Table of contents
There is a statistic circulating right now that says document management specialists face a 68% probability of automation. It is being shared on LinkedIn, discussed in team meetings, and quietly making people anxious. I want to talk about it, not to reassure you that everything is fine, but to tell you that the conversation is pointing in the wrong direction entirely.
I have been a document controller. Not in title only. In the actual work of it. Transmittals, metadata schemas, revision cycles, contractor submissions that never arrived in the right format, naming conventions that three different engineers all interpreted differently. I spent years in that role before moving into document management, then senior document management, and eventually into information management and BIM advisory for major energy infrastructure programmes. I understand what the job actually contains, not just what the job description says.
So when I read that 68% figure, I nodded. Not in panic. In recognition. The automation probability is real, but it is only real for one version of the role.
The Part of the Job That AI Will Replace
There is a version of document control that is entirely at risk. If your daily work is receiving a file, checking it has a document number, uploading it to a system, and sending a transmittal notification back to a contractor, AI can do that now. Not in five years. Right now.
While working as a product owner on a major infrastructure programme, I was part of a team developing an intelligent document checker integrated as an API into our common data environment. The system could read incoming documents, validate metadata against a predefined schema, identify missing fields, detect naming convention violations, and reject non-compliant submissions automatically. It was fast. It did not need a salary. It worked.
If your entire value as a document controller sits in that activity, the automation risk is not 68%. It is higher.
The Part That AI Cannot Replace
Here is what the statistic does not capture.
A contractor submits a document marked for information. The system is configured to expect it for review. The workflow has changed mid-project, agreed verbally in a coordination meeting last Thursday, but nobody updated the configuration. The system declines the submission automatically. The contractor raises a delay event. The project manager wants to know why the DMS is creating problems.
That conversation, the one that explains the deviation, adjusts the expectation, prevents the delay claim, and maintains the relationship with the contractor while quietly updating the system logic behind the scenes, is not automatable. The intelligence required to know when to override the rule, when to follow it, and when to flag it as a systemic problem is built on context that no configuration file contains.
That is not a filing function. That is judgment. That is the part of the role that the senior document manager does every day without anyone noticing, right up until it stops being done.
The Real Threat
The real threat is not AI replacing document controllers. The real threat is what happens when project managers and operations directors believe that the filing function can be automated without replacing the judgment layer. They will buy an intelligent document processing tool, switch off the human oversight, and discover six months later that the documents are all in the system and none of them are right.
The question nobody is asking is whether the engineers, project directors, and asset managers who work alongside document controllers actually understand what those people do. The answer, in my experience, is often no.
Ask yourself when, in your current project, someone sat down and defined the document structure that will support operations for the next 30 years. Was it at FEED? Detailed design? At handover? Or was it treated as an afterthought until the as-built moment arrived, and everyone was scrambling to reconstruct a paper trail that should have been intentional from the beginning?
That is the problem. Not the AI.
Document controllers who thrive in the next five years will not be the ones competing with automation. They will be the ones who can explain why the structure exists, who owns what information and when, and what a deviation from the process costs downstream. That person becomes more valuable when AI arrives, not less. They become the one who tells the AI what to do, and who catches it when it is wrong.
Are You Safe? Use This Table to Find Out
Be honest with yourself.
| What you spend most of your time doing | AI risk level |
|---|---|
| Uploading documents and checking naming conventions | High. This is automatable today. |
| Running transmittal logs and chasing outstanding submissions | High. Workflow automation handles this. |
| Configuring review workflows and managing metadata schemas | Medium. Tools assist, but judgment is still needed. |
| Advising project teams on document structure from FEED onwards | Low. This requires experience and contextual understanding. |
| Managing contractor relationships and resolving submission disputes | Low. Collaboration and verbal communication cannot be configured. |
| Defining the information management strategy for handover to operations | Very low. This is consultancy-level work. No tool replaces it. |
If most of your time sits in the top two rows, the automation risk applies to you. The solution is not to resist the tools. It is to move into the work that the tools cannot reach.
The Final Word
AI will not replace good information management. It will expose the absence of it. The document controllers who understand why a structure exists, who can see three steps ahead of a contractor's submission, and who know that a verbal agreement on Tuesday needs to be reflected in the system by Wednesday morning, those people are not being replaced. They are becoming essential.
The ones at risk are the ones who never understood why they were doing what they were doing. That has always been true. AI has simply made the question harder to ignore.
At SøDera we work with energy and infrastructure companies to build the information management foundation that holds up whether the team is ten people or a thousand. If your document management is reactive rather than intentional, or if you are about to automate a process that has never been designed properly, that is exactly the conversation we should be having.
Stay informed
Get new insights in your inbox
When we publish a new article on RDS, document management, or asset data strategy, we will send it to you directly. No noise, no marketing. Just the article.
Unsubscribe any time. We do not share your email.