> For the complete documentation index, see [llms.txt](https://documentation.ardoq.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://documentation.ardoq.com/core-features/ai/ardoq-ai-overview.md).

# Ardoq AI Overview

This page provides a comprehensive overview of all **AI and Generative AI capabilities available in the Ardoq platform**. This is a living page and will evolve as Ardoq's AI capabilities continue to expand.

### How Ardoq Uses AI (At a Glance)

Ardoq uses AI to **accelerate modeling, democratize insights, and make architecture data more useful** while keeping customers firmly in control. Learn about our [Ardoq AI Principles](https://www.ardoq.com/ai-and-innovation/principles) and follow our Ardoq AI Labs [ProductBoard](https://portal.productboard.com/zvnzutc4rg7xnc7pp8uiz7s7/tabs/14-our-ideas-) tiles to keep up with our AI roadmap & experiments.

### Ardoq AI Capability Maturity Tags

Each capability below is tagged with its current maturity:

* **GA** – Generally Available to customers
* **LA** – Limited Availability: production grade but phased rollout
* **Beta** – Early Access (Open Beta)
* **Labs** – AI experiment or technical preview capability

### Agentic AI

#### Omnipresent AI Assistant (GA)

![](https://2484004798-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FnlnB7uZ29iLVLzqGbCym%2Fuploads%2FtzBk5EHwpRe9vHoadaCM%2Fassistant.png?alt=media)

Ask questions about your architecture in plain English and receive answers grounded in your live data. The Assistant also supports agentic modelling of Ardoq content with human in the loop.

**Used for:** Application analysis, dependency questions, impact analysis, Agentic Ardoq modelling

**Learn more:**

* [Blog: Launching AI Chat Assistant: A Conversational Interface for Your Enterprise Architecture](https://www.ardoq.com/blog/launching-ai-chat-assistant?utm_source=beamer\&utm_medium=standalone\&utm_campaign=AI-Chat-Assistant---Available-in-Open-Beta\&utm_content=textlink)
* [KB Article: Omnipresent AI Assistant](https://help.ardoq.com/en/articles/510373-omnipresent-ai-assistant)
* [Latest improvements to the AI Assistant](https://help.ardoq.com/en/articles/510373-omnipresent-ai-assistant#h_b6aeca8d60)
* [Demo video](https://www.youtube.com/watch?v=vn5BXkDqHsE)

#### Build your own Custom Agents (Open Beta)

![](https://2484004798-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FnlnB7uZ29iLVLzqGbCym%2Fuploads%2FF2wHPSCmboVYAZwZD6js%2Fcustomagents.png?alt=media)

Define your own AI agent inside Ardoq with its own instructions, tools and behavior. Write a detailed prompt once, then run it by name from the AI Assistant. Agents follow your step-by-step procedure, gather data from your architecture graph and return a structured result. They can also propose changes, which are staged in a Scenario for human approval.

Admins create and manage agents. Any user with access can run them, always with their own permissions.

**Used for:** Architecture governance and review, data quality checks and clean-up, operational insights, repeatable analysis such as application rationalization

**Learn more:** [KB Article: Build your own Custom Agents](https://help.ardoq.com/en/articles/866857-build-your-own-custom-agents)

#### Ardoq Custom Agents

![](https://2484004798-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FnlnB7uZ29iLVLzqGbCym%2Fuploads%2FvPPKdJt1J9jPpiaoBdLu%2Fagents.png?alt=media)

Run purpose-built AI agents that analyze your architecture and produce structured, repeatable outputs, right from the AI Assistant.

The Agents extend the [AI Assistant](https://help.ardoq.com/en/articles/510373-omnipresent-ai-assistant) beyond ad-hoc questions and answers. An Agent runs a predefined, multi-step procedure on your behalf.

**Used for:** Each agent is purpose-built for a specific enterprise architecture task, so you get consistent, repeatable results every time you run it.

