Processing...
NIX developed an AI-powered solution for a ticket resale marketplace. This comprehensive LLM predicts ticket prices at event venues with 88% accuracy, resulting in a 76% increase in investments for the company.
Retail and E-commerce
AWS, Data Engineering, AI, Cloud Services
Anthropic Claude, AWS Bedrock, Lambda, S3, CloudFront
Our client is a ticket resale marketplace for events, operating in the North American and European fields. To gain a competitive advantage and fill the niche, they came up with the idea of developing an AI-powered platform that can provide users with accurate data-driven predictions of future ticket prices based on factors like venues, dates, area of the venue, seating, and the popularity of the event.
The client presented the concept to potential investors, and one of them, AWS, recognized the immense potential and opportunity to expand its long-term user base and decided to fund the development of a proof of concept (PoC). AWS also advised the client to engage NIX, one of its time-tested partners with extensive experience in AI and large language models (LLMs), for rapid development and execution of the client’s needs.
Price Break Levels utilize the visualization of the venues and the different sectors to provide price thresholds at which ticket prices change significantly and when users can make decisions on purchasing or selling their tickets.
Ticket Price Forecast, based on a variety of critical factors, including venue specifics, seating sections, and event time.
The main complexity during production was the highly expanded scope of the project and required features, which could not be accommodated within the limited PoC scale, prompting us to devise new workarounds within the budget. Additionally, the crucial need for the data engineering pipeline for processing and data storage could not be covered in the PoC scope. Despite this, our team succeeded in testing various approaches and hypotheses and finding the best solution for the client.
Since the client set the ticket prices for the different sections and settings manually, the main objective was to determine the possibility of an AI solution to automate their workflow. This AI solution would use historical data and data analysis to accurately predict ticket prices based on the type of event, venue, dates, sector, and setting.
When approaching AI and LLM implementation, our specialists selected the Claude AI agent for its efficiency, high performance, availability, and seamless integration with the AWS infrastructure. This allowed us to build a solution that wasn’t just accurate but also conversational. Instead of receiving dry, mathematical answers, users get a human-like response, which makes the experience more engaging.
The concept of price break levels was designed to give users a clear, visual understanding of ticket pricing dynamics. The initial idea was to allow users to upload a simple image of an event venue’s seating chart to a chatbot. The system would then use AI and LLMs, enhanced with computer vision, to identify different sections and provide accurate price forecasts for each.
The main challenge was the lack of essential data in a basic venue image, which made it impossible to make accurate predictions. We overcame this limitation by implementing the entire dataset within the PoC, enhancing the system with structured data in JSON files. These files consisted of statistical info on maximum, average, and minimum prices for each section.
This allowed the LLMs to access and compare this crucial data, leading to a much higher level of accuracy. Essentially, the final solution analyzes the venue’s visual layout and cross-references it with a comprehensive, data-driven JSON file that clusters sections by their price similarities. This combined approach ensures the Price Break Levels feature provides the client and their users with reliable and valuable insights.
When developing accurate ticket price forecasting, we analyzed available historical data and found the most relevant and similar examples. To ensure its accuracy, our team meticulously filtered the data by specific attributes like sector, row, seating, and time remaining until the event. We then sent it to the LLM, which enhanced the information with its existing knowledge base before generating a comprehensive response.
A key technical challenge was the absence of a robust, real-time data pipeline within the PoC’s scope. We addressed this by leveraging Amazon S3 JSON files to store all the necessary data. However, the raw data was too unorganized for the AI agent to use effectively, so our team performed extensive data cleaning and deduplication to ensure the information was structured and reliable.
Similarly, we encountered performance issues with processing a large number of sizable JSON files, which significantly increased execution times in AWS Glue. To address it within the PoC limitations, we refactored the AWS Glue job code and optimized several data-processing methods, leveraging the pandas library for more efficient handling of large datasets. These changes reduced processing overhead and made working with voluminous JSON files more manageable.
The successful completion of the PoC delivered the client a compelling roadmap for the full-scale development of the product. Within the one PoC, we actually tested two different functionalities and two different AI approaches: LLM-based and classic AI. The client chose the classic AI option, and the PoC engaged new partners and attracted additional investments crucial for further development.
Cloud-based Solutions to Enhance HR and Behavioral Health Platforms
Healthcare
AI-Driven Application for Mental Health Support in the US
Global Betting Leader Reduces Latency by 25% with AWS Migration
Gambling & Casinos
AWS-powered Personal Safety iOS App
Entertainment
AI Integration: Streamlining Accreditation Workflow in Education
Education
Mobile App to Streamline Teacher-Student Workflows
Discover NIX AI chatbot solutions to boost engagement and efficiency. Custom development from PoC to enterprise. Learn more here!
Discover AI agent development solutions tailored to your needs. Our expertise in AI agent development helps businesses automate tasks and drive innovation.
Partner with NIX for expert machine learning development. Our ML experts leverage the latest tech stack to create ML solutions tailored to your business needs.
Reduce your AWS costs by 30% and enhance cloud performance with NIX. We take into account your short and long-term needs to maximize business results. Explore!
Harness the power of generative AI with our end-to-end services that include AI strategy, model training, deployment, and integration.
NIX provides end-to-end AWS migration services—from AWS cloud migration consulting to post-support. Learn how to grow cloud performance and keep costs in check.
At NIX we provide top-notch cloud consulting services empower clients to adopt cloud and gain agility, cost-efficiency and security of their infrastructure.
AWS consulting and development services from NIX is your way to join the cloud seamlessly and cost-efficiently. Stay on 24/7 with a certified partner.
NIX is a software engineering company in the USA that offers enterprises digital transformation consulting services to embrace the future and growth.
Schedule Meeting
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.
