Table of contents
About the client
Fundacja Nowa Muzyka Katowice organises Tauron Nowa Muzyka, an annual electronic music festival held in Katowice, Poland.
Now in its second decade, the festival attracts Polish and international audiences across four days of techno, house, electronica, and experimental music at the Katowice Culture Zone, a landmark post-industrial cultural venue. Despite a strong artistic reputation and a loyal fanbase, the festival’s organisers identified a pressing challenge heading into 2026: audience growth had plateaued.
Business challenge
Festival lineup planning is a complex, high-stakes process. Artistic directors must balance genre diversity, booking costs, audience appeal, and day-by-day programme flow, while also assessing which combinations will resonate with different demographic groups. Historically, this process relied entirely on experience and intuition, with no systematic way to model how lineup choices would affect audience composition or how to tailor promotion for different segments.
The Tauron Nowa Muzyka team needed a tool that could generate concrete lineup proposals based on genre, budget, and strategic direction; profile the audience segments each configuration would attract; suggest targeted promotional approaches; and estimate the proportion of new versus returning attendees, all usable by non-technical staff without any setup.
Solution
Future Processing was engaged to design and build a purpose-fit AI planning tool on AWS. The AWS Partner selected a fully serverless architecture to match the festival’s usage pattern: intensive during planning cycles, idle in the off-season. This approach eliminated fixed infrastructure costs and the need for ongoing server management.
The frontend is hosted in Amazon Simple Storage Service (Amazon S3) and distributed globally through Amazon CloudFront (CloudFront), with all traffic secured via HTTPS and routed through an Origin Access Control policy. The backend is a REST API exposed through Amazon API Gateway (API Gateway), backed by an AWS Lambda (Lambda) function running Node.js 20.x. Lambda handles lineup generation requests and retrieval of generation history.
AI inference runs through Amazon Bedrock (Bedrock), AWS’s fully managed service for foundation model access, using Claude Haiku 4.5 via Bedrock’s EU cross-region inference endpoint. Accessing the model through Bedrock means no external API keys are required, authentication is handled entirely through AWS Identity and Access Management (IAM), with no credentials to rotate or secrets to store.
Every generation request and its result are stored in Amazon DynamoDB (DynamoDB). A caching layer returns stored results for repeated queries within 90 days, eliminating redundant inference calls. To handle the 10–20 second generation time without holding open HTTP connections, the solution uses an asynchronous job pattern: Lambda returns an immediate HTTP 202 response and re-invokes itself to perform inference while the frontend polls for the result.
The entire stack, Amazon S3, CloudFront, API Gateway, Lambda, DynamoDB, and AWS Secrets Manager (Secrets Manager), is defined in a single AWS Cloud Development Kit (AWS CDK) stack written in TypeScript and deployable with a single command.
Results and Business Impact
The application was designed, built, and deployed to production in 3 weeks, from initial scoping to a live, production-ready system.
Festival organisers can now generate and compare complete lineup scenarios in under 20 seconds, a process that previously required days of manual research and internal deliberation.
The total monthly running cost remains negligible, low enough to be treated as a minor operational line item in any festival’s budget, with Amazon Bedrock accounting for the largest share and all other AWS services contributing only marginally. The serverless model ensures the festival pays nothing during off-season periods when the tool sits idle. No server management, patching, or infrastructure monitoring is required.
The architecture is designed to extend beyond a single event: the same platform can be reconfigured for other festivals or cultural programmes by updating the context layer alone, with no changes to the underlying infrastructure.
For the first time, we can explore dozens of lineup scenarios in a single meeting instead of weeks of analysis and debate
Technology Stack
- Frontend: Vanilla HTML / CSS / JavaScript
- CDN & Hosting: Amazon CloudFront, Amazon S3
- API Layer: Amazon API Gateway (REST)
- Compute: AWS Lambda (Node.js 20.x)
- Database: Amazon DynamoDB
- Secrets Management: AWS Secrets Manager
- AI / LLM: Amazon Bedrock (Claude Haiku 4.5)
- Infrastructure as Code: AWS CDK (TypeScript)
- Region: eu-central-1 (Frankfurt)
About Future Processing
Future Processing is a tech strategy and software delivery company headquartered in Gliwice, Poland. With over 25 years of experience and 900+ projects delivered, the company takes a consulting-led, outcome-based approach to turning business ambitions into measurable results. Future Processing helps organisations assess AI readiness and implement AI solutions in a structured, measurable way, applying AI across the full delivery lifecycle, from advisory and architecture through to deployment and ongoing support, for clients across Europe and beyond. Future Processing is an AWS Advanced Partner (Consulting Partner).
Core Technologies
