Signs of the Future: Designing Airport Signage with the Ingenuity of ChatGPT-4.0(including video)
Testing ChatGPT-4.0 image recognition with this image of an airport departure hall (but without telling it!)
Last month, OpenAI launched image recognition on ChatGPT 4.0 and what better way to test it than with an image of an airport departure hall. Firstly, to test the level of recognition and secondly, to make some recommendations and improvements to ‘the situation’ as well as propose four signage designs.
The image I used (see above) is widely available on Unsplash and shows a generic airport departure hall with passengers queueing to check-in. I purposely drafted a vague prompt (not something I would generally recommend!) with no mention of airports or departure halls, and the results were incredible. Here’s how I did it and the results of the test.
Getting Started
With the image of the airport departure hall, I initiated a conversation with ChatGPT-4.0 in default mode. This mode displays an image icon in the prompt field. To test the image recognition abilities I used the following prompt, “Give me five ways in which we could improve the situation in this image”; no mention of airports, queues, departures or signage. Purposely vague.
Within seconds ChatGPT-4.0 suggested the following:
Certainly. Based on the image you provided, which appears to show an airport check-in area, here are five ways to improve the situation:
Automated Check-in Kiosks: Introduce more self-service kiosks where passengers can check in, print boarding passes, and even tag their own luggage. This reduces the reliance on manual counters and speeds up the process.
Queue Management: Implement a better queue management system, perhaps using digital displays to direct passengers to the next available counter. This ensures a smoother flow of people and reduces crowding.
Mobile Check-in: Promote the use of mobile check-in apps that allow passengers to check in and receive a digital boarding pass on their smartphone, reducing the need for them to queue at the counters.
Dedicated Counters: Allocate specific counters for different services, such as for passengers with no checked baggage, families, or passengers with special needs. This allows for faster processing tailored to specific requirements.
Staff Training: Ensure that staff are well-trained to handle common issues swiftly and that there are enough staff members during peak times to manage the volume of passengers.
Implementing these suggestions could enhance the efficiency and user experience in the airport check-in area.
Not bad considering I’d offered no context in my prompt and, although some of the recommendations seem obvious such as ensuring there are enough baggage trolleys, there were others including a QMS and dedicated counters for different passenger segments that were more intelligent.
Testing Recommendations with Itself
The next step was to test the recommendations provided with the model itself so I selected Queue Management System (QMS) as an experiment. I thanked the model (I believe we should always keep it polite!) and asked the following:
Thank you! Could you provide an overview of what a queue management system could look like?
Here are the results:
Certainly! A queue management system (QMS) is a set of tools and processes used to manage and optimise the flow of customers or clients, whether in physical locations (like banks, hospitals, and service centres) or online platforms (like customer support websites). Here's an overview of what a robust queue management system could look like:
Digital Displays: Install digital displays that show which counter or service point is available next, directing customers efficiently. This ensures customers are attended to in the order they arrive.
Ticketing System: Provide customers with tickets that have a number on them. As service points become available, the next number in line is called, either audibly or displayed on screens.
Virtual Queuing: Allow customers to join a queue using their mobile device or a kiosk. They can then be notified via SMS or app notification when it's their turn, allowing them to use their waiting time more productively.
Differentiated Queues: Create separate queues for different services or customer types (e.g., express lanes, priority services for premium customers, special needs, etc.) to ensure efficient and tailored service.
Real-time Monitoring: Implement systems that allow managers to monitor queue lengths and waiting times in real-time. This can help in reallocating resources or staff as needed.
Feedback Mechanism: Install feedback terminals or use mobile apps where customers can rate their waiting experience. This can provide insights into areas of improvement.
Predictive Analysis: Use historical data to predict busy times and adjust staffing levels accordingly. This ensures that there are enough staff during peak times and avoids wastage during off-peak times.
Online Booking: For services that don't require immediate attention, offer an online booking or appointment system. This spreads out customer visits throughout the day and reduces unexpected crowding.
Customer Education: Use signage, digital displays, or staff to educate customers on how the system works, ensuring everyone knows where to go and what to do.
Integration with Other Systems: Ensure the QMS can be integrated with other tools, such as CRM systems, to offer more personalised services, or with workforce management tools to adjust staffing levels.
Incorporating these elements into a queue management system can dramatically improve the efficiency of service delivery, reduce waiting times, and enhance the overall customer experience.
Again, some generic QMS principles but one or two good ideas including virtual queuing and a feedback mechanism. As a customer experience consultant, the more feedback, qualitative or quantitative, the better. Of course, we could use ChatGPT or other LLMs to create insights from that data but that’s for a future post.
From airport departure hall to actionable and realistic recommendations.
Creating Digital Signage with ChatGPT 4.0 and Dalle-e 3.0
I was keen to test Dall-E 3, ChatGPT’s latest image creation tool and, based on the QMS recommendations, selected digital signage to focus on.
Subscribers to ChatGPT 4.0 Pro or Enterprise have access to Beta functionality and plugins for extra functionality.
Dall-e 3.0 is the latest image creation tool from OpenAI.
Using the model’s own copy, I started a new conversation, having selected Dall-E 3.0 from the dropdown available to Pro and Enterprise users of ChatGPT (see above) and submitted the following prompt:
Design four digital displays that show which counter or service point in the airport is available next, directing customers efficiently. This ensures customers are attended to in the order they arrive.
It takes approximately 30 seconds for the model to create and display the images, but it did indeed provide me with what I asked for; four digital signs, three of which were shown in the context of an airport environment.
One thing to be aware of is that images created are naturally futuristic and not always realistic. Furthermore, the spelling is questionable - something that seems to be an inherent problem with AI image creation tools right now.
However, the images created were what I’d asked for and here are the results.
Airport service counter digital signage proposed by Dall-e 3.
With my user experience hat on I had a problem as none of the proposed designs are what I would consider to be accessible or very well designed. SOme of the fonts were to small, content heirarchy rules not always adhered to and the colours, garish.
Finally, I asked the model to redesign the bottom right option for me using only black and white - take a look at the video below to see the final result!
Although, high level in parts, I hope this has given you an interesting insight into what ChatGPT can help you with in the airport environment. Do remember that you would need a subscription to Pro or Enterprise to get some of the functionality featured.
