The Decision Layer · Deterministic · Traceable

The Decision Engine for the AI Era

Sertainly compiles business logic into deterministic decision APIs — versioned, governed, and traceable. AI can reason. Business decisions must be deterministic.

Request
POST /evaluate

“package_id”: “expense_policy”,
“case”: {
  “request”: {
    “category”: “MEAL”,
    “amount”: 185,
    “attendees_count”: 3
  }
}
Response
{
  “verdict”: “compliant”,
  “reason_codes”: [“MEAL_UNDER_LIMIT”],
  “trace_id”: “tr_8f2a...c3d1”
}

Your business logic has no home

Every organisation has rules that govern decisions — pricing logic, approval thresholds, eligibility criteria, compliance requirements. Those rules live everywhere and nowhere. AI and agents make this worse: without explicit, governed rules, every decision is an improvisation.

Buried in code

Business rules are scattered across services, duplicated in every app. Changing a threshold means deploying code. Nobody can audit which version of which rule is in effect.

Locked in documents

Thousands of pages of policy, compliance rules, and operating procedures exist as PDFs. No system reads them. No agent follows them. They're shelfware — until someone gets it wrong.

Hardcoded into prompts

AI agents improvise rules from context windows. Different prompts produce different answers. There's no versioning, no audit trail, no governance, and no guarantee of consistency.

Decision infrastructure for the AI era

Sertainly is a dedicated architectural layer where business logic is compiled, governed, deployed, and executed as deterministic APIs. Every decision is traceable to its rules, its source documents, and its inputs.

Compile

Upload documents or write rules directly. AI extracts logic and compiles it into versioned, immutable decision packages.

Govern

Review, test, and approve every rule before it reaches production. Full audit trail from source document to runtime artifact.

Deploy

Publish as REST APIs and MCP tools. Pin versions per environment. Roll forward or back instantly.

Execute

Every call returns a deterministic verdict, reason codes, and a full trace. Same input, same output — always.

Two capabilities. One platform.

Sertainly combines two capabilities that are usually separate: turning documents into executable rules, and running those rules as decision infrastructure.

AI-powered rule compilation

Upload policies, contracts, handbooks, or compliance documents. AI extracts every rule, threshold, and routing condition — then compiles them into formal decision rules. Nothing reaches production without review and approval.

  • Source document to executable rule
  • AI-assisted extraction with human governance
  • Versioned, immutable decision packages
  • Full provenance — every rule traces to its source

Decision execution infrastructure

A dedicated decision layer in your enterprise architecture. Compiled rules run as deterministic APIs that any system can call — agents, applications, workflows, or enterprise systems.

  • Deterministic decision APIs — REST + MCP
  • Version-pinned deployments per environment
  • Full trace on every decision
  • Zero AI at runtime — pure deterministic execution

From rules to production API

AI does the work at build time. The runtime is pure computation.

1

Compile

Upload documents or write rules. AI extracts and compiles them into versioned decision packages.

2

Govern

Review every rule. Test with AI-generated scenarios. Approve before production.

3

Deploy

Publish as REST APIs and MCP tools. Pin versions. Roll forward or back instantly.

4

Decide

Any system calls the API. Every call returns a deterministic verdict, reason codes, and a full trace.

One engine. Three reasons to use it.

Agent Builders

AI agents cannot safely improvise business decisions. Sertainly gives them deterministic decision APIs they call via REST or MCP — action gating, routing, validation, and structured data collection with needs_info.

  • Deterministic verdicts — same input, same output, always
  • REST + MCP — works with any agent framework
  • Action gates — approve or deny before execution
  • Sub-millisecond evaluation — no inference at runtime
Explore agent integration →

Developers

Write decision rules locally in BDL. Validate and test on your machine. Publish to Sertainly. Evaluate via REST or MCP. Change rules without redeploying — business owns the logic, engineering owns the integration.

  • CLI-first workflow — author, validate, test, publish
  • REST + MCP runtime — single POST, deterministic result
  • Version-pinned artifacts — no silent logic drift
  • AI optional — no inference at runtime
See the developer platform →

Enterprise

Make business decisions a governed system in your architecture — not application code scattered across teams. Sertainly provides the full pipeline from source document to production decision API, with approval workflows, RBAC, and regulator-ready audit trails.

  • Document-to-rule conversion with full provenance
  • Governance — RBAC, approval workflows, audit trails
  • Deployment options — our cloud, your cloud, or on-prem
  • Zero AI at runtime — deterministic, always
Explore enterprise →

Not another rules engine

Traditional rules engines embed logic in applications. Sertainly runs decisions as infrastructure.

DimensionTraditional Rules EnginesSertainly
Where rules liveEmbedded in application code or databaseIndependent decision layer
How rules are authoredManual coding by engineersAI-extracted from documents + human review
Governance modelCode review / PR-basedApproval workflows, RBAC, audit trails
Runtime modelCoupled to the applicationStandalone deterministic API
Deployment independenceRules deploy with the appRules deploy independently, version-pinned
TraceabilityApplication logsFull trace per decision with source citations
DeterminismDepends on implementationGuaranteed — compiled, no AI at runtime
Agent readinessNot designed for agentsREST + MCP, structured needs_info, action gates

Decision infrastructure for regulated industries

Where compliance, auditability, and consistency are non-negotiable — business logic deserves to be infrastructure, not improvisation.

Healthcare

In healthcare, a wrong decision isn't just costly — it has consequences. Enforce prior authorization, formulary rules, and billing compliance deterministically with full trace for every decision.

Learn more →

Insurance

Every claims decision should be defensible. Compile underwriting guidelines and claims rules into executable logic. Every decision links back to the source clause that produced it.

Learn more →

Banking & Financial Services

Every credit decision you make is examinable. Deterministic loan eligibility, KYC/AML enforcement, and credit decisioning with version-pinned rules and regulator-ready audit trails.

Learn more →

Business logic is infrastructure.

Stop burying decisions in code, documents, and prompts. Make them a governed system in your architecture.

For DevelopersFor Enterprise