What you might have learned from the guide to buying property in Spain is that the market is booming. With more and more people looking to relocate to warmer climates and slower-paced cities, houses are in high demand. But this also means that it’s getting harder to get your offers to stand out.

How do you make sure that potential buyers notice your offer? You should pay attention to photos.

It’s no secret that in the era of digital, images take a leading role in seller-buyer relationships. The principle is relevant for all spheres, from retail to real estate. Conversion rates in real estate marketing are highly affected by the photos and their quality. According to an expert in the field of website usability Jakob Nielsen, images are the most effective way to present information on a website. It’s so because people react to them instantly. Artificial Intelligence (AI) tools can help you to manage images.

AI for Managing Images

Let’s begin with why images are important. Statistics prove that the number of photos and their quality affects the dynamics of sales: homes with one photo spend 70 days on the market on average, while homes with 20 – only 32. High-quality photos help to sell homes 32% faster.

Since images are of high importance, employees have to spend a lot of time to choose the right ones. This takes up to 40% of their working time. Therefore, automation is necessary. Before we move to the AI role, let’s look at daily images workflow:

1. Discarding inappropriate photos

2. Choosing a cover photo

3. Setting up an order of displaying photos

4. Creating tags and captures for photos.

If an employee ignores all these, he risks displaying inappropriate images as a cover photo. This affects product attractiveness. Imagine a lot with a beautiful sea view from a terrace and another one – with an old horse in an unkempt garden. The latter would be opened more rarely.

AI Image Recognition

AI image recognition services take the responsibility to detect inappropriate photos and exclude them. The AI market offers ready-made solutions that are optimized for real estate market needs. They are:

1. https://restb.ai

2. https://inten.to

3. https://www.sentisight.ai

4. https://imagga.com

5. https://vize.ai

We used one of the services for a training data set that later one became a trained model, implemented in daily operations. We began with automated tagging of an image and recognition. Tags were used to:

• List priority tags that help to choose photos as the cover

• Make up the list of photos with inappropriate tags that can’t be the cover or displayed in a slideshow at all.

• Apply certain features based on tags automatically

• Prepare a template basis for a detailed description

• Set up a photo slideshow order according to tags

• Discard unacceptable images.

Once we managed to collect the necessary number of images, we prepared the necessary data for automated tagging and launched the process. As the results show, 2-4 tags per photo are of the highest accuracy. After that, all the tags were classified, algorithms programmed to determine rendered photos were launched.

When the trained model was deployed, we primarily used it for images auto-detection, automated creation of the displaying order, naming files, and photos description generation. One of the main goals was to automate the process of eliminating inadmissible photos for the cover. Strict rules show the best results. The system doesn’t use an image as the cover if it’s considered to be unacceptable with a minimum of 60% probability.

In the end, the share of bad photos used for cover reached 4% after two months. For comparison, the parameter was 34% before the AI-based algorithm.

Auto-Tagging for SEO

Tags are useful for SEO-optimization. We use a small trick to obtain high traffic: we create landing pages for various types of queries. The common problem is that some aren’t labeled with all the tags they could have. This leads to the fact that the final output is not full. High conversion rates depend on the more in-depth queries. As for the landing pages, search engines may not show the page in case just some items are placed on it.

Automation allows us to eliminate human forgetfulness. To begin with, our company turned to the Virto Commerce catalog module that allows integrating digital data. Then, we did the following:

1. Added features and tags from the list of generated tags that are useful to search engines and clients.

2. Combined certain tags together and used them as a group of synonyms.

As a result, we managed to double the number of items in the output. It means that we used a minimum half of important, useful data that wasn’t used before. Another significant achievement is the facilitation of writing product descriptions. Once the photo is uploaded, a copywriter can see all the tags, which can be utilized. It speeds up the process of writing texts since the copywriter spends about half of the time identifying the key features of a product.

All in all, AI opens new horizons and allows more time for more meaningful tasks.