Artificial intelligence applied to archiving

How to bring value to DAM thanks to artificial intelligence

Every employee in the field of communication is daily struggling with the creation of content that captures the attention of the public. He knows that he has in his archive photos, texts, videos that could help him to create new contents but, at the thought of looking in folders and under folders looking for a particular image or a specific text, prefers to experience something new every day. With an inevitable increase in costs.

Tagging or adding information to existing business resources is still a cost in terms of time that not all companies are willing to face. Modern technology, with the newly developed artificial intelligence algorithms, also meets needs such as to recognize whether an image represents a person, a landscape or both and consequently enrich digital resources with new information.

Imagine that you want to create content that talks about the fashion color of the moment, and you have to search your server for all the images that have that color inside them.

How many of these have the color you are looking for in the file name? How many of these have been inserted in folders whose name shows the color you are looking for? If the answers to these two questions are "I don’t know" maybe it’s time to start thinking about incorporating a Digital Asset Management into your tools and populate it by exploiting all the potential of artificial intelligence.

Digital Asset Management (DAM) and Artificial Intelligence (AI): a few questions.

- How can we be sure that the tags and metadata associated by the AI to the resources in the DAM comply with the company guidelines?

In a DAM a certain taxonomy can be set strictly related to the business environment. This, however, could lead to the loss of other valuable information when AI asks for new tags not present in the preset taxonomy. A further alternative would be to instruct the AI so that it automatically increases the number of tags remaining in the guidelines initially set.

- Can you get tags in different languages than the DAM system?

Yes, it is possible in two ways: by directly using the vocabularies proposed by the provider of the AI algorithm applied to the DAM or by collecting all the tags created and translating them with the technology proposed by Google. This second option is usually considered in those cases where it serves to have tags in a minor language, usually not provided by major AI providers.

- I tag associati alle risorse possono essere esportati in altri sistemi come ad esempio un CMS?

Sì. Grazie alle API messe a disposizione dal DAM, è possibile esportare informazioni come tag e metadati per gestire canali di uscita con molta facilità. Un tag relativo ad una categoria di prodotti associato ad una determinata risorsa può fare in modo, secondo le impostazioni del CMS, che questa venga pubblicata nella categoria corretta senza un ulteriore intervento manuale.

- È possibile istruire un algoritmo AI utilizzando contenuti specifici dell'azienda?

Sì. Dato un set campione di immagini con le rispettive descrizioni è possibile addestrare il motore di AI affinché applichi tag simili per immagini che riportano contenuti simili. Questo consente inoltre, in fase di ricerca di ottenere nei risultati risorse che potrebbero trovarsi in cartelle diverse all'interno dell'archivio digitale.

Share