Systems | Development | Analytics | API | Testing

Two years ago, GenAI in production usually meant a single LLM serving a single endpoint. In 2026, it usually means much more. The applications shipping in front of users today are compound AI systems: orchestrated pipelines of retrievers, embedders, dialogue models, classifiers, code interpreters, SQL executors, and tools, with a single user request fanning out to several model calls across the stack.

An engineering leader asked me a question a few weeks ago: could we read their business workflows out of API traffic instead of asking people to document them? I said it should be possible. Then I tried it. A few engineers know how the system really works. They know which calls make up a work order and which checks happen after a write. That stuff rarely makes it into the test plan. Usually it’s in somebody’s head. Sometimes it’s in several heads, with slightly different answers.

The fundamentals of React Native over-the-air updates shouldn't sit behind a paid tier. Today we're shipping three additions to Bitrise CodePush: delta updates that help significantly reduce the size of your updates, CodePush code signing so you can secure every package you push to your users’ devices, and a dedicated Bitrise CI Step that publishes updates straight from your workflow. All three are available to every CodePush user, including everyone on the free plan.

Apple announced the M6 and M5 Pro on August 25, 2026, both inside the redesigned Mac mini. Bitrise Build Hub will be the first GitHub Actions runner alternative to add them to the fleet. That means your builds run on the newest Apple silicon before any other GitHub Actions runner can offer it. That's not a one-off. We've shipped new runners fast on every Apple silicon generation since M1 all the way up to M4 Pro, and M5 Pro and M6 are next.

A VuGen script is C. Some scripts are a recorded journey with helpers around it, others are a framework that happens to send a few requests. OctoPerf 17 ships a LoadRunner migration playbook for AI agents whose first job is to tell you which one you have. Real prompts, real output, script downloadable. Target: JPetStore, our public MyBatis demo shop. Every step is reproducible. One Action.c, HTTP recording level, eight transactions walking a purchase. Around them: It replays green in VuGen.

Customer service AI orchestration connects AI-driven intake to the backend systems and people who actually resolve a request, not just the chatbot that receives it. Every request should trigger an end-to-end resolution, not stop at an automated response. The real challenge with AI in customer service isn’t adoption; it’s fragmentation.

A platform can connect a dozen systems and still behave as if it connects none of them. The listings service reports one count of active properties, the analytics dashboard reports another, and finance keeps a third in a spreadsheet. Each number is correct inside its own system. None of them agrees, because every system carries its own definition of a Listing, a Property, and an Agent. This is usually diagnosed as an integration problem, so teams add another connector.

When information is spread across disconnected systems, it becomes difficult and time-consuming to find the right information, make informed decisions, and move work forward. Take, for example, planning a weekend getaway with friends: you have to review airline portals, manage a shared Google Doc for the itinerary, orchestrate a flurry of WhatsApp chats, and book an Airbnb reservation. Making decisions with this much information sprawl becomes exhausting.

Seven AgentSpot agents now augment every step of our content development pipeline, shifting our team’s focus from manual operational tasks to high-value content development and strategy. Our eLearning development team at ThoughtSpot manages around 50 courses and hundreds of videos across six learning paths on ThoughtSpot University. We build enablement content for external customers: business users, business analysts, data experts, and administrators.