Structured, typed, reliable
You define the schema. The AI Agent fills it. Always typed, always structured, always ready for the next rule downstream.
AI Capabilities Build with AI.
Build with AI.
Decide with certainty.
DecisionRules puts AI where your rules already live. Build them faster in plain language, reason over documents and unstructured input, and connect the AI tools your team already uses. Everything stays deterministic, versioned, and auditable.
Decision Table ready. Deterministic, versioned, ready to publish.
ABCDE Age Annual Income Credit Score Decision Rate age ▾ annual_income ▾ credit_score ▾ decision ▾ rate ▾ 1 ≥ 21 ≥ 30,000 ≥ 720 = APPROVE = 6.9 % 2 ≥ 21 ≥ 30,000 BTW 650 | 719 = APPROVE = 9.4 % 3 ≥ 21 ≥ 30,000 ≥ 650 = REFER = n/a 4 ANY ANY < 650 = DECLINE = n/a
Trusted by Industry Leaders All Over The World
The AI surface
AI on top, a deterministic engine underneath
DecisionRules is a stateless evaluation engine that responds to REST API calls. AI works on top of that engine to help your team author rules, reason over unstructured input, and connect your own AI tools.
AI Assistant
Your teammate inside the editor. Build, test, and explain Decision Tables in plain language.
AI Agent
A native rule type that reads documents and narratives and returns structured, typed output.
MCP for your AI stack
Build, run, and manage rules in DecisionRules directly from the AI tools your team already uses.
AI Assistant
Your right hand in the editor: analyst, guide, and builder in one
You do not need to know the platform inside out. You need to know your use case. The AI Assistant does the work with you, building, testing, and explaining your rules so you can stay focused on the logic.
- Builds and edits Decision Tables and Scripting Rules
- Functions, test data, and summaries on demand
- Answers from the docs and Academy
Built-in AI Expert
Ask anything about the platform or your rules. The Assistant answers from the official docs and Academy, and highlights the exact button to click.
AI Rule Builder
Describe your logic in plain language or paste your data. The Assistant builds a working Decision Table or Scripting Rule, from idea to execution in seconds.
Test Data Generator
Stop crafting test inputs by hand. The Assistant generates realistic test data on demand, covering edge cases, row triggers, and uncovered cases.
Function Builder
Need a complex calculation in a cell? The Assistant writes the function expression, explains existing ones, and helps you debug.
Rule Summarizer
Turn complex rule logic into plain-language summaries anyone on the team can understand.
Measured, not claimed
We measured
before we shipped
Two medium-complexity Decision Tables, built from scratch: a 15 to 20 condition rule and a multi-step calculation. Timed by hand, then with AI.
- Three experience levels: newbie, professional, expert.
- Same specs, real stopwatches, with and without AI.
- Result: ~60% less authoring time, up to 3x productivity.
MINS
240
180
120
60
0
~60%
Faster rule authoring
ExpertProfessionalNewbie
MANUAL AUTHORING WITH AI ASSISTANT
AI Agent
Thinking rule type that reads, reasons, and returns structured data
Decision Tables and Scripting Rules follow explicit logic. The AI Agent understands context. It reads what no other rule type can, documents, narratives, and signals, and turns them into structured output your rules can act on. Not a chatbot. Not an integration. Built into the engine.
- Decision making
- Deep analysis of documents
- Report writing
- Policy validation
Reason against documents
Attach a policy, a pricing guide, or a product manual. The AI uses it as live context when it makes its decision.
Show the reason behind it
Explainable AI returns a confidence score, a plain-language reasoning paragraph, and source fragments tied back to the exact input text.
Choose your own model
Connect Gemini, OpenAI, Anthropic, Vertex AI, or Microsoft Foundry. Switch models per rule, without touching your schema or your Decision Flow.
Cut cost with caching
Identical inputs return the same result instantly, without calling the model. Lower cost and latency, toggled per rule.
Keep flows running
Every AI Agent node has a dedicated error output. Route failures to a fallback, a human queue, or a default value, and carry on.
MCP for your AI stack
Use DecisionRules from any AI tool, through MCP
Your business rules live in DecisionRules. Now the AI tools your team already uses can work with them directly, through the Model Context Protocol (MCP).
- Real, governed rules you can run, not just suggestions
- Deterministic, versioned, auditable
- You approve every change
The main advantage
Author and run real rules from your AI tools
Connect any MCP client to your authenticated space, and let it create, update, test, and execute rules without leaving the tool your team already works in. What comes out is not a throwaway chat suggestion. It is a real, deterministic, versioned rule you can audit, and you review every change before it goes live.
- Real, governed rules, not chat output
- Deterministic, versioned, auditable
- You approve every change
Pull in managed templates and functions
Give your AI tools access to predefined templates, reusable components, and structured content.
Search the docs in plain language
Query the full DecisionRules documentation from inside your AI workflow, with answers and direct links.
Why DecisionRules
Why you still need DecisionRules, even with AI
A pure AI model can make a decision, but it cannot promise the same one twice. Here is the difference between a chat that suggests logic and an engine that runs it.
A generic AI copilot
- Suggests logic in a chat window
- You copy, paste, and hope it holds
- No version history or audit trail
- Output can change between runs
- No gate between suggested and live
DecisionRules AI
- Rules built and run inside the engine
- Every rule versioned and testable
- Full audit trail you keep
- Deterministic output, the same every time
- You own the architecture, no lock-in
Every rule you create stays deterministic, versioned, testable, and carries a predictable cost, so the decision behaves the same today as it will a year from now.
AI-accelerated migration
// legacy rule engine
existing logic + IBM ODM exports
↓ AI reads, consultants finish
DecisionRules
→ deterministic
→ versioned
→ testable
AI-accelerated migration
AI-accelerated migration off legacy engines
Migrating decision logic off a legacy rule engine is usually slow and manual. Our team uses AI to accelerate the heavy lifting, reading your existing logic, including IBM ODM exports, and rebuilding it in DecisionRules, while our consultants validate and finish the work.
You reach a working, deterministic rule set faster than a hand migration. Available as a Professional Services engagement, scoped to your rule set.
Talk to us about migrationYour models, your data, your cloud
Bring your own model
Connect your own model for both the AI Assistant and AI Agent. Gemini, OpenAI, Anthropic, Vertex AI, or Microsoft Foundry, encrypted and scoped to your Space.
Data residency
US, EU, or on-premise. You choose where your data lives.
No public training
Your rules and data are never used to train AI models.
Deploy anywhere
Public Cloud, Private Managed Cloud, or Self-Hosted on Docker and Kubernetes, including the on-premise AI Engine.
Stateless & auditable
A scalable REST API. AI accelerates authoring; the engine stays predictable and auditable.
Enterprise security
SOC 2 ISO 27001 SSO full RBAC
Get started
See what measured AI looks like
We'll help you design a clear path to centralized, transparent, and automated decision-making.
Need help?
Frequently Asked Questions
We're here to help with any questions about DecisionRules, from AI features to deployment, pricing, and supported platforms.