered logic that used to need code.
Every decision, one flow, your rules.
Crux AI Decision Engine puts machine learning scorecards, business rules and live data into a single visual workflow, so your team can build, test and change decisions without waiting in a development queue.
Used for credit underwriting, fraud checks, onboarding and approvals, in banks, lenders, insurers and public services.
Most institutions decide twice. A model gives a score in one system, a rule turns it into an answer in another, and nobody can see both at once. We put them in the same flow, so a change in policy is an afternoon's work and a question from an auditor has a one-page answer.
What happens when a decision runs
Every decision follows a flow your team designed. The engine runs it in real time and keeps the evidence, so the answer always comes back with its reason attached.
An illustrative flow. Every run is logged with its inputs and the rule path taken.
1.Receive
Your origination, CRM or onboarding system calls the engine through a REST API or webhook.
2.Enrich
Internal records, bureau and third-party data are pulled in and validated before anything is scored.
3.Score and apply policy
AI scorecards predict risk, then your business rules turn the score into an action.
4.Respond and record
The decision and reason codes return in real time, and the full path is logged for audit.
The platform
Built so the people who own the policy can change it
Six things we care about, in the order our clients usually bring them up.
1
No-code workflow designer
Drag a block, set a rule, test it in the simulator and publish. Policy changes happen when the business needs them, not when a developer has capacity.
2
Scorecards and rules in one flow
A model score is just another step in the flow, so prediction and policy are managed, tested and audited together instead of being glued across two tools.
3
Minimal IT dependency
Business teams configure and edit flows themselves under role-based permissions. Developers handle the first integration, not every rule change.
4
API-first, cloud-agnostic
REST APIs and webhooks connect your loan origination, CRM and onboarding platforms. Run on cloud, on-premise or hybrid.
5
Versioning and governance
Every change is attributed, timestamped and versioned, and each live decision traces back to the flow version that made it. Roll back to a previous version when you need to.
6
Real-time monitoring
Watch decision distribution, rule hit rates and model drift live, with alerts when thresholds are crossed.
Wherever a decision repeats
Lending & credit
- Credit underwriting
- Loan origination
- Risk-based pricing
- Collections strategy
Risk & compliance
- Fraud detection
- KYC / AML screening
- Digital identity verification
- Regulatory compliance checks
Operations & eligibility
- Customer onboarding
- Insurance eligibility
- Expense approval
- Procurement authorisation
One platform, or two tools held together by tape
Many institutions run models in one system and rules in another. That works until policy has to change quickly, or someone asks why a customer was declined.
Separate scorecard tool + rules engine
Crux Decision Engine
Scores and rules kept in different tools
Scores and rules in one visual workflow
Rule changes queue behind developers
Business teams edit, test and publish themselves
Audit evidence pulled from several systems
One decision trail with inputs, version and rule path
Integrations built and maintained per tool
One API for upstream and downstream systems
Licences and support spread across vendors
A single platform to license and support
In development
Coming
Coming
Coming
What we are building next
Advanced version comparison
Side-by-side comparison of flow versions, so reviewers see exactly what changed before they approve.
Smarter automated alerts
Configurable notifications when thresholds, anomalies or conditions are met, routed to the right people.
Questions
Things people ask first
What is an AI decision engine?
A decision engine automates decisions such as approving a loan, flagging a transaction or routing an application. An AI decision engine combines predictive models, like credit or fraud scorecards, with business rules and live data in a single flow, so each decision uses both statistical prediction and your policy.
What is the difference between a rules engine and a decision engine?
A rules engine applies if-then logic that people write. A decision engine does that too, but also calls predictive models, pulls in external data, orchestrates several steps and records why each outcome happened. Crux puts scorecards and rules in the same workflow so they do not need separate tools.
Can business teams change decision flows without developers?
Yes. Analysts build and edit flows in a visual editor, test them in a simulator and publish with role-based approval. Developers are only needed for the initial system integration, not for every rule change.
How does the Decision Engine connect to our systems?
Through REST APIs and webhooks. Your loan origination, CRM or onboarding platform sends a request and receives the decision and reason codes in real time. Deployment can be on cloud, on-premise or hybrid.
How are changes governed and audited?
Every change is attributed to a user and timestamped, flows are versioned, and each live decision can be traced to the flow version, inputs and rules that produced it. This supports internal audit and regulator review, and your compliance team sets the approval policy.
Put decision control back in your team's hands.
In a short demo we walk through one of your real decisions, such as credit approval or onboarding, and show how it would be built, tested and governed.