Why you can’t innovation without content strategy and content structure

Frequently, clients say they want to be innovative on their website or intranet (or another information product) and have a culture of innovation in their organization. What does it mean to innovate with content? What needs to be in place to be innovative?

Having seen inside the world of information management in many different organizations in the last 20+ years, you’ll be surprised at what is actually innovative. You’ll be surprised at what companies are not doing to support their content ecosystem.

Organizations absolutely must assess the risks of AI before starting projects or unleashing an AI-driven product into the world. You must grapple with privacy, hallucinations (inaccurate, nonsensical, or factually incorrect information), security, and harm to people and the environment. You must insure that your AI product best serves your users and employees.

Let’s dive deeper.

How companies are not, and cannot be, innovative with content

Companies who wish to innovate on top of their content are struggling with fundamental abilities. It’s not surprising. Executives have relied on the “next best thing” to solve their content problems, and typically don’t put much money into content development. With the advent of generative AI, this problem is becoming more and more pronounced.

Companies don’t have common metadata or crosswalks

Metadata tells us what the content is and what it is about. We use fields like title, description, author, date, content type, topic, and audience to ensure that applications can better understand the content and display it accordingly in an interface. Without consistent metadata across systems and departments, content becomes siloed and search results are sub-par.

Crosswalks map fields between systems. When you haven’t established crosswalks, different systems can’t talk to each other. Combining data into a data lake or some sort of repository creates confusion. One system might have the field of “Audience” and another system might have the field of “Content Target.” They might be different field types, or use different taxonomy terms. Without mapping these fields together in a crosswalk, there’s no way to know that they are the same thing.

Without metadata standards, generative AI doesn’t know how to connect metadata across systems. More fundamentally, any personalization, analytics, or reporting will also be unwieldy and inaccurate.

Companies don’t have a shared taxonomy

A taxonomy aligns how an organization names, categorizes, and structures concepts. I still see content management systems, document management systems, and other management systems with poorly created taxonomy facets that are used only within that system. I also see different teams using the same systems and using different taxonomy facets for the same thing. If we take the Audience example again, one team could use the Audience field and facet while another team creates their own taxonomy, puts the Audience terms into that facet, and uses those Audience terms.

• Companies need taxonomy to label content, to tell the application what the content is about, who it’s for, which department owns it, and for many more reasons

• Companies need taxonomy to relate content together and disambiguate between different content

• Companies need taxonomy to support search and search results filtering

• Companies need taxonomy to improve analytics, personalization, and reporting

This is the start of innovation. If you don’t have this, you can’t build on top of it. It’s a house of cards. You’ll miss opportunities for personalization, content reuse, search, and browse. Any generative AI tool won’t be able to make connections between content, or it might try in a speculative way. But if you’re dealing with critical information, do you want someone else to speculate?

Companies don’t have domain models

Domain models reflect how a company represents its core ideas, relationships, and concepts in structured form. Without a domain model, your organization has no shared understanding of how things are related together. What problems does it create?

• As an organization, you haven’t agreed on it internally and there are no set definitions

• Because you haven’t modeled your domain, you can’t properly explain it to users or sufficiently support them in their interactions with you

• Any generative AI tool won’t be able to make connections between content, or it might try in a speculative way

Companies don’t have business process models

Content doesn’t exist in a vacuum—it supports operations and your organization. Your business not only needs to map out its content workflows and governance, but it needs to map out the business processes so that you can properly support users and internal staff. If you haven’t mapped your business processes, you’ll have these problems:

• Internally you won’t have clarity on what you’re actually doing (as opposed to what you’re supposed to do)

• You won’t be able to identify where you can improve these processes

• You can’t explain the process to users or sufficiently support them

• You can’t identify content to help users through the process

• Any generative AI tool won’t be able to make connections between content, or it might try in a speculative way

Companies do not have content governance

Without content governance, information systems die slow deaths. Governance provides the rules of the road: who owns content, how it’s maintained, what good looks like. Not having governance means that your content can be:

• Difficult to navigate

• Out of date

• Inaccurate and incorrect

• Impossible to search

• Off-brand

• A resource drain

• A liability

You simply cannot put content with these kinds of problems into a generative AI tool and expect it to fix the problems. It’s still “garbage in, garbage out.” You must work to create a solid body of content that is governed, so that you can innovate with that content.

Companies do not know enough about their users and their content needs

Without user research, mapping content to the user journey and tasks, analytics reporting, and structured feedback loops, companies guess at what structure and what content is needed. I’ve seen organizations produce a lot of content that is then never read, never accessed, and is a resource drain when it needs to be reviewed and kept up to date. Content is an psychological albatross that employees carry every day.

Companies don’t reduce or eliminate bias

Unstructured, historical, or legacy content is riddled with unconscious bias. Content needs to be updated to remove gender bias, racial bias, and other historical inequities. Without reviewing and revising content, any innovation sitting on top of biased content will also be biased and perpetuate harm.

Companies simply don’t have the maturity they need

Many organizations have never treated content as a strategic asset. They've underinvested in content systems, governance, and skill development. Now, they want to leapfrog to AI innovation without building the foundation.