Improving the EG Booking Experience

#

As part of a project for EG, we explored ways to improve the booking process for exhibitors applying for stands at Engestofte Julemarked.

The existing workflow relied heavily on manual communication between exhibitors and administrators. Applications were submitted, reviewed manually, and stand placements were often coordinated through email conversations.

The goal of this project was to investigate how software could streamline that process while improving the experience for both exhibitors and administrators.

Live Demo

#

🔗 Deployed Project: https://eg-booking-project-five.vercel.app

Designing a More Realistic Booking Flow

#

Instead of treating stand booking like a simple checkout process, I wanted the system to reflect how event applications are handled in reality.

The booking flow was redesigned so exhibitors first submit information about themselves and their business before choosing stand locations.

This allows the system to understand:

• Who the exhibitor is
• What products they sell
• What type of stand they may require
• Which locations might be most suitable

After completing the application, exhibitors select three prioritized stand wishes rather than a single stand.

This mirrors the real-world process where organizers need flexibility when assigning placements.

The flow now works like this:

1. The exhibitor fills out the application form
2. The application is analyzed
3. Stand recommendations are generated
4. The exhibitor selects three prioritized stand wishes
5. The application is submitted for review
6. An administrator processes the request

Interactive Venue Maps

#

One of the biggest additions was replacing generic stand selection with venue-based floor plans.

Instead of selecting from a list, users can browse the actual venue areas:

• Outdoor Area
• Hestestalden
• Kostalden
• Jagtstuen
• Laden

Each area contains its own interactive floor plan with clickable stand locations.

This makes the booking process feel much closer to the real event layout and gives exhibitors a much better understanding of where their stand will be located.

The experience ended up feeling surprisingly similar to booking seats in a cinema, but adapted for a large event with multiple buildings and stand categories.

AI-Assisted Stand Recommendations

#

One of the most interesting parts of the project was experimenting with AI-assisted stand recommendations.

The goal was never to let AI make final placement decisions. Instead, AI is used to help understand the exhibitor’s application.

For example, if an exhibitor writes that they sell:

Port wine, sparkling wine, and specialty beverages

the AI can classify the application into categories such as:

Wine
Indoor Beverages
Premium Food & Drink

These classifications are then combined with rule-based logic.

The recommendation engine considers factors such as:

• Indoor vs outdoor suitability
• Stand availability
• Product category distribution
• Similar vendors nearby
• Stand size requirements

This hybrid approach turned out to be much more reliable than relying entirely on AI.

The AI helps understand the exhibitor, while deterministic business rules handle the actual recommendation logic.

If the AI service is unavailable, the system can fall back to keyword-based categorization to ensure the booking flow still functions.

Stand Management and Priorities

#

The booking system keeps track of:

• Available stands
• Booked stands
• Stand categories
• Pricing
• Indoor and outdoor locations
• Recommended stands
• User-selected stand wishes

Exhibitors are required to select three prioritized stand wishes:

1. First choice
2. Second choice
3. Third choice

If a stand is removed, priorities automatically adjust.

This reflects how many real-world event applications are handled and gives administrators more flexibility when assigning placements.

Administrative Workflow

#

An administrative interface was also added to support the review process.

Administrators can:

• Review exhibitor applications
• View stand preferences
• See AI-generated recommendations
• Approve stand assignments
• Suggest alternative locations
• Manage incoming requests

At the current stage, authentication has not yet been implemented, so users are automatically treated as administrators for testing purposes.

AI-Generated Email Notifications

#

Another feature currently being explored is AI-assisted communication.

When an application is approved, booking information can be sent to an AI model that generates a personalized confirmation email.

The workflow looks like this:

1. Admin approves an application
2. Booking details are provided to the AI
3. The AI drafts a confirmation email
4. The administrator reviews the draft
5. The email is sent to the exhibitor

The goal is not to fully automate communication, but to reduce repetitive administrative work while still allowing human review.

What I Learned

#

This project highlighted how quickly complexity grows once a workflow becomes realistic.

Building stand selection itself was relatively straightforward. The difficult part was understanding how all the pieces connect:

• Applications
• Venue layouts
• Recommendations
• Administrative review
• Communication
• Business rules

I also learned that AI works best when given a clearly defined responsibility.

Using AI to classify exhibitors and assist recommendations felt practical and reliable. Allowing AI to make final placement decisions would have introduced significantly more uncertainty and made the system harder to trust.

Another takeaway was the importance of user experience. Small improvements such as interactive maps, stand recommendations, and clear application workflows can dramatically improve how a system feels to use.

Future Improvements

#

The project is still a prototype, and several important pieces are still missing:

• JWT authentication
• PostgreSQL database integration
• Persistent booking data
• Persistent exhibitor applications
• Production-ready email delivery
• More advanced recommendation rules
• Better administrative tools

With those additions, the project could move much closer to becoming a production-ready exhibitor booking platform.

Conclusion

#

This project was an interesting example of how software can improve a business process without fundamentally changing how people work.

The goal was not to automate everything, but to reduce friction, improve communication, and make the booking process easier for both exhibitors and administrators.

What started as a stand booking prototype gradually evolved into a much more complete workflow involving applications, recommendations, administrative review, and AI-assisted communication.

More importantly, it reinforced something I have noticed in several projects: the hardest problems are often not technical. They are understanding workflows, understanding users, and finding ways to make their jobs easier.