The no longer ignorable role of Information Architecture in AI

The essential role of optimal information is too often underestimated in digital transformations. The focus is mostly on technology and tools. That's a problem, because all those tools and technology really only do one thing: pump around information. If that information is bad, no cool (AI) tool can fix it.

If the data in your organisation is fragmented, outdated or full of inconsistencies, artificial intelligence (AI) does not solve the problem; it is then just a faster way to generate reliable crap. Without reliable data, no organisation or digital transformation can be successful.

Architecture is also about information

For digital transformation to succeed, an (enterprise) architecture is crucial. Enterprise Architecture (EA) is often the domain of IT architects, who focus on applications, technology and sometimes business processes.

However, a crucial element is missing in many organisations: a well-established Information Architecture (IA). This neglect has always been a risk, but with the rise of (generative) AI, the risk has now increased exponentially. Without IA, your investments in AI are likely to lead to disappointment - and the infamous “garbage in, garbage out” scenario.

Read my earlier blog on Enterprise architecture

A tool is just a tool

We see it often: organisations invest in the latest AI tools, hoping for revolutionary insights, but run into problems. The data is fragmented, inconsistent, unstructured and simply not usable. This is a direct result of the lack of a solid Information Architecture.

Information architecture involves the organisation, structuring and labeling of both data and content so that they can be effectively found, understood and used. It includes defining:

• Data models - How data is defined and related.

• Content typologies - What types of content there are and how they are classified.

• Taxonomies and metadata - How information is labeled and categorised for discoverability, reuse and consistency.

• Data governance - Policies and processes to ensure data quality, security and compliance.

• Content governance - Policies and processes to keep content relevant, up-to-date and usable.

In short, information architecture ensures that the right information, at the right time, is available to the right person and in the optimal structure and form for the interface with which that person is communicating. It is the basis for informed decisions, efficient processes and an optimal customer experience.

IA as a unifying factor in enterprise architecture

A successful EA is not a collection of individual components, but a coherent whole. The IA plays a connecting role in this, between the:

• Business/Process architecture - What information is needed to support business processes and achieve strategic goals?

• Information architecture - What structure and semantics of information is needed, what are the information sources and how do the information flows?

• Application architecture - What applications manage what information, and how do they integrate with each other?

• Technology architecture - What technologies are used to store, process and distribute the information?

Without this connection, a silo structure is created, where data and content are locked in departments, experts' heads and systems. This leads to inefficiencies, errors and missed opportunities.

AI reinforces the need for IA

The rise of AI makes a well-appointed IA even more urgent. AI models, such as generative AI, rely on large amounts of high-quality data to learn and deliver effective results.

• Data quality - AI models are prone to noise, inaccuracies and inconsistencies in the data. A good IA provides data governance and quality control.

• Contextual understanding - AI needs to understand the context of the information to generate relevant insights. A defined IA with metadata and taxonomies helps AI interpret the information correctly.

• Accessibility - AI must have access to the right data and content. An IA ensures that information is structured and can be easily found.

• Ethical considerations - An IA can help identify and mitigate biases in the data, which is crucial for ethical AI applications.

If you want to deploy artificial intelligence (AI) based on your own data and content - which is often the most valuable application - a lack of information architecture (IA) is a guaranteed recipe for failure. You will have to clean up, organise and enrich data, which is a costly and time-consuming operation.

Garbage in, garbage out ...

The phrase “Garbage In, Garbage Out” originated in the early days of computer science. In a 1957 article about U.S. Army mathematicians working with old-fashioned IBM mainframes, the phrase “Garbage In, Garbage Out” was used to describe how poor quality input data leads to poor quality output.

This saying has only become more relevant in our increasingly data-driven world. Our decisions based on data are only as good as the reliability of that data.

No trust in data

Salesforce published a report last month on a dilemma related to data-driven decision-making: trust in data is declining, while reliance on data is increasing.

Although 76% of business leaders say the rise of AI increases their need to be data-driven, only 36% say they are confident in the accuracy of their company's data. And that confidence in data has dropped 27% since 2023 (!).

With increasingly powerful AI tools, including AI agents acting autonomously, the quality of the underlying data is more important than ever.

Getting started with Information Architecture

Setting up an IA is a complex task, but the investment is worth it. Here are a few steps to get started:

• Start small - Start with a specific business domain or process.

• Engage stakeholders - Work with business owners, content owners, data stewards and IT architects.

• Define data & content models - Identify key data and content types and define their relationships.

• Implement metadata and taxonomies - Label and categorise information so it can be easily found.

• Establish data & content governance - Define policies and processes to ensure information quality and consistency.

• Make IA a part of your EA - Integrate IA into the overall Enterprise Architecture.

The time for neglect is over. Information architecture is no longer a nice-to-have, but a crucial pillar of a successful Enterprise Architecture - and especially in the age of AI. Start setting up your information architecture today and lay the foundation for a data-driven and intelligent organisation.