BookMyShow has deployed more than 80 domain-specific AI agents on Databricks’ Genie platform. Business teams now query company data in natural language instead of filing requests with a central data team. The company says data engineering tickets have dropped 90% as a result. The numbers come from a Databricks case study published September 29, 2026. It is a vendor document, not an audited report.
Either way, the scale of the operation is real. BookMyShow claims 100 million-plus monthly active users. It says it sells roughly 22 million tickets a month across movies, live events, and sports. Chief Technology Officer Noel Curtis, who joined the company from US consumer tech, is the story’s narrator throughout.
From queue to self-service
The before-state will sound familiar to any large enterprise. BookMyShow ran a data warehouse on Amazon Redshift and a lake on S3 with Athena, Glue handling ETL. Every analytics request from marketing, finance, or live entertainment went through a central data team. That team wrote the SQL, delivered results, and moved to the next request. Capacity set the pace of analysis.
Two details sharpen the story. First, a 2023 AWS blog post showcased that same Redshift-and-Athena stack as a model migration. BookMyShow’s then-CTO co-authored the post, which claimed 80% lower analytics costs. Second, the case study credits Curtis with pushing the move to Databricks, Unity Catalog, and Genie. The bottleneck was never the old stack’s price — it was who could ask questions of it.
The design choice
BookMyShow deliberately skipped the single general-purpose assistant. It built 80-plus Genie Agents, each grounded in one team’s data context. Those domains span marketing, customer lifecycle management, live events, cinemas, and movie intelligence.
Genie One serves company-wide queries; power users graduate to Genie Code for deeper analysis. The marketing and CLM agent alone tracks over 3,500 daily offers. Those span payment promotions, promo codes, and loyalty programs.
The pattern extends outward too. Publisher integrations that took months of custom engineering now ship in about a day through Databricks’ OpenSharing. An IPL ticketing app went from one franchise to four within days. A single analyst built the Flask service on a serverless SQL warehouse.
What to watch
Every metric here is self-reported by the vendor and customer together. The 100 million MAU figure, in particular, sits well above BookMyShow’s last independent peak of 70 million in 2022. The 90% ticket reduction has no baseline or methodology attached. And the roadmap — real-time personalization, fraud and bot defense, natural-language support — reads as aspiration more than commitment.
The durable lesson is organizational, not technological. BookMyShow’s central data team now maintains pipelines, catalogs, and governance while business teams answer their own questions. That inversion — platform team up front, self-service behind it — is where the case study generalizes. India’s DPDP Act drives the governance emphasis. It will force the same reckoning on every consumer platform holding personal data.
Editor’s Note
This news item draws on a Databricks customer case study on BookMyShow, distributed on September 29, 2026. CIOL published syndicated coverage the same day. TechRecast independently verified Noel Curtis’s role as CTO and the 2022 monthly active user peak from Economic Times coverage. The prior AWS-based architecture comes from a 2023 AWS blog post co-authored by BookMyShow’s then-CTO.
All efficiency metrics, agent counts, and user figures are vendor and company claims. Databricks is the case study’s publisher and stands to benefit from its promotion.