**Learn more:**

* [KB Article: AI Custom Agents](https://help.ardoq.com/en/articles/721047-ai-custom-agents)
* [KB Article: AI Assistant](https://help.ardoq.com/en/articles/510373-omnipresent-ai-assistant)

### AI-Assisted Modeling

These capabilities help teams move from blank pages to usable architecture models faster.

#### AI-Generated Capability Maps (GA)

![](https://2484004798-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FnlnB7uZ29iLVLzqGbCym%2Fuploads%2FKVsFBuefO8hrrD0l1vuD%2Fcapmaps.gif?alt=media)

Automatically generate multi-level business capability maps based on your organization's context, industry, or imported data.

**Used for:** Capability-based planning, heatmaps, cost and risk analysis.

**Learn more:** [Blog: New Ardoq AI Features To Automate Value Streams And Capability Modeling](https://www.ardoq.com/blog/new-ardoq-ai-features-to-automate-value-streams-and-capability-modelling)

#### AI-Generated Business Value Streams (and AI linking to Value Streams) (GA)

![](https://2484004798-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FnlnB7uZ29iLVLzqGbCym%2Fuploads%2F8NQG3paHWkGPht2e2O8w%2Fvaluestreams.gif?alt=media)

Create draft business value streams in seconds, providing a structured starting point for refinement and alignment. Automatically suggest and create links between architecture elements (applications, capabilities, initiatives) and business value stream stages.

**Used for:** Strategy-to-execution traceability, transformation planning

**Learn more:** [Blog: AI for Business Value Streams](https://www.ardoq.com/blog/new-ardoq-ai-features-to-automate-value-streams-and-capability-modelling)

#### AI-Generated Business Value Propositions (GA)

![](https://2484004798-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FnlnB7uZ29iLVLzqGbCym%2Fuploads%2FX32yOP9PEeTgqJ3v0QlB%2Fvalueprops.gif?alt=media)

Generate draft business value propositions connected to capabilities, value streams, and initiatives.

**Used for:** Framing business outcomes, stakeholder communication, value realization discussions

**Learn more:** [Blog: AI for Business Value Streams](https://www.ardoq.com/blog/new-ardoq-ai-features-to-automate-value-streams-and-capability-modelling)

#### AI Process Modeling (ShiftX) (GA)

![](https://2484004798-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FnlnB7uZ29iLVLzqGbCym%2Fuploads%2F9gUs7mCHIiKiVCNnJ0we%2Fshiftx.gif?alt=media)

Convert text descriptions or uploaded images into structured, connected process models. Refine existing process models using natural language instructions, such as adding steps, changing flows, or clarifying logic.

**Used for:** Process documentation, operational analysis, linking processes to applications

**Learn more:**

* [Help Center Article: Leverage AI to generate process drafts in Ardoq ShiftX](https://help.ardoq.com/en/articles/407282-leverage-ai-to-generate-process-drafts-in-ardoq-shiftx)
* [Blog: Conversational AI in Ardoq ShiftX: EA & Process in Collaboration](https://www.ardoq.com/blog/conversational-ai-ardoq-shiftx)
* [Smarter Process Modelling with Ardoq ShiftX](https://www.ardoq.com/features/shiftx)

#### AI Visual Importer (GA)

Upload diagrams, whiteboards, or slide screenshots and automatically convert them into structured architecture data.

**Used for:** Migrating legacy diagrams, accelerating documentation

**Learn more:**

* [Blog: Ardoq AI Visual Importer: From Visuals to Structured Data, Instantly](https://www.ardoq.com/blog/ardoq-ai-visual-importer)
* [Blog: Q4 2025 Ardoq AI Roundup: Governing, Guiding & Growing the AI-powered Enterprise](https://www.ardoq.com/blog/q4-2025-ardoq-ai-roundup)
* [KB Article: Getting Started with AI visual importer](https://help.ardoq.com/en/articles/331048-lucidchart-visual-importer)
* [Lucidchart + Ardoq: Your Diagrams, Now Powered by Live EA](https://www.ardoq.com/integrations/lucidchart)
* [Demo video](http://youtube.com/watch?v=yJUuUEw3J2Q\&feature=youtu.be)

### AI-Powered Querying & Insights

These capabilities make architecture insights accessible through natural language.

