What Are You Waiting For? The Tale of Doing Nothing

What Are You Waiting For? The Tale of Doing Nothing

Is this you??

You have a lot of content. A large corpus. It might be structured using DITA or some other XML schema. It might be in Word or Google Docs. Maybe you have some of each.??

You know your content is redundant. Writers have copied and pasted plenty of times over the years, making tweaks along the way. You are experiencing content drift – many versions of the same content that are all just a bit different from each other. You have no single source of truth.??

Over time, you’ve had to make changes to the content: updates, new releases, sunsetted features, and so on. Since you’ve lost track of all of the versions and variations of the same content, some of the versions are updated, others are not. You really can’t be sure.??

In terms of metadata, you have some. It is incomplete. As your product line has expanded, you haven’t had time to create more metadata and keep it organized. You intend to go back and fix it, but another urgent project comes up and your attention has to be put elsewhere.??

The metadata that you have is unevenly applied. Some of your files are tagged thoroughly. Others, minimally. Again, you really aren’t sure.??

If one or more of these statements apply to your content, first know that you are not alone. As our companies have aged and expanded, we have created more and more content for more and more products and services. Without strict governance and impeccable content stewardship, it is very easy and incredibly common for things to start teetering out of control.??

In the time before artificial intelligence (AI), most content people understood what they needed to do to solve their runaway content problems. While not a menial or inexpensive task, managers understood how to make the case to senior leadership for the money and time to make things better. It could take years, but it was an inevitability for most medium and large enterprises.??

Then, from seemingly out of nowhere, along came AI. And overnight, the content world became paralyzed. AI, Large Language Models (LLMs), and now Retrieval Augmented Generation (RAG) tools quite literally stopped everyone in their tracks.??

If AI is here, what should we do? Should we:?

  1. Skip structured content and move straight to AI solutions???
  2. Continue structuring our content and then implement AI??
  3. Do nothing because who knows what is going to happen next??

In my experience at Content Rules, Inc. , most companies select choice three, do nothing.??

The mantra is: “Since we don’t know what comes next, and what we do now might be totally wasted, let’s just sit tight and watch how things develop.”?

Interlude: Moose and Doing Nothing?

I have a dog named Moose. Moose is a very large, slobbery rescue who weighs in at 105 pounds. Once upon a time, we had to take Moose to the vet for some stitches. When he was done, the doctor put a cone around his neck to keep him from reaching the stitched area.?

Big black dog named Moose with cone around his neck.
Moose

Moose decided that he could no longer walk because he had the cone around his neck. He was simply stuck in place. He would not move. It was like his legs stopped working.??

He didn’t know what was coming next, so he did nothing. Just like content teams.?

Doing Nothing Doesn’t Help?

We all know what happens if you have runaway content and you decide to do nothing about it. The problem becomes exacerbated. Your content doesn’t magically fix itself while you wait to see what AI is going to do next. Quite the opposite. The hole gets deeper.??

Things never improve when you do nothing about a problem.??

What to Do Instead of Nothing?

While you are taking a wait-and-see approach, it makes sense to maximize the results that you will get with your future AI solution.??

Many factors negatively affect the quality of results generated from AI, including:?

  • Redundancy throughout the content?
  • Conflicting or inaccurate information??
  • Poor quality or absence of metadata?

When these factors are present, the AI solution cannot be trained, validated, or used effectively. You cannot trust the results from an AI solution that has been trained on your messy enterprise content as it is today.??

The good news is there are things you can do to your content to prepare it for more effective AI implementations. These things include:?

  • Chunk your content into small units of information?
  • Eliminate inconsistencies??
  • Remove duplicate versions of the same content?
  • Fix inconsistent terminology, grammar, and style?
  • Tag content with semantically rich metadata?

Rather than continue to suffer from runaway content, get to work on fixing it now, before you have your AI engine. That way, your corpus is ready when your solution arrives.??

And just think of all the problems you’ll solve at the same time.??

Let's work together to get your content AI-ready. You can find me at Content Rules, or drop me a message here on LinkedIn.

要查看或添加评论,请登录

Val Swisher的更多文章

社区洞察

其他会员也浏览了