# Put AI agents to work. Keep your policies in charge.

Source: https://aletyx.ai/platform/

Aletyx Platform

Agents that only draft and assist don’t change your numbers. Connect them to your systems, let your policy decide what they can do, and bring a person in when it calls for review.

[Explore the Platform](https://aletyx.ai/platform/#agents-in-practice) [Talk to a founder](https://aletyx.ai/contact/)

Agents under policy

## Even the CEO has to follow the policy.

The discount policy shown on Home returned 10%. Here, revised policy v2.0 returns 15%. The CEO asks for 25%, but the answer stays 15%. Every answer carries a record of the inputs, the version, and the rule that decided it.

[Watch an agent follow the policy](https://aletyx.ai/platform/walkthroughs/?demo=in-practice#see-it-in-action)

Your policy governed the action. The proof is recorded.

> Our critical decisioning now runs on Aletyx within Plaisse, enabling us to continue modernizing our legacy Drools/BRMS infrastructure **in weeks, without business disruption**, and providing a scalable, AI-enabled foundation for our future decisioning capabilities.

**Darshan Mehta** Managing Director, Application Development at Pennymac

Aletyx Platform starts at $25,000/year. [Compare editions](https://aletyx.ai/pricing/)

[Rules & decisions](https://aletyx.ai/platform/#decisions) [Releases & evidence](https://aletyx.ai/platform/#governance) [Flow & integrations](https://aletyx.ai/platform/#workflows) [People & approvals](https://aletyx.ai/platform/#people-and-work) [Full walkthroughs](https://aletyx.ai/platform/walkthroughs/)

Rules & decisions

## Delegate decisions. Keep your policy in charge.

Put agents to work on pricing, eligibility, and approvals. Your team writes the policy. Applications and agents run it: same inputs, same outputs.

Check the boundaries before your customers do. Review the rules and run saved tests before anything is published. AI Assistant is built in to explain a policy, draft it from a document, and build the tests.

### Your policy becomes an executable model.

See how the policy and the order become the answer.

### Inspect the rules behind the answer.

Business orders below 10 items receive 15%; larger orders receive 20%. Your team can review each rule.

### Test the edges of the policy.

Seven saved tests cover both customer types and the 10-item threshold. Explore a case to see its inputs and result.

[Explore decision modeling and testing](https://aletyx.ai/platform/walkthroughs/?demo=create-with-ai#see-it-in-action)

Releases & evidence

## Review changes before they reach your customers.

A loan locked in 2022 is judged by the rules in force in 2022. Aletyx keeps every published version available and routes each request to the one that applies.

Changes reach production through environments and approvals you control. When someone asks why a result happened, the execution record shows the inputs, the version, and the rule. Your team can defend the answer.

### Keep policy versions available when you need them.

This mortgage example keeps three published versions serving, each with its own effective date.

### Run the version that applies to the request.

A routing policy selects the mortgage rules in force on the loan's lock date.

[Explore execution records in detail](https://aletyx.ai/platform/walkthroughs/?demo=run-in-aletyx#see-it-in-action)

Flow & integrations

## Get work done across the systems you already use.

A Flow is a visual map of how work moves. It connects your enterprise systems, calls your policy, and coordinates the next action. Agents get a way to complete work across your business, with policy and human approval built into the path.

AI can help build the Flow, then validate, test, and publish it. Your own agent can do the same through MCP, the open standard for connecting agents to tools.

### Need BPMN too?

Teams that use BPMN can also model and run their business processes in Aletyx Platform, including branching paths and human approvals.

### Start with the information your business needs.

An equipment request enters the Flow through an integrated form.

### Let your rules choose the next step.

The Flow calls the equipment-review decision. Requests over $1,000 need a person to review them.

### Keep routine work moving. Route the exceptions.

Requests that need review go to a human task; the other branch can continue directly.

[Watch AI build and publish this Flow](https://aletyx.ai/platform/walkthroughs/?demo=delegate-to-ai#see-it-in-action)

## Your agent can build it, too.

Connect the agent you already use through MCP. It can create the workspace, write the policy and the Flow, test them, and publish. Published policy skills let any agent run your policy and return the result with its record.

[Explore the agent workflows](https://aletyx.ai/platform/walkthroughs/?demo=define-with-your-agent#see-it-in-action)

One request. Models, tests, and a published application.

People & approvals

## Some decisions need a person. Your policy says which.

When policy calls for human review, the request lands in one task inbox with the context a reviewer needs. Nothing risky moves forward until someone has decided. Routine work keeps moving.

People and agents work in the same flow, and your team builds it together. Live multi-user editing and shared data tables keep everyone on the same version.

### Put the exception in front of a person.

The equipment request appears in the task inbox with the unit, version, and status behind it.

### Give your reviewer the context to act.

Open the task, inspect its inputs, and add your response in the integrated form.

### Keep the work moving.

Submit the review so the Flow can continue. You can explore the task yourself.

[Watch a reviewer handle the task](https://aletyx.ai/platform/walkthroughs/?demo=review-in-the-inbox#see-it-in-action)

## Want to see every step?

Explore the full interactive walkthroughs for authoring, collaboration, agents, and execution. Choose the workflow that matters to you.