#### AI Query Builder (GA)

![](https://2484004798-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FnlnB7uZ29iLVLzqGbCym%2Fuploads%2FIu9Fa1XehEAEmj1VawCi%2Fquerybuilder.png?alt=media)

Automatically generates custom data queries based on natural language input, which users can then refine.

**Used for:** Faster reporting, advanced analysis without query expertise

**Learn more:**

* [KB Article: Getting Started with AI Query Builder](https://help.ardoq.com/en/articles/528056-getting-started-with-ai-query-builder)
* [Product Portal](https://portal.productboard.com/zvnzutc4rg7xnc7pp8uiz7s7/c/541-ai-initiative-use-natural-language-to-create-custom-queries-for-reports)

#### AI-Powered Viewpoint Creation (GA)

![](https://2484004798-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FnlnB7uZ29iLVLzqGbCym%2Fuploads%2F5QiPc9gachUh8i3sF6Xu%2Fviewpoint.gif?alt=media)

Generate draft viewpoints based on natural language prompts, automatically selecting relevant components, relationships, and perspectives.

**Used for:** Rapid stakeholder views, impact analysis, executive-ready visuals

**Learn more:** [Help Center Article: How to create viewpoints using AI](https://help.ardoq.com/en/articles/171731-how-to-create-viewpoints-using-ai)

#### AI-Assisted View Configuration (Beta)

![](https://2484004798-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FnlnB7uZ29iLVLzqGbCym%2Fuploads%2FlDBR1ccngAoFZ2850qON%2Fviewconfig.gif?alt=media)

Apply conditional formatting and labels directly in a view by describing what you want to see. For example, color applications nearing end of life, or show owner and lifecycle status on each application.

**Used for:** Highlighting important data, surfacing key properties, and making complex visualizations easier to interpret.

**Learn more:** [KB Article: AI-Assisted View Configuration (Beta)](https://help.ardoq.com/en/articles/718246-ai-assisted-view-configuration-beta)

#### AI-Powered Viewpoint & Report Descriptions (GA)

![](https://2484004798-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FnlnB7uZ29iLVLzqGbCym%2Fuploads%2FjzPJ4uRepIP1OG0WVHMs%2Fdescriptions.gif?alt=media)

Automatically generate clear, consistent descriptions for viewpoints and reports, grounded in the underlying architecture data.

**Used for:** Documentation, stakeholder communication, consistent reporting

**Learn more:** [KB Article: AI-Generated Report & Viewpoint Descriptions](https://help.ardoq.com/en/articles/407079-ai-generated-report-viewpoint-descriptions)

#### AI Reference Creation with Retrieval Augmented Generation (RAG)

Automatically suggests relationships between components (for example, applications and capabilities) using AI, based on existing architecture data and context.

**Used for:** Faster modeling, reduced manual linking, and more complete architecture views without deep EA expertise.

**Learn more:** [KB Article: AI-based reference creation in Ardoq with simple Retrieval Augmented Generation](https://help.ardoq.com/en/articles/127002-ai-based-reference-creation-in-ardoq-with-simple-retrieval-augmented-generation)

### AI for Knowledge Access & Enablement

These capabilities help users learn, navigate, and adopt Ardoq more effectively.

#### Aria: Ardoq's Technical Support AI Agent (GA)

![](https://2484004798-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FnlnB7uZ29iLVLzqGbCym%2Fuploads%2FsLtHddhKuJwO73p9bSZw%2Faria.png?alt=media)

**Aria** is **Ardoq's Technical Support AI Agent**, designed to help you get quick answers and guidance while using Ardoq.

