From Text to Image: Crafting Custom Training Visuals with Dall-E

Scroll to see how to create images like these for customer scenarios, quizzes for new hires, and more. 

During the “AI and Learning Management: A Roadmap to AI Strategy Success” breakout at the recent CHART conference in Charlotte, North Carolina, facilitators Rachael Nemeth of Opus Training and Brea Watkins of Levity Live explored how the integration of artificial intelligence is transforming the way training and development are approached in the hospitality industry. Although we had a wide-ranging discussion on the subject, the section that I found most interesting was how Dall-E on ChatGPT has emerged as a pivotal asset by providing the ability to generate diverse, high-quality images from text prompts which can then be used for hotel or restaurant training materials.

The advantages to using a tool like Dall-E are numerous:

• Cost-effective – Using Dall-E will be more affordable than hiring a professional photographer or purchasing stock photos (a subscription to GPT-4, which includes Dall-E, starts at only $20 per month).
• Time-saving – Generating images using AI is typically faster than organizing a photo shoot (and you don’t have to interrupt your teams in operations by taking up time/space for these shoots) or searching for suitable stock photos.
• Customization – AI image generators allow you to create highly customized images that match your specific requirements.
• Consistency – By using an application like Dall-E, you can ensure that all images in your training materials have a consistent style and quality, creating a cohesive and professional look.
• Privacy – Using AI-generated images eliminates the need to photograph real employees or guests, which protects their privacy and avoids potential legal issues. You won’t have to uncomfortably ask your employees to sign a waiver giving up the rights to their image in perpetuity, and you won’t have to keep using anybody’s face in a job aid even after they have long left the company.

For trainers new to Dall-E, the process is straightforward. You begin by considering the type of image needed for your training session, and the more detailed your description, the better. If you require an image depicting a welcoming hotel lobby, for example, your prompt might include specifics about the decor, lighting, and even the presence of guests and staff. After inputting your requirements, Dall-E generates an image that you can then refine by providing further instructions, much like iterative feedback during a photo shoot (turn this way, move a little to the left, etc.).

Using Dall-E effectively requires practice and precision in crafting prompts that will get you exactly what you need. AI tools, while advanced, are not mind readers and operate best with detailed guidance, including:

• Defining the concept – Clearly state what the image should represent (define the context of the training).
• Including specifics – Detail the setting, characters, and actions within the scene.
• Offering guidance on tone – Indicate if the image should be formal, casual, instructional, or engaging.
• Being iterative – Use the feedback loop to refine the images and specify adjustments as needed.
• Remembering legality and sensitivity – Be mindful of legal and ethical guidelines, particularly when depicting sensitive scenarios.

Dall-E on ChatGPT is revolutionizing hospitality training by providing an accessible, cost-effective means of generating tailor-made visual content. Its use not only conserves financial resources but also injects a level of customization and relevance into training modules that were previously unattainable for companies with limited budgets. As AI continues to advance, embracing tools like Dall-E will empower hotel and restaurant professionals to elevate their materials and, consequently, their training effectiveness, ensuring they stay at the forefront of innovation within the hospitality industry.

Below are examples of pictures created by Dall-E that you could use in a class on dealing with upset guests. You can change your prompts to ask for different elements of the scene until you get exactly what you need for your session. In my first attempt, it’s not clear who the guest is and who the employee is (and the backward computer screen doesn’t help alleviate this confusion). Attempt #2 is a little bit better, but I decided to ask again to get a horizontal image that will fill up a 16:9 PowerPoint slide.

You could use Dall-E creations like these two images to quiz new hires on proper uniform standards. For example, you might ask, “What items should these employees be wearing or holding to complete their uniforms and be ready to work at their stations?”

This image might be used in a training session with housekeeping staff. “If you walked into one of our rooms and experienced a scene like this one, what are the first three actions you should take before you start cleaning?”

You can also create job aids like this one showing the proper building blocks of a burger. The first version was created by Dall-E, and for the second version, I added labels in a photo editor to make it even more helpful for new team members.