The Digital Transformation Blog | Ardoq

The AI-First Ardoq Has Arrived: A Letter from Our Founder & CEO

Written by Erik Bakstad | May 28, 2026 8:01:35 AM

Today, we're relaunching Ardoq as the AI-first EA platform where AI automates thousands of hours of EA work, and every AI output is grounded in your living architecture graph.

We launched Ardoq in 2013 with a simple but ambitious mission: help big companies move like small companies by democratizing enterprise architecture (EA) insights. Back then, EA was locked in the hands of a few specialists, buried in static documents that were outdated on arrival. We believed that if every architect and business leader could see the same dynamic, connected picture of their organization, they'd make faster, smarter, data-backed decisions. We built proprietary graph technology into Ardoq to make that possible. And it worked.

In the AI era, that mission is more critical than ever.

The recent rise of vibe coding is the clearest signal of what's changed. The ability to generate rich, custom outputs conversationally is genuinely magical. It opens up an enormous space of possibility for more people to build, analyze, and create than ever before. When the cost of generating content, code, or a business case drops toward zero, the volume of what gets produced explodes.

But here's the problem nobody is talking about loudly enough: when the cost of generating a recommendation goes to zero, the cost of a wrong one doesn't.

Vibe coding produces outputs that are delivered in a confident tone and deliver the illusion of accuracy. Underneath, there can be bugs, edge cases, hallucinations, and data leaks that nobody caught because nobody knew to look. For building a prototype or drafting a one-off analysis, that's a reasonable trade-off. For deciding which mission critical applications to decommission, which technology investments to make, or how to integrate an acquisition — it isn't. A beautiful architecture diagram that maps the wrong dependencies doesn't just look wrong. It sends a $50M transformation in the wrong direction, with everyone's sign-off on it.

This is the problem Ardoq was built to solve. And AI makes solving it more urgent, and more possible, than ever before.

Graph as our Differentiator

Ardoq's proprietary architecture graph has always been different. It's live, governed, and schema-aware. It knows your organization's structure, not just the surface of it.

When AI reasons on top of that, it doesn't guess. It knows what's real, what changed last Tuesday, and what will break if you make the move you're considering. Every output comes with a traceable reasoning chain — measurable at the board level. That's not a product feature. It's a structural advantage that no amount of prompt engineering on a stale export can replicate.

Ardoq is the AI-First Platform For Enterprise Architects


We're not adding AI features to a legacy tool that was once on-premise and pushed into the cloud. Every Ardoq AI feature is grounded in your living architecture model. Every insight, every recommendation, every action an AI agent takes in Ardoq comes with a source, a traceable reasoning chain, and a clear owner.

This is what AI-first means at Ardoq. Not AI bolted on, but AI is built into the foundation, reasoning on the most accurate picture of your enterprise that exists anywhere.

Ardoq AI Announcements For Spring 2026

We're announcing a new generation of Ardoq AI capabilities that are built on the graph, governed by design, and ready for the decisions that actually matter.

Omnipresent AI Assistant — GA

Ask anything about your architecture and get an instant, accurate, contextual answer from wherever you are in Ardoq. This isn't a search bar. It's a conversational interface that reasons across your entire data model in real time. No exports, no prompting tricks, no hoping the AI has the right context. Just answers you can stand behind.

AI Import Builder — GA

Connect any third-party data source to Ardoq in minutes — no technical setup required. This is an AI-powered capability that reads third-party service documentation and automatically configures a connection to import data into Ardoq. You can use this as a self-service workflow; provide authentication credentials, and Ardoq's AI agents handle the configuration. This will eventually evolve into a shared connection library (that respects customers’ privacy) — any successfully configured connection can be verified and reused across the Ardoq customer community.

Custom Agents — Open Beta

Build and deploy your own AI agents in Ardoq, scoped to your architecture, your metamodel, and your workflows. This is the first time any EA platform has given customers the ability to create purpose-built agents that run on their own live architecture data. We're not building one agent and calling it done. We're building a platform where your architects become AI builders — agents that know your environment, not the internet's best guess at it. This will roll out to our Partners in Q2 and continue on to customers thereafter.

