Years ago, I was sitting in on a Monday morning meeting at a financial services company. Our customer-facing website had real usability problems, and the conversation was getting heated. Then one of our most senior digital leaders, Ethan, waved his hand and said, oh, that's just a governance issue, and the conversation ended. We left that room with completely different ideas about what to do next and nothing changed. That moment has stayed with me for many years because Ethan wasn't wrong. It was a governance issue. But using the word governance as if we all understood what it meant gave everyone in the room simultaneous permission to stop thinking about it.

Let's start at the beginning, because if we do not agree on what governance actually means, everything that follows falls apart. When most people hear the word governance, they think bureaucracy, committees, policies, the thing that slows everything down. That is not what governance is. Governance is the framework of decision-making rights, processes, and people responsible for your platform. Three questions sit at the heart of it: who has the right to decide this (the person who is accountable), what process must be followed (a minimum step, a review, an approval, a sign-off before content goes live), and how will we know if it's working (define what good looks like, and set a review date).

This is the governance framework I keep coming back to. When it works, nobody notices. The platform just works and employees find what they need. When it breaks, everyone feels it. Employees stop trusting the intranet, no one knows who is responsible, duplicate pages multiply, contradictory information spreads, and if you add AI on top of that, AI does not fix bad content, it amplifies it at scale. That is why governance is not bureaucracy. It is the foundation that makes everything else trustworthy.

Here's what I've learned watching governance models hold or fail over many years: the model on paper is almost never the model in practice. There is always a person, often the person watching this, who absorbs the gap, who manages in every direction, who compensates with extra effort so the rest of the organization can function smoothly. That work is the most important leadership skill in digital workplace management, and it is also a lot. Governance covers a lot of ground, so in this video we focus on two specific problems that make digital workplaces break down, and that Staffbase Content Pro was built to solve. These two problems feed each other, and they are both getting worse right now as AI raises the stakes for content quality.

The first problem is content quality. In most organizations, intranet content accumulates faster than it gets maintained: eight duplicate pages, nineteen contradictory pages, and twelve not updated in a year. Here's a real example: one page says thirty vacation days, another says twenty-one, both are live, and employees don't know which one to trust. At scale, this is nearly impossible to manage manually, and if you plan to use AI on top of your intranet, it's an even bigger problem, because AI will only scale whatever is already in there, accurate or not. AI content governance gives you an automated eye on the platform, surfacing outdated, duplicate, or off-brand content before it damages employee trust.

The second problem is authoring. The people who hold the most valuable knowledge in your organization, HR experts, department heads, and subject matter experts, avoid publishing because the process is too complex. They know the answer, they just can't build the page. So editorial teams become the bottleneck, or nothing gets updated, or it gets published inconsistently, which takes us straight back to problem one. AI page generation moves your subject matter experts from author to editor: they bring the knowledge, Content Pro handles the structure.

The good news is that you do not have to carry all of this by hand anymore. Content Pro does not replace governance, you still lead it, decide it, own it, but AI-powered content governance takes the manual burden off the parts that were never going to scale by hand. If you are doing this work well, you are leading organizational transformation, whether anyone recognizes it or not. Governance is leadership work, not administrative work, and Content Pro automates the parts that we're never going to scale by hand.

AI Content Governance is an improvement to the feature that was released in the last release, Q1 2026, and you can now use AI to further monitor pages in the Content Governance tab in Pages. The use case here is similar to before: you maintain quality, correctness, and freshness of pages at scale, but now you can also use AI. Editors and up can view details for their own pages.

You now have two AI issue types: Duplicate Pages and Contradictory Pages. Duplicate Pages contain semantically very similar content, and Contradictory Pages are pages that are logically confusing. The AI searches through pages in the same space and finds potential issues, then gives a short explanation about why the pages have been flagged. There is an audience check, meaning content varies based on audiences and, obviously, between spaces. There might be multiple HR pages within multiple different spaces with different information regarding vacations or policies, so searching for content in the same space only avoids incorrectly flagging legitimate variations.

In the Content Governance tab inside Pages, you can see Duplicate Pages and Contradictory Pages listed as the two new issue types. Clicking on one shows an AI-generated insight, for example: both pages describe user story mapping in agile with nearly identical content and examples. From there, you can go into the page and edit it to avoid duplicate content, combine the pages, or separate them out. For Contradictory Pages, an example insight might say both pages state different numbers of free vacation days, thirty versus twenty-one, showing the two pages where that applies so you can edit directly. Duplicate Pages and Contradictory Pages are also listed in the top metrics, with tooltip explanations for easy oversight.

The AI service used is Microsoft Azure OpenAI. There have been multiple improvements since the Q1 release: the top metrics now include not updated in one year, broken links, broken widgets, no owner, no views, and expired reminder. You can click an issue to open a side panel with further details, for example for broken widgets or pages with no owner; some pages may show multiple issues at once, such as both a broken widget and a contradictory page. There is also a new page owner column in the overview table in addition to the editor column, letting you see whether an owner has been assigned and who the page editors are.

New actions in the action menu let you show details, assign an owner, or dismiss an issue. Assigning an owner takes you to the page settings; dismissing an issue removes it from the overview, useful if, for example, a page genuinely doesn't need an owner. You can also filter the overview by issue type at the top, for example showing only pages with no owner, or combining filters such as duplicate pages and no owner, and the top metrics realign automatically as you filter.

AI Page Generation allows you to generate well-structured pages inside the Content Designer for Pages with AI, saving time and effort with fast content creation. Managing editors can create a new page, then prompt the AI in natural language. They can optionally add a file, PDF, or doc, upload up to five images to be used inside the page, and reference an existing page. Once they click send, the page generates: the file helps generate the text, the images are added into the page, and the existing page is used for structure, so a well laid-out page can be attached as a structural example. The page generates based on the specifications provided and is then fully editable.

In Studio, after creating a page and choosing a space, clicking Create takes you to the AI generation screen. You can start from scratch, which opens the content designer to add blocks and text as normal, or enter a prompt along with your information, which is the more common path. You can attach an existing page as a structural example, upload a PDF as your text base, and upload a few images to include on the page. Only one PDF can be attached at a time; adding a second produces an error saying a maximum of one document is allowed, though multiple images are still fine. Clicking Send generates the page.

The generated page inserts the uploaded images along with text drawn from both the PDF and the prompt, organized into the requested sections, for example an information section and a bullet-point list of best practices. From there, you can edit the page as needed, save it as a draft, and publish it when ready to make it live.

A few good things to know: the AI service used is Microsoft OpenAI, and the feature is only available in the Content Designer, not the Classic Editor. Content can be generated in most languages, based on the language of the prompt or of your Staffbase platform. There is a file size limit of 25 megabytes, and it is not possible to regenerate a page directly, though generated content can be adjusted afterward, or you can delete it and create a new one with a different prompt or different uploaded content.

Best practices for prompting: use short, natural language, describe your goal and what you want specifically, and feel free to reference blocks, formatting, or design, since the AI will understand that. Provide as much context as you can, including about company language, so the AI gets to know you better. The Organization Voice feature can also provide additional background for the AI in these areas.