Behind the Vision: How We Built AI into ADOIT

Adding AI without a defined purpose does not create value. Features that are bolted onto a tool for novelty can easily add complexity instead of reducing it. Our approach in ADOIT is different. AI in ADOIT is seamlessly integrated into the workflows of enterprise architecture, bringing smart functionality to modelling, planning and analysis, and ensuring every capability addresses the actual needs of architects.

This principle is guided by architecture thinking. We begin with the problem, examine the data and compliance requirements, and then decide where AI can shorten a task or raise the quality of the outcome. Each feature is measured against its impact on actual work. Architects remain in charge, and AI works in the background to remove obstacles and speed up the progress.

Our roadmap for AI in ADOIT began with listening to enterprise architects. Across organizations, three recurring challenges stood out.

1. Architects work in highly complex environments. Business units run on different systems, many of them legacy, some cloud-based, others supplied by vendors. The result is a network of dependencies and integration issues that demands constant attention.

2. Architects also face growing time pressure. New business requirements arrive quickly, and teams must adapt while maintaining stability and governance. Balancing innovation with control has become one of the hardest parts of the role.

3. Finally, there is a need for guidance. EA tools are powerful, but their breadth can be overwhelming. With so many methods and frameworks available, even experienced architects look for direction on where to start and which practice fits best.

These pain points shaped our design decisions. Instead of creating generic assistants, we identified use cases with the greatest impact: requirement suggestions, roadmap planning, modelling support, conversational queries and end-of-life lookup. Each capability is built to help architects make progress faster, providing guidance that complements their expertise.

The value of AI in ADOIT comes from a set of principles we apply to every feature. These principles ensure that the technology helps architects in practice and produces results they can trust.

Embedded in workflows
AI is available inside the modelling, planning and analysis functions that architects already use. There is no need to switch tools or copy information into a separate environment.

Driven by repository data
Insights are based on the content stored in the architecture repository. This keeps results relevant to the organization’s actual portfolio rather than generic outputs from a language model.

Hint: Is your repository growing out of control? See how ADOIT Forms can help you keep it clean and well-governed.

Transparent and controllable
Each suggestion shows its source and can be reviewed before use. Architects always decide what to apply, adapt or dismiss.

Supporting the expert
AI highlights options and patterns but does not take decisions on its own. It acts as an assistant that extends the architect’s view, while responsibility for choices stays with the user.

The ideas behind our approach are already translated into working features. ADOIT includes AI capabilities that are active in the tool today, ready to be used.

In ADOIT, we use AI to help Enterprise Architects manage complex application portfolios more efficiently. Instead of manually scoring and assessing each application, AI automates evaluations, learns from past investment decisions, and provides probability-based recommendations. This not only accelerates portfolio analysis but also ensures transparency, while leaving the final choice in the architect’s hands.

During modelling, ADOIT recommends elements and relationships that fit the context of the diagram. This ensures greater consistency, reduces errors and helps diagrams reach a useful state faster.

Stakeholders can type questions in plain language and receive answers directly from the repository. This lowers the barrier to engaging with architecture content and makes insights available to people who are not daily users of the tool.

ADOIT automatically surfaces end of life information for applications and technologies. Architects see which components are nearing retirement and can address risks before they grow into larger problems.

These features have one thing in common. They provide guidance and suggestions at the point of need, while leaving the final decisions with the architect in charge.

What is live in ADOIT today marks only the beginning. Our roadmap builds on these foundations and will extend AI into areas where architects see the greatest need. The goal is not to add more features for their own sake, but to give teams support that makes planning and decision-making more effective.

• Evolving prioritization models to highlight where resources should be directed first

• Expanded pattern recognition to detect recurring structures and suggest reuse

• Deeper GenAI integrations in selected areas, supported by governance and compliance safeguards

• Enhanced capability mapping connected more tightly with initiatives, applications and investments

These upcoming enhancements will deepen the role of AI in enterprise architecture without changing the principle that defines it in ADOIT: features that guide, suggest and accelerate.

The AI in ADOIT is not designed to impress with novelty. It is shaped by the challenges that architects face every day and built to remove friction where it matters most. This focus on real needs is what gives the features lasting value.

Instead of following hype, we have made AI a working part of the tool. It appears in the same places where decisions are prepared and models are built, offering suggestions that support the actual work. The result is a platform that strengthens practice rather than distracting from it.

This approach defines the path forward. AI in ADOIT is built to extend the practice of EA, giving teams support they can rely on as they manage complexity and plan change.