DAM and AI

Artificial Intelligence is accelerating the rate of increase in asset metadata

Evolved Media are working with and integrating to Elvis DAM a number of AI (artificial intelligence) tools which can offer in-DAM updates to asset metadata on a variety of types. Keyword tagging is just the beginning. Tools such as Amazon's Rekognition and Google Vision are very successful in identifying objects in images with position information. Other services are further enriching by identifying objects styles, material patterns, colours etc.

Evolved Media are already working to integrate these tools into clients workflows. They are opening up interesting possibilities for workflow automation, such automatic cropping based on object detection. Image correction and enhancement. In the past these processes have been slow and labour intensive and where large volumes of assets are involved, this can increase overall cost significantly.

Initially when we talk with clients, conversations around AI start with auto tagging  type features. How to recognise objects. If it's fashion, then blouse, dress, trousers, handbag etc. This then quickly moves on to styles and patterns. There are now a number of off-the-shelf services providing great information back on imagery. Clients are often focused on enhancing their e-commerce sites. Usually to provide visual search and features like find me this style etc. 

But more and more, we see the use of this AI technology goes deeper into the workflow, and when hooked into DAM broadens out the use of both in the workflow. From buyers and merchandisers tools where immediate input from AI can help build styles and collections, to identifying imagery as its being selected, and mapping it straight into the client's product data and identifying all the new shots and views as they arrive so everyone can see progress as new ranges are commissioned and brought to market. AI has an important place in all the steps in the workflow. From beginning to end.

We can already see as well, that this enhanced layer of imagery can create feedback loops between the creative teams and commercial teams, where the business will better understand types of images that might work better on an e-commerce site to increase sales of particular products. To feed back into briefs for shoots on what types of shots and views are required.

Retail was traditionally a bricks and mortar venture. With the internet opening as a retail platform in 1994, Amazon and eBay quickly jumped on board soon after in 1995. Twenty-four years on, we have 12-24 million eCommerce sites trading, making for a noisy and competitive environment.

Clients such as Next, roots can be traced back to the late 1800s, the modern branded store and directory catalogue we recognise began in the 1980s. The brand opened its e-commerce site in 1999 and has become the largest retail brand in the UK. Its success can be owed to many factors, but the brand and directories stay true to the core customer base by offering lifestyle-based fashion and homeware choices, built carefully around consumer preferences and modern design.

But the delineations around e-Commerce are blurring, we are entering the age of “Experience-Driven, UnifiedCommerce” where everyone is permanently connected to the internet and sales are experience-led and personally tailored.


Evolved Media | Case Studies | ID:1863 | v2.1 | Last modified 2019-02-25T09:38:00