How Sala Entertainment Put AI to Work for 150,000 Customers, and Its Own Engineers
Sala Entertainment needed to support a fast-growing user base without a fast-growing headcount, and keep its engineering team shipping without sacrificing code quality. Two AWS Bedrock-powered systems solved both at once.
Real results. Not just projections.
Two bottlenecks, one growing pain.
Sala Entertainment runs one of Saudi Arabia's most popular ticket booking and entertainment platforms, letting more than 150,000 active users discover events, book tickets, manage loyalty rewards, and top up digital wallets, in both English and Saudi Arabic. That kind of growth is a good problem to have, and it's still a problem: more users means more support questions, and a bigger, faster-moving codebase means more pressure on code quality at exactly the moment there's less time to enforce it manually.
The first bottleneck was customer-facing. Sala's app has a lot of surface area, ticket booking, wallet balances, loyalty programs, spread across multiple pages, and customers were struggling to navigate it. That pushed a steady stream of repetitive questions onto the support team, and because Sala serves a genuinely bilingual audience, a generic FAQ page was never going to cut it.
The second bottleneck was internal, and less visible to customers but just as costly. Every pull request against Sala's Magento 2 backend needed a manual code review, and those reviews were taking 30 to 60 minutes each, with best practices, security standards, and coding conventions applied inconsistently across reviewers.
Two AI solutions, built to work together.
Rather than treating these as separate problems needing separate vendors, we designed both solutions on the same foundation, Amazon Bedrock, deployed in the AWS Middle East (Bahrain) region, with serverless, event-driven architecture so both scale automatically without ongoing operational overhead.
The customer-facing solution is a conversational AI assistant built on AWS Bedrock Agent with Nova Premier, embedded directly in the Sala booking app, with Amazon Transcribe handling voice input for customers who'd rather ask than type. It answers the repetitive questions, booking status, wallet balance, loyalty program details, directly and in the customer's language.
The internal-facing solution triggers automatically on every GitHub pull request against Sala's Magento 2 backend, analyzing code quality, security, and adherence to Magento best practices in seconds rather than the 30 to 60 minutes a manual review used to take. It doesn't replace human judgment on architecture decisions, but it catches the consistent, checkable issues immediately, before a reviewer's time gets spent on them.
What actually changed.
On the customer side, the chatbot now handles a meaningful share of Sala's support volume directly, in either language, without the support team growing headcount to match user growth.
On the engineering side, automated code review turned a 30-to-60-minute manual bottleneck into a process that completes in seconds, a 99% reduction in review time, freeing engineers to focus on the judgment calls that actually need a human and letting the team ship with more consistency and confidence.
One AI foundation, across every touchpoint.
Sala's ticket booking and entertainment platform sits at the center of a multi-brand operation, so the chatbot and code reviewer needed to hold up under real, everyday production load, not a pilot environment. Deployed on the same Bedrock foundation in Bahrain, both systems scale automatically as usage grows across the platform.
Bilingual support, built into the app itself.
The chatbot lives directly inside the app customers already use to book tickets and manage their wallet, not a separate help center they have to go find. Ask a question by typing or speaking, in English or Saudi Arabic, and get an answer in the same conversation.
“By leveraging AWS Bedrock, we transformed both our customer experience and development workflow, delivering AI-powered support to 150,000+ users while our developers ship faster with confidence.”
- Sala Entertainment
What This Approach Gets Right
One foundation model platform can solve two very different problems
Sala's customer support and code review needs looked unrelated on the surface, but both were solved by the same Bedrock foundation, deployed once and scaled to two use cases.
Bilingual support has to be built in, not bolted on
A generic FAQ page was never going to serve a genuinely bilingual, Arabic and English audience. The AI assistant answers in whichever language the customer used to ask.
Automating checkable review issues frees humans for judgment calls
The code reviewer doesn't replace human architecture decisions, it clears the repetitive, checkable issues immediately so reviewers spend their time where it actually matters.
Want AI That Actually Fits Into How Your Team Already Works?
SnapTec builds production-grade AI solutions on AWS, from customer-facing chatbots to internal developer tooling, tied to a real business or operational outcome, not deployed for novelty.
Talk to Our AI Team