Business challenge

JASONGRECH spent months sifting through images to gather inspiration and saw an opportunity to improve the process, combining data insights with intuition to produce stunning collections.


Analyzing unstructured images and comments from social media, the solution can predict fashion trends and understand the audience to guide decisions about the creative direction.


600% faster

research phase, accelerating the process from 28 to 4 days on average

Boosts confidence

in the creative direction of a collection by backing up decisions with data

Speeds storyboard creation

allowing the designer to focus on showstopping details such as custom dyes and accessories

Business challenge story

Slow and subjective creative process

As both artists and entrepreneurs, high-fashion designers like JASONGRECH need to walk the line between avant-garde and marketable. They want to set the trends, but they also need the right colors and cuts to appeal to the masses. However, the creative process can be slow, opaque, subjective and messy. In the past, Grech and his small design team would create a database of images of the previous season’s colors, fabrics, necklines and other details. He also collected images that inspired his creativity on a personal level, usually photos of architecture. After weeks or months of sifting and sorting the images, Grech would settle on a color palette and general direction for the collection. And the team would frequently make last-minute changes in the weeks and days leading up to a big show, second-guessing their work as they watched competitive lines unfold around the world. JASONGRECH aimed to simplify this process and make design choices with greater confidence, freeing up more time for show-stopping details such as matching accessories and custom-dyed fabrics.

IBM Watson gave us a direction and helped us evaluate risk and maintain confidence in the collection.

Jason Grech, Fashion Designer and Founder, JASONGRECH

Transformation story

Cognitive tools spark innovative designs faster

Top designer JASONGRECH is transforming his creative process with a first-of-a-kind Watson AI solution designed to distill hundreds of thousands of fashion images and social media conversations into usable inspirations for upcoming collections. The solution collects images from Instagram, Pinterest, Twitter and fashion archives, using visual recognition technology to identify and categorize key elements: faces, the human form, colors, patterns, fabrics and more. By analyzing the occurrence of these elements over time—and how the images are shared and discussed—the system can predict upcoming trends in color and style. It can also pull in images from outside the fashion world, such as architecture, and find similarities in lines, curves and textures to serve as inspiration for clothing. Plus, the solution analyzes unstructured, natural language social media content to understand which styles generate the most buzz among industry influencers, including the hemlines, necklines, dress lengths and more. Parsing comments and metadata such as shares and likes, the solution can also produce insights about the people behind the posts—whether they’re whimsical or serious, brash or reserved, earnest or sarcastic—which makes it possible to segment and understand the target audience.

Results story

Information-gathering phase accelerated

The first-of-a-kind Watson AI solution accelerates the information-gathering phase of the creative process by 600 percent, from 28 to 4 days. With the insights and predictions from the cognitive tools, JASONGRECH can define the creative direction of a collection with greater confidence, using data to support decisions about color palette and other design elements. And the solution allows the designer to finish storyboards faster and focus on showstopping details such as beadwork, custom dyes and accessories.

JASONGRECH is approaching the creative process in a completely new way with a first-of-a-kind Watson AI solution, gaining a competitive edge with the ability to predict where the fashion industry will go next rather than reacting to the work of other designers. By analyzing images and social media buzz for trends and inspirations, the designer can move with greater confidence and speed, with more time to spend on showstopping elements such as intricate embroidery and beading that strengthen his reputation as a couture fashion icon.

In the world of high fashion, it’s not unusual to spend a full year putting together a collection, with weeks or months of the creative process devoted to finding inspiration and building storyboards. JASONGRECH’s small team worked for months, manually gathering images and doing research to inform the creative direction. Now, the team can get the same insights from a Watson AI solution that can understand and analyze images and social media content in a matter of hours or days, providing recommendations and predictions about fashion trends in upcoming seasons.

The Watson AI solution collected 10 years of images and unstructured text from fashion archives and social media sources, including Instagram, Pinterest and Twitter. The data includes hundreds of thousands of images of runway looks, natural language conversations about fashion, and social media platforms’ metadata, such as numbers of likes and shares.


Based in Melbourne, Australia, JASONGRECH is a couture brand specializing in high-quality bridal and red-carpet gowns. The brand is the creation of designer Jason Grech, who started his career in fashion in 1997. Today, the JASONGRECH workshop and showroom are located in a heritage-listed stable building in North Melbourne, where Grech and his small team create stunning gowns known for their fine details and delicate embroideries.

Solution components

Take the next step

To learn more about IBM Research, please contact your IBM representative or IBM Business Partner, or visit the following website:

View more client stories or learn more about IBM Watson