You can use Aria to:

* Ask questions about how to use Ardoq
* Get guidance on enterprise architecture best practices
* Understand modeling approaches and recommended patterns
* Find relevant help articles and documentation without manually searching the Help Center

Aria is available directly in the Ardoq platform through the Technical Support Chat/Messenger and is always there when you need help.

**Used for:** User onboarding, self-service support, modeling guidance

**Learn more:**

* [KB article: Aria: Ardoq's Technical Support AI Agent](https://help.ardoq.com/en/articles/545473-aria-ardoq-s-technical-support-ai-agent)
* [KB Article: Getting EA Best Practices Guidance with Aria](https://help.ardoq.com/en/articles/532569-getting-enterprise-architecture-best-practices-guidance-with-aria-ardoq-s-technical-support-ai-agent)

#### AI Semantic Search (GA)

![](https://2484004798-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FnlnB7uZ29iLVLzqGbCym%2Fuploads%2FLrc0BXCUR7HO1WFAlqA8%2Fsemantic.png?alt=media)

Intent-based discovery, allowing you to find insights using natural language, matching the meaning of your query rather than just exact words.

Available to all organizations with AI features enabled. Admins turn it on under **Organization settings > Search settings**. Users can also take a search straight to the AI Assistant from the search bar. See [Semantic quick search](https://help.ardoq.com/en/articles/682139-semantic-quick-search) for the full guide.

**Used for:** Rapid data finding, asking architecture questions

**Learn more:** [Product Board Tile](https://portal.productboard.com/cjkps8hu4ybuwz6h9kzrumxp/c/535-search-get-search-results-based-on-meaning-and-intent-not-just-keywords)

### AI Data Enrichment & Documentation

These capabilities improve data quality, consistency, and usability.

#### AI-Powered Component Descriptions (GA)

![](https://2484004798-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FnlnB7uZ29iLVLzqGbCym%2Fuploads%2FWSDG3tYqym2CgGE0HN9V%2Fcompdesc.gif?alt=media)

Automatically generate consistent, structured descriptions for architecture components based on their relationships.

**Used for:** Documentation, portfolio reviews, stakeholder communication

**Learn more:** [KB Article: AI-Powered Component Descriptions](https://help.ardoq.com/en/articles/150511-ai-powered-component-descriptions)

#### AI-Assisted Surveys (GA)

Use AI to help generate survey questions and structure responses into usable architecture data.

**Used for:** Data collection from application owners and business stakeholders

**Learn more:**

* [KB Article: Create references from apps to business or tech capabilities in Surveys](https://help.ardoq.com/en/articles/212247-create-references-from-applications-to-business-or-technical-capabilities-in-surveys-with-ai)
* [KB Article: How to Automatically Generate Component Descriptions in Surveys](https://help.ardoq.com/en/articles/150511-ai-powered-component-descriptions#h_7b77572340)

#### Foundation Insights Agent (Open Beta)

![](https://2484004798-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FnlnB7uZ29iLVLzqGbCym%2Fuploads%2FI2z95RmlZCVd5io5Gdvc%2Finsights.gif?alt=media)

Foundations Insights helps EAs notice important data patterns sooner, such as identifying hidden risks or inconsistencies, so teams can have better conversations and make better choices.

**Used for:** Data discrepancy flagging, proactive assessment and improved architecture quality.

**Learn more:** [Introducing Foundation Insight Agents](https://help.ardoq.com/en/articles/619112-introducing-ardoq-foundation-insights-agents)

### AI Integrations & External Access

These capabilities allow secure, governed AI access to Ardoq data from external tools.

#### AI Gateway (Model Context Protocol, MCP) (GA)

![](https://2484004798-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FnlnB7uZ29iLVLzqGbCym%2Fuploads%2Fmeunnx8Gz766zGQBiADp%2Fmcp.gif?alt=media)

Provides **read-only**, permission-aware access to live Ardoq architecture data for external AI tools such as Microsoft Copilot or Claude.

