Transforming Theater Research
Executive Summary
Playbill, the iconic 142-year-old theater magazine, partnered with Mission to develop an AI-powered research assistant that revolutionizes how content creators access decades of theatrical history. Mission delivered a chatbot using Amazon Bedrock that provides accurate, cited responses while maintaining the data integrity critical to Playbill's editorial standards and reputation.
About Playbill
Playbill produces the show programs distributed at theaters across America. The renowned brand has served the theater industry continuously, building deep archives of shows, performers, and productions that span generations of theatrical history.
Scale faster without breaking your cloud
Background
Playbill maintains extensive archives documenting actors, shows, and theatrical productions spanning decades. Content authors face daunting research challenges when writing about performers with 40, 50, or 60-year careers across multiple productions.
Challenge
Researching information for Playbill's media website proves increasingly difficult as the depth of their historical knowledge grows. Writers need to verify details about actors with multi-decade careers, cross-reference show histories, and confirm facts across disparate data sources. The research process often requires four, five, or even twenty separate searches to gather comprehensive information for a single article. Accuracy is paramount as well—the publication can not afford a single hallucination or error, particularly regarding sensitive information like whether someone has passed away. The gating challenge centers on data accuracy and preventing any false statements that would damage Playbill's credibility and trust with the theater community.
Why Mission
Jon Goldman, Chief Technology Officer of Playbill, has worked with Mission for seven years across multiple engagements. The long relationship has flourished from consistent quality delivery across projects ranging from basic AWS hosting to advanced AI development. Mission's track record keeps Goldman from considering other partners. The value proposition extends beyond technical capability to genuine trust built over years of collaboration.Why AWS
AWS provides the foundational services that make Playbill's research bot possible. Amazon Bedrock delivers the large language model capabilities required for natural language processing and intelligent responses, while Amazon Aurora offers the database performance needed to query decades of theatrical data efficiently. AWS Glue consolidates disparate data sources into unified formats accessible to the AI system. Container services enable responsive front-end interfaces. This combination of AI, database, and integration services positions AWS as the platform capable of handling Playbill's unique requirements for accuracy and citation.
Solution
Mission is delivering a proof-of-concept research chatbot using Amazon Bedrock as the foundation for natural language processing. The chatbot pulls from four data sources, including a graph database and web search, using a multi-agent pattern that employs 39 different tools to dynamically create and iterate queries across data sources.
The team works closely with Playbill to understand the unique requirements of theater research and the critical importance of accuracy. Together they explore Playbill's data landscape, discovering depths of information that the organization hadn't fully catalogued.
The solution incorporates multiple data sources including Amazon Aurora databases, containerized front-end services, and AWS Glue for consolidating disparate information. A key differentiator has been the citation functionality—every response includes references to source materials, allowing writers to verify information and understand exactly where each fact originated. This citation approach addresses Playbill's zero-tolerance policy for hallucinations by making responses traceable and verifiable.
Mission's collaborative process helps Playbill understand what all is possible with this chatbot, along with future AI possibilities. The team approached data discovery as an exploratory journey, helping Playbill understand the full scope of archives available for integration. Container services provided the UI foundation while AWS Glue unified data from multiple locations. The architecture balances performance with accuracy, creating a system that can serve production editorial workflows and other potential AI applications.
Cost consciousness has guided the technical approach. Mission works within Playbill's budget as a non-enterprise company, focusing on leveraging pre-existing database services rather than expensive new processing models.
Results
The proof-of-concept so far has demonstrated surprising accuracy that impresses even Playbill's most knowledgeable theater staff. Functional completeness and accuracy have been primary goals, to which the chatbot is returning results in less than 3 minutes for even the most complex queries. Writers discover the chatbot can answer questions correctly and then provide additional context beyond their own expertise. This depth proves the system's potential value for research workflows.
The next phase focuses on moving from POC to production. The team will optimize response times by reformatting data structures and improving how Amazon Bedrock accesses information. Speed improvements will unlock the chatbot's practical value for daily editorial workflows. Once deployed internally, the system looks to dramatically reduce article research time and improve fact-checking accuracy.
Playbill sees the research bot as a huge advancement into AI territory that has not yet been seen with live theater. The goal is making AI a companion to theatrical arts rather than attempting to replace human performers or critics. Playbill looks forward to faster article production, better verification workflows, and editors empowered with deeper knowledge at their fingertips.
AWS Services
- Amazon Bedrock
- Amazon Aurora
- AWS Glue
- Amazon Elastic Container Service (ECS)
- Amazon S3