Skip to main content

OpenAI integration for image analysis

The configuration of the AI solution and all workflows were integrated directly into digital asset management and are independent of the actual AI service. This means that users do not have to operate an additional tool and the company can change the model if necessary without having to re-implement automatic tagging.

By connecting OpenAI, we have not only changed the model, but also expanded the scope of automation and optimized the process. The new image process is now even faster, delivers more accurate results and also saves money. This creates the basis for further applications, e.g. mass analysis of images or automatic text generation based on image metadata.

1. Automatic basic indexing with AI

After uploading to the DAM system and the subsequent duplicate check, new images can be sent to OpenAI via an interface. There, the images are analyzed and returned with the appropriate keywords. The keyword quality is much better than in the previous model and OpenAI also recognizes relevant company-specific terms, for example yellow buses are tagged as “Postbus”.

2. Checking the image world conformity

OpenAI checks each image for conformity with the company’s imagery. Based on a specific definition, the model recognizes whether an image is a brand image or a pure product image and sets the corresponding feature on the image. Only brand-compliant images are then automatically exported by censhare to the Frontify Experience Hub. OpenAI also provides a detailed explanation, which makes the final check much easier for employees.

3. Creation and translation of the image description

The image descriptions, which were previously created manually, are now also generated automatically by OpenAI. Using predefined prompts, OpenAI creates the descriptions in the exact style and text length required. After final approval, censhare then automates the translation of the texts using the company’s connected Systran system and saves all texts directly to the respective image.

4. Flexible and convenient final inspection

Employees can check and edit the results before they are saved in the system. If necessary, they can delete or add keywords, adjust the description texts and correct the image conformity. After saving, all information is processed and added directly to the image in censhare. The system creates new keywords and automatically recognizes and references existing keywords and synonyms.

80

faster keywording

17000

images managed

95

cost reduction

A well thought-out DAM concept as the basis for current and future requirements

The company manages more than 300 campaigns a year in a censhare Content Hub and manages all images for all applications. Not only do the images need to be quick and easy to find, they also need to be maintained with metadata and delivered to third-party systems via various interfaces. To reduce the workload for those responsible, we implemented automated basic keywording with AI, which significantly streamlined the process. However, the results were not convincing in terms of content.

With OpenAI, we worked with the company to find a model that better meets the content requirements for keyword quality and image description. We not only updated the model, but also optimized the processes based on previous experience and tailored them to the possibilities of OpenAI. The AI now also creates image descriptions for each image, which are also translated automatically. And each image is checked for conformity with the company’s visual world.

Of course, all the advantages of the previous AI integration have also been retained: Searching is quick and easy, the system performs much better than before and is intuitive for users to use without prior knowledge. Automatic basic keywording with artificial intelligence saves the company time and money, as all basic keywords are automatically recognized and assigned.

”Artificial intelligence takes the image management at Swiss Post to a new level - OpenAI provides more context, better searchability and maximum efficiency.

Michael EberleProduct Owner, Swiss Post

Requirements

Swiss Post has long relied on a central content hub with censhare, also fort he management of the extensive image library. We redesigned the keywording system and then integrated Amazon Rekognition into the content hub in order to keyword the images quickly and automatically. However, the AI service was not only too expensive for continuous operation, but also did not deliver satisfactory results.

After further tests, we switched the model to OpenAI and used the experience gained from the previous implementation to optimize the process in the system and add new automations, for example for creating the image description and checking image conformity.

Implemented with
censhare

Use Case
DAM
Content Hub

Implemented for
Die Post

Read Use Case
Download PDF

System Solution

censhare

censhare has fully integrated digital asset management, product information management and content management to help you to master your content. Companies like Allianz, Lands’ End, Dyson, Christie’s and hundreds more rely on censhare for on-brand, always up-to-date content, taking advantage of every opportunity to reach the right customers at the right time.

Learn more about censhare

censhare.com

About Swiss Post

The Swiss Post is a public company that is 100% owned by the Swiss Confederation. In 2021, the company shipped around 1,8 billion letters and 202 million packages, transported 135 million passengers with PostBus and managed over CHF 111 billion in customer assets with PostFinance.

post.ch

Ready to go the path with us? We look forward to exchanging ideas with you.

Contact us