**Key characteristics:**

* Read-only access
* Fully permission-aware
* Supports both API-token and OAuth authentication
* Grounded in live architecture data
* No autonomous actions

**Used for:** Asking architecture questions from enterprise AI tools

**Learn more:**

* [Blog: Ardoq First to Market With MCP GA: Ask, Discover, and Act Through AI](https://www.ardoq.com/blog/announcing-mcp-ga)
* [KB Article: Using Ardoq MCP Server](https://help.ardoq.com/en/articles/415375-using-the-ardoq-mcp-server)
* [KB Article: Context Engineering: Getting the most from Ardoq MCP](https://help.ardoq.com/en/articles/413191-context-engineering-getting-the-most-from-ardoq-mcp)
* [KB Article: MCP SSO / OAuth](https://help.ardoq.com/en/articles/676638-mcp-single-sign-on)

#### Prompt Engineering Playbook for Enterprise Architects

A practical guide to writing effective, context-aware prompts when using Ardoq AI and external AI tools (such as Copilot or Claude) with architecture data. The playbook covers best practices, example prompts, and governance considerations to help teams get consistent, reliable results from AI, without compromising control or trust.

**Learn more:** [Prompt Engineering Playbook (Guide)](https://content.ardoq.com/ai-prompting-guide)

### AI Trust, Governance & Controls

#### AI Lens (GA)

![](https://2484004798-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FnlnB7uZ29iLVLzqGbCym%2Fuploads%2FAPVCGB3GOZIqlX0w5uzD%2Failens.gif?alt=media)

AI Lens gives organizations a clear, enterprise-wide view of how AI is used, from embedded and SaaS AI to emerging AI agents.

Built on Enterprise Architecture, it helps teams govern AI, manage risk and ownership, and stay compliant using a single control plane.

With pre-built EU AI Act assets and the flexibility to support any regulation or internal policy, AI Lens enables confident, compliant AI adoption.

**Used for:** AI governance, EU AI Act readiness, risk and compliance assessments, executive reporting.

**Learn more:**

* [AI Lens Demo](https://www.youtube.com/playlist?list=PLN64rGvDdWmnc0itkqd7WbRb3_V0LkcFm)
* [Blog: AI Governance Dashboards Every EA Should Use](https://www.ardoq.com/blog/ai-governance-dashboards-eas)
* [Web page](https://www.ardoq.com/solutions/ai-lens)
* [KB Article: How to Manage AI Innovation with Ardoq](https://help.ardoq.com/en/articles/143599-how-to-manage-ai-innovation-with-ardoq)
* [KB Article: Introducing AI Lens](https://help.ardoq.com/en/articles/455687-introduction-to-ai-lens)
* [Webinar: Governing the AI Surface Area, From Agents to Architecture](https://www.youtube.com/watch?v=TvlEXsGx6As)
* [AI Lens: Enterprise AI Governance Purpose, Scope and Rationale](https://help.ardoq.com/en/articles/442813-ai-lens-enterprise-ai-governance-purpose-scope-and-rationale)
* [AI Lens: Enterprise AI Management: Getting Started guide](https://help.ardoq.com/en/articles/440929-getting-started-with-ai-lens-enterprise-ai-management)
* [AI Lens: Enterprise AI Management: Metamodel](https://help.ardoq.com/en/articles/440631-ai-lens-enterprise-ai-management-metamodel)
* [AI Lens: Enterprise AI Management: Purpose, Scope & Rationale](https://help.ardoq.com/en/articles/442091-ai-lens-enterprise-ai-management-purpose-scope-and-rationale)
* [AI Lens: Enterprise AI Governance: Metamodel](https://help.ardoq.com/en/articles/442811-ai-lens-enterprise-ai-governance-metamodel)
* [AI Lens: Enterprise AI Governance: Purpose, Scope & Rationale](https://help.ardoq.com/en/articles/442813-ai-lens-enterprise-ai-governance-purpose-scope-and-rationale)
* [AI Lens: Enterprise AI Governance: Getting Started guide](https://help.ardoq.com/en/articles/442961-getting-started-with-ai-lens-enterprise-ai-governance)
* [How to Leverage Ardoq to Support the EU AI Act](https://help.ardoq.com/en/articles/448393-how-to-use-ardoq-to-track-compliance-with-the-eu-ai-act)

#### Ardoq AI Security FAQs

Ardoq's AI capabilities are designed to support enterprise governance and regulatory requirements.