[Choose a full walkthrough](https://aletyx.ai/platform/walkthroughs/)

Already running Drools or KIE Server?

## Enterprise support for the systems you depend on.

Aletyx Platform runs on enterprise-supported builds of Drools, Kogito, and KIE Server. The rules and workflows you already trust keep running while you put AI agents to work.

[Explore modernization](https://aletyx.ai/modernization/)

[

### Enterprise Drools

Run your business rules with enterprise support, including existing Drools v7 applications.

](https://aletyx.ai/enterprise-drools/)[

### Enterprise Kogito

Build and run cloud-native decisions and workflows with enterprise support.

](https://aletyx.ai/enterprise-kogito/)[

### Enterprise KIE Server

Keep existing rules, processes, and integrations running while you modernize at your own pace.

](https://aletyx.ai/enterprise-kie-server/)

Frequently Asked Questions

## What teams ask before they start.

01 What does Aletyx Platform cost?

Aletyx Platform starts at $25,000 per year. Every edition includes the core Platform: rules, Flow, collaborative modeling, forms, data tables, a unified task inbox, and native MCP. Editions differ in capacity, compatibility, governance, and support. [Compare editions](https://aletyx.ai/pricing/).

02 Where does Aletyx Platform run?

On your own infrastructure. Aletyx Platform ships as a container that you deploy and operate in your environment.

03 Which AI models does it use?

AI Assistant is a service operated by Aletyx. It uses third-party AI models together with Aletyx’s own systems. Your own agents can use the popular model providers that support MCP and agent skills, so you keep the provider you already work with.

04 What happens to our data?

The Platform runs in your infrastructure, so your policies and records stay with you. When someone uses AI Assistant, what they submit goes to Aletyx to produce the answer, and that is all it is used for. We don’t train on it, and we don’t let our model providers train on it. It is yours, and so is the output. Keep personal data out of it.

05 Can we use our own AI agent?

Yes. Native MCP connects your agent to Aletyx Platform so it can create workspaces, build models, test, and publish. To run a published policy, your agent uses that policy's skill to call the decision service. Aletyx evaluates the rules; the agent returns the result and can provide links to the model and execution record. [Watch an agent build through MCP](https://aletyx.ai/platform/walkthroughs/?demo=define-with-your-agent#see-it-in-action) or [run a published policy through its skill](https://aletyx.ai/platform/walkthroughs/?demo=run-with-your-agent#see-it-in-action).

06 Can an agent override the policy when someone asks it to?

A request in the conversation does not change the published rules. In the discount example, revised policy v2.0 returns 15% for a Business customer ordering five items. Asking for 25%, even as the CEO, leaves that policy result unchanged. To change the outcome, your team must change the model and publish a new version. [See the policy-controlled result](https://aletyx.ai/platform/walkthroughs/?demo=in-practice#see-it-in-action).

07 How does Aletyx choose which policy version to run?

Published unit versions remain available, and a routing policy chooses which version applies to a request. For example, when a lender needs to use the rules in force on a loan's lock date, routing can select that historical version instead of today's rules. The execution record identifies the version that ran, its inputs, and the rule behind the result. [Explore versions and routing](https://aletyx.ai/platform/#governance).

08 How do Flow and decisions work together?

Decisions apply your business rules to determine an outcome. Flow connects enterprise systems, calls those decisions, and coordinates the next steps for people and AI agents. Teams that need BPMN can also model and run their business processes within Aletyx Platform.

09 What happens when a request needs human approval?

Your policy and Flow define when work goes to a person. In the equipment request example, requests over $1,000 need review. The Flow creates a task in the unified inbox; a reviewer sees the request, responds through an integrated form, and completes the task so the Flow can continue. Routine requests follow the other branch. [Follow the human approval](https://aletyx.ai/platform/walkthroughs/?demo=review-in-the-inbox#see-it-in-action).

10 Can we keep our existing Drools, Kogito, or KIE Server workloads?

Aletyx Platform includes enterprise-build options for [Drools](https://aletyx.ai/enterprise-drools/), [Kogito](https://aletyx.ai/enterprise-kogito/), and [KIE Server](https://aletyx.ai/enterprise-kie-server/). Start with the workloads you already depend on and plan the right modernization path for your versions and deployment.

11 Is AI Assistant included?

AI Assistant is built into Aletyx Platform, and your team can turn it on or off. Using it requires the [Aletyx AI Assistant](https://aletyx.ai/ai-assistant/) subscription, priced separately per user. You can also use the assistant on its own in the free Aletyx Playground.

12 What happened to Decision Control and Decision Control Tower?

They are now part of Aletyx Platform. Model authoring, testing, execution, and enterprise governance come together in one offering, alongside decisions, Flow integrations, forms, and tasks.

Start with one workflow

## Where could your agents do more for your business?

Bring one workflow you want to improve, or the system you need to modernize. We'll help you find a practical first step, with your rules and your team in control.

[Talk to a founder](https://aletyx.ai/contact/)