How IA and Architecture Meet: What 'AI by Design' Requires From Information Structure

On March 7th, we mark World Information Architecture Day, a global moment dedicated to how information is structured, connected, and made meaningful.

At Mercedes-Benz.io, this conversation comes at a pivotal moment. As our products become increasingly intelligent, Information Architecture (IA) is the foundation that enables AI systems to behave predictably, understand context, and earn trust.

To celebrate this day, we brought together three MB.ioneers: Wolfgang Hermann, Cláudio Medina, and Moritz Blum, to explore how Information Architecture (IA) and Architecture meet, and what "AI by Design" requires from information structure. Their perspectives reveal one core truth: in an AI powered era, meaning must be designed before intelligence can be delivered.

Information Architecture as the Foundation for Machine Understanding

As AI-driven experiences become the default across products and workflows, the role of Information Architecture evolves from a human-centred guidance system into a critical layer of machine infrastructure.

For Wolfgang, Information Architecture becomes "the operating system for meaning", and without structure, AI turns into a generic suggestion box. With it, AI becomes a knowledge amplifier that works in alignment with organisational goals and user needs.

Cláudio reinforces that AI does not eliminate the need for information structure but amplifies it. High-quality metadata, consistent vocabulary, and explicit semantic relationships prevent hallucinations and anchor AI responses in real time. Information Architecture gives AI the clarity it needs to locate context, interpret intent, and produce grounded outputs.

Moritz frames Information Architecture as a contract between what users need and how AI behaves. If the structure doesn't define user tasks, the AI will guess and guessing does not lead to clarity, but to cognitive friction.

What Breaks When Information Architecture Is Missing: Trust, Truth, and Orientation

Poor Information Architecture doesn't just create messy interfaces; it destabilises user confidence and system behaviour:

• Moritz sees orientation as the earliest failure point. Before trust breaks, users lose their sense of place: What changed? Where am I? What can I do next?
• For Wolfgang, the first casualty is trust. Without structure, AI outputs become inconsistent and unpredictable, and users stop believing what they see.
• Cláudio highlights hallucinations as the natural outcome of semantic ambiguity. When AI lacks clear meaning, it fills gaps with assumptions, leading to misinformation.

Once orientation collapses, no amount of AI sophistication can restore confidence.

How Architecture Makes AI Understandable and Human

AI feels human-centred not because of its intelligence, but because of its predictability.

Cláudio highlights transparency as a design requirement. Information Architecture enables AI to reference sources, explain its logic, and follow consistent fallback paths. Structure turns AI from a black box into a traceable collaborator.

Wolfgang emphasises the role of structural thinking: templated outputs, clear workflows, and curated context create a predictable behavioural pattern – and predictability builds trust.

Moritz adds that architecture grounds AI in the visible interface state. When AI suggestions align with what's already on screen, users can see the "why" behind a recommendation. Without structure, AI becomes magical – and magic is not trustworthy.

Designing for Meaning: From Content to Context and Intent

"Designing for meaning" is one of the most important principles in modern Information Architecture and one of the most misunderstood.

Moritz sees meaning through the lens of user tasks. If a piece of content doesn't support a decision or action, it's decorative noise. Clarity emerges when everything in the system has a job.

For Wolfgang, meaning is curated, not simply organised. It begins with understanding exactly what a specific role needs in a specific moment.

Cláudio defines meaning as structuring relationships and intent, not just objects. Information Architecture must articulate why something exists and how it connects to everything else.

Where AI Supports IA: Making Complexity Navigable

When Information Architecture sets the structure, AI becomes a guide rather than a guessing engine.

Wolfgang explains that structured information allows AI to apply consistent logic, turning complex workflows into navigable experiences.

Cláudio describes the partnership between the two: Information Architecture provides the map, while AI helps users interpret it in real time. Instead of forcing people to navigate complexity alone, AI becomes an adaptive layer that reduces cognitive load.

Moritz adds that AI shines when it's grounded in structure. It surfaces the right thing at the right moment but only because Information Architecture has defined what "right" means in the first place.

IA Principles That Matter More in an AI Era

AI does not replace Information Architecture; it raises the bar for it, where these three principles become essential:

• Curation over indexing: Exhaustive information is not meaningful, but curated context is.
• Semantic clarity: Ambiguity is manageable for humans, but dangerous for machines.
• Transparent state management: AI-driven changes must always be visible, reversible, and understandable.

Small Structural Changes, Big System Impact

Even small Information Architecture improvements can dramatically improve AI systems.

Cláudio advocates for standardised vocabulary. When different teams label the same concept differently, AI cannot reliably understand or retrieve information.

Moritz champions task-based labels. Naming things by what users want to do (not by internal feature names) improves clarity without changing interaction patterns.

Wolfgang highlights defining role accountabilities. This single decision unlocks clarity around required knowledge, workflows, and expected outputs, becoming the blueprint for AI behaviour.

Persistent Information Architecture Misconceptions

Despite its growing importance, IA still suffers from common misconceptions:

• Information Architecture is not just navigation or taxonomy. It is the discipline of making complexity intelligible.
• Search cannot compensate for poor structure. Even AI-powered search fails without semantic foundations.
• IA is not documentation. It is the decision-making logic behind what appears where, when, and why.

Information Architecture Is Becoming the Architecture of AI

At Mercedes-Benz.io, as we build intelligent products, Information Architecture is no longer an optional design layer, it is the structural foundation that makes AI explainable, contextual, and trustworthy.

In an era where systems think, Information Architecture provides:

• the meaning
• the guardrails
• the transparency
• the predictability
• the human-centred logic

that transform machine intelligence into meaningful experiences.

AI may be powerful, but without Information Architecture, it has no understanding. With Information Architecture, it becomes a collaborator.

Although AI possesses significant capabilities, it lacks comprehension unless paired with Information Architecture. When combined with Information Architecture, AI transforms into a cooperative partner.