#### Ardoq AI Principles

We're building deeply embedded, context-aware AI that helps EAs model change, predict outcomes, and guide business decisions. This isn't AI for show. It's AI for workflow.

#### Permission-Aware AI

AI respects Ardoq's role-based access controls and data permissions at all times.

### Frequently Asked Questions

#### How do I enable AI features in my Ardoq instance?

Admins can enable AI features via Preferences > Organization Settings > Features.

If this is not available, AI must first be toggled on for your organization. Please contact your **Customer Success Manager (CSM)**, who can enable AI for your Ardoq instance. AI features enabled work at the organizational level, meaning all eligible users will have access to them once activated.

#### Can I choose not to use AI features in Ardoq?

Yes. If you prefer a **no-AI experience**, you can ask your CSM to ensure AI features remain disabled for your instance.

Ardoq fully supports customers who choose not to enable AI and provides the same core EA platform functionality without AI assistance.

#### Which AI services power Ardoq AI features?

Ardoq uses a combination of **leading large language models (LLMs)** and proprietary technology.

Key points:

* Ardoq does **not** train AI models on customer data
* Customer data is processed only to generate responses within your environment
* AI responses are grounded in Ardoq's **graph-based architecture data**

Specific model providers may evolve over time as Ardoq continues to adopt best-in-class AI technology.

#### Does Ardoq AI modify my data automatically?

No. Ardoq's AI features generate **suggestions, drafts, and recommendations**. All changes must be explicitly reviewed and approved by a user before being saved.

Ardoq AI operates in a **controlled, non-destructive way**.

AI-generated outputs are created inside a **separate Scenario**, not directly in the live model. Users must **review, validate, and deliberately merge** any AI-generated content before it is saved to production data.

#### Is Ardoq AI permission-aware?

Yes. All AI features respect Ardoq's existing **role-based access controls**.

AI cannot surface or infer information that a user does not already have permission to access.

#### Is Ardoq AI Gateway (MCP Server) read-write or read-only?

While Ardoq's Model Context Protocol (MCP) implementation does not write directly to your live production workspace, it is not strictly read-only. It features a data population tool that allows AI clients to create components and references using natural language. To ensure safety, this data is staged in an Ardoq Scenario where you can review, validate, and merge the changes when you are ready.

#### Can I build AI agents using the Ardoq AI Gateway (MCP Server)?

**Yes.** Ardoq's MCP Server enables you to build **context-aware AI agents** that can reason over your architecture data and use Ardoq as a trusted source of truth.

Using MCP, AI agents can:

* Query architecture, application, capability, and dependency data
* Reason across relationships (for example, impact analysis, capability gaps, risk hotspots)
* Combine Ardoq context with other enterprise data sources to support multi-step analysis and decision-making

Today, MCP access is **read-only**, meaning agents can retrieve and reason over Ardoq data but **cannot directly modify it**. We are actively working on read/write capabilities.

When write operations are introduced, they will follow Ardoq's **human-in-the-loop model**. Any proposed changes will be generated in a **separate Scenario** and require **explicit user review and manual merge** before affecting live architecture data.

#### Is my data used to train AI models?

No. Customer data is **never used to train external AI models**.

Your data remains within your Ardoq environment and is used only to generate responses at query time.

#### Where can I learn more about Ardoq's AI governance and security?

You can review Ardoq's AI governance approach in the following resources:

* [Ardoq Information Security Documentation](https://www.ardoq.com/information-security)
* [Ardoq AI Principles](https://www.ardoq.com/ai-and-innovation/principles)