Foundation Insights Agent — GA

Already in early access and getting strong adoption, this agent automatically detects data quality issues, gaps, and risks in your architecture — and builds you a prioritized action plan. The architects using it today are spending less time hunting for problems and more time fixing them.

AI Query Builder — GA

Already in early access, this new feature lets users create powerful queries without having to learn complex syntax. Simply type what you want to find in plain English, and Ardoq automatically builds the matching Advanced Search, ready to run, review, or save as a report.

MCP Single Sign On — Technical Preview

Connect Ardoq securely to any AI tool in your stack using industry-standard authentication. One approved connection. No security exceptions needed. Enterprise IT gets the governance. Your architects get the flexibility.

Agents for OOTB Solutions — Technical Preview

We're embedding purpose-built AI agents across Ardoq's out-of-the-box solutions — automating the analysis, mapping, and decision support that used to take your team days. Value stream mapping, application rationalization, disaster recovery planning — the work that's been manual for too long is starting to run itself.

Data Ingestion Agent — Technical Preview

Pulls data from documents, contracts, and existing sources and maps it straight into Ardoq automatically. Whether it's a software contract PDF, a spreadsheet from a legacy system, or documentation from an acquisition — it goes into your model, not your inbox. You will also be able to use the AI Assistant and the MCP server to safely populate content into Ardoq.

AI Web Search — Technical Preview

Your architecture should reflect what's actually happening in the world, not what was true last year. This agent pulls current, real-world information from the web and brings it into your architecture work — so your models stay relevant without extra manual effort.

AI Semantic Search — GA

Finding the right asset used to require knowing exact names, IDs, or fields—slowing non-experts and hiding value behind insider terms. AI-powered Semantic Search lets you ask in your own words and instantly understands you, getting to the right asset in seconds with a more context-aware navigation system.

What This Means For Our Customers


We expect that 40% of routine EA tasks can now be automated by Ardoq AI. That number moves the needle, but what matters more is what your team does with the time they get back.

Right now, too many architects are maintaining the map. Updating records, chasing data owners, answering the same questions about what's running and what it costs. That work is necessary — but it's not why you hired an EA team. When AI handles the maintenance, your architects handle the judgment. They're in the room when the $50M transformation decision gets made, not catching up on documentation afterward.

We've seen this play out in real customers. At Tenneco, Ardoq is delivering a 292% ROI — not from a demo, but from a live deployment that gives their team an architecture model they could actually trust and act on.

That's the outcome we're building toward for every customer.

What We Won't Compromise On

We've thought carefully about how to build AI that enterprise teams can actually trust. A few commitments we're making publicly:

  • Every AI answer comes with a source. We don't surface recommendations without showing the reasoning chain behind them. If an insight can't be traced back to your architecture data, it doesn't ship.
  • Governance is built in, not bolted on. Ardoq's AI is permission-aware. It only reasons on data your users are authorized to see. Human approval is built into every agent action that matters.
  • We are ecosystem-neutral and flexible by design. Ardoq is not bound to a single AI vendor or tech stack. We select the best model for the task, optimize for cost, latency, and privacy, and support privacy-controlled AI access. Our AI architecture is designed to adapt as the AI market evolves. We also make sure our AI works regardless of how you customize your metamodels and Ardoq fields.

We anchor everything to the graph. This is the one we're most firm about. Generic AI guesses. Ardoq knows. The difference is the graph and we're not going to let that connection drift.

This is Just the Beginning

The capabilities we're announcing today are the foundation of something bigger. We're actively expanding what agents can do, deepening the integration between AI and every part of the Ardoq platform, and investing in making the graph even richer — so AI has even more to reason on.

If you're an existing customer, you'll start seeing these capabilities in your product over the coming weeks. If you're evaluating EA platforms for the first time, there's never been a better moment to see what a live, AI-powered architecture graph can do for your organization.

We built Ardoq to make enterprise architects more powerful. Thirteen years in, that mission hasn't changed. The tools just got a lot better.