Article
Automatic tags and asset categorisation: the work inside a DAM
AI-generated tags on DAM assets: why I use tools already on the market and what really matters for them to help the people searching every day.
Published 15 November 20264 min readAIDAMTagging
If you work with lots of photos, videos and documents, you know how it goes. A file uploaded in a hurry ends up in a folder with a random name. A few months later nobody remembers what is inside. For BaseCamp Studio I build features in the DAM, the software where the company organises its digital content. One of them is automatic tag generation on assets. I will not go into technical details here: I will tell you how I think about it and what I look at.
Tools already on the market, not a model from scratch
I generate the tags with AI tools already on the market. I did not build my own model, and for a company it would not have made sense. A model from scratch is a research project. Integrating an existing tool is work that ships. The real choice is not which AI is the most impressive. It is how to fit it into the work of the people who use the DAM every day.
What tags are for
A tag is a short label. It lets you see at a glance what is in a file, without opening it. It also helps classic searches, the keyword kind. If the right tag is there, you find the content with the filters you already use. If it is missing, that file barely exists as far as search is concerned.
- They help you find your way when browsing a large archive.
- They make the filters and keyword searches you already know useful.
- They do not replace searching by meaning: they sit alongside it.
Tags, categorisation and semantic search
Tags are only one piece. In the same DAM I developed AI enrichment of assets, which extracts and categorises information from images, text, PDFs and videos. There, every suggestion goes through a person before it becomes asset data: I cover it in AI asset enrichment. Then there is searching by meaning, which I describe in my case on semantic search in the DAM. They are three connected things, but they answer different needs.
Before switching them on in your company
I will not promise time savings or accuracy I have not measured. Before talking about tools, I would ask how you search for your content today and where you get stuck. Then who needs to find it, and with which words. If the problem is that nobody can find the files, automatic tags can be a simple first step. If you need to search by meaning instead, it is better to start there.
If you want to understand where AI can fit into your archive, see my AI consulting service. Or get in touch with an example of a file you cannot find today: I would start from that.