ENG2609. Building Automations Live (TCH, NAA)
One hour. One real accounting problem. One working automation, built live in front of the room.
This is a live-coded build session with no canned demo, no pre-written code, and no safety net. Using Claude Code and the most relevant AI tooling available at the time of the conference, we'll go from a blank screen to a functioning tool, with the audience along for every step: the prompts that work, the ones that don't, the course-corrections in real time, and the moments where the tech does something genuinely surprising. Expect the energy of a cooking show and the stakes of a live performance.
Along the way, you'll see the decisions that matter most. How to scope a problem small enough to ship. When to trust the AI and when to intervene. How to connect to the systems accountants actually use. And what "good enough to use on Monday" really looks like. The specific build will reflect whatever tooling and capabilities are most relevant the week we take the stage, because in this space, six months is a lifetime.
The real takeaway isn't the tool itself. It's an unfiltered, end-to-end look at what a single practitioner with the right AI tools can now build in the time it takes to run a morning stand-up, and a concrete sense of what that shift means for how firms should be thinking about software, staffing, and service delivery in the year ahead.
Learning Objectives:
- Describe the current capabilities and limitations of AI coding tools such as Claude Code, including where human judgment and domain expertise remain essential to producing reliable output.
- Apply a repeatable approach to scoping, prompting, and iterating on an AI-built tool, so they can take a workflow problem from idea to working prototype without writing code from scratch.
- Analyze build-versus-buy decisions in a world where AI-native development has dramatically lowered the cost and time required to build custom tools, and determine when building in-house now makes more sense than purchasing commercial software.
