On June 8 and 9, Adobe held its annual digital experience event in London, with over 3,000 participants. Smile was there, and brings you the key learnings of the event!
Generative artificial intelligence for professional use
Adobe has unveiled its "Adobe Firefly" implementation of image-generating AI, and gives its vision of what generative artificial intelligence should be in a professional context.
Let's take a simple use case: creating a banner to promote content on the web. Starting with a photo, Adobe shows that generative AI is capable of carrying out retouching that is difficult or impossible to do manually. Select a part of the image, ask to add an object, and in a few seconds the object appears perfectly integrated into the image, with its shadows and reflections. Then, to obtain a panoramic image, Firefly is also capable of enlarging the image by adding a completely fictitious landscape to the left and right, which blends perfectly with the original photo in the center. An operation almost impossible to perform manually!
Select an area on the left of the photo, and ask to add a "boat"...
... Adobe Firefly does the rest. The result is convincing! Note the reflection in the water, which is correctly rendered.
The tool's integration with Adobe Express enables rapid production of deliverables. In just a few clicks, banners are automatically generated in formats adapted to each terminal: desktop, tablet, mobile.
Adobe emphasizes its "ethical" approach to generative AI, as Adobe Firefly is trained on a set of images from Adobe Stock whose authors have previously agreed to the use of their creation in an artificial intelligence. This means that professionals can publish the images generated without fear of being suspected of copyright infringement.
For a future version, Adobe is working on making it possible to train Adobe Firefly on a customized set of images. An exciting feature! It would enable a brand to train Adobe Firefly on a set of photos from its entire product range, so that Adobe Firefly could then generate images of the products in situation. No need for a studio visit, images would be generated in a matter of seconds. This would save an enormous amount of time, and offer tenfold creative possibilities!
Data as a lever for business growth
Adobe has presented details of its new "Real-Time Customer Data Platform" (CDP). The Adobe Real-Time CDP is a new solution designed to collect profiling data on users across multiple sources: e-commerce, website, social media, stores... so that it can then be used to segment customers and activate them across different channels: emails, sms, targeted advertising…
Prisa Media was able to testify to its use of Real-Time CDP. Prisa Media is a Spanish media group with 30 brands in the press, sports and entertainment, and music sectors. In total, the group has 230 million devices connected every month and 250,000 paying subscribers.
At Prisa Media, Real-Time CDP is capable of recognizing the identity of Internet users as they move from one site to another of the group's brands, and links all the data collected to a single customer file. In 9 months, Prisa Media was able to deploy Real-Time CDP across five sources and five activation channels. Adobe insists that all data is collected, processed and usable in real time.
We were able to test the platform on site, and the user interface is easy to learn. Connectors are already available to accelerate data feed, for example the connector with Adobe Commerce has been available since November 2022.
Real-Time CDP can be used with other Experience Platform solutions. We saw a demonstration of the use of Real-Time CDP with Adobe Journey Optimizer to trigger actions and engage the customer based on Real-Time CDP information. In this case, Journey Optimizer has access to all Real-Time CDP information in real time, enabling it, for example, to react immediately to an event such as the geolocation of a customer near a point of sale, triggering the immediate sending of a push notification or SMS.
