Coursera Spring Sale
40% Off Coursera Plus Annual!
Grab it
Learn how an AI agent successfully handled 60% of data questions when a head of data went on paternity leave in this 52-minute conference talk from the Computer History Museum's Coding Agents Conference. Discover the practical journey of deploying "Wobby," an AI analyst that became the unexpected backup for a logistics SaaS company with a 2.5-person data team. Explore why the team abandoned their web UI in favor of Slack integration, understand why BIRD benchmark scores proved irrelevant for real-world success, and examine how they built an evaluation system that caught actual failure modes rather than synthetic ones. Gain insights into critical technical decisions including context engineering, metadata design, and latency optimization, while also learning about equally important non-technical considerations such as channel design, user onboarding, and building trust with skeptical business users. Understand the hard-won lessons from taking an AI agent from concept to daily operational use, including what worked, what failed spectacularly, and what would be done differently in future implementations. Get a practitioner's perspective on agent deployment challenges and solutions from a software engineer specializing in AI and ML engineering who builds text-to-SQL agents and advises early-stage startups on AI/ML products.
