Module 1: Four ways AI pays · Lesson 1.4· Free

1.4 Start with the client, not the code

Watch alongside the Lesson 1.4 video. Most AI tools jump straight to building. EAI starts the way a good consultant does: with your client.

Key terms3
Context

Everything the AI knows about your situation: the client, the problem, the goal. Better context gets a better build.

e.g. Who Coastal Snacks is, the second-visit problem, and what good looks like.

Prompt

The message you type to the AI. In the course, an opening prompt sets up a whole conversation.

e.g. The opening prompt for the opportunity hunt, under the 3.3 video.

Client first

EAI starts by understanding the client and the problem, then builds. The usual way is to ask for a feature straight away.

e.g. "Help me understand the problem before we build anything."

TranscriptClick a time to jump there

Jumping straight to building

This is my favourite lesson. Most AI tools jump straight to building. EAI starts the way a good consultant does: with your client.

The usual way vs the EAI way

Here’s the difference. The usual way, you type “build me a maintenance form”. You get a form. It might even look nice. But nobody asked whether a form fixes the second- visit problem. The EAI way, you tell it about your client. Who they are, what’s going wrong, and what good looks like. EAI asks questions back, builds up the context, and then builds the right thing. Same as you’d do sitting across the table from a client.

Here’s how I open

Here’s how I open. “I want to build an app with you. I’m working with a client called Coastal Snacks. They run 240 vending machines across Sydney, and one in three maintenance call-outs needs a second visit. Help me understand the problem before we build anything.”

Notice what’s in there. The client’s name. What they do. The problem, with a number on it. And I’ve asked it to understand the problem before it builds.

Watch what it does

Watch what it does. It asks questions back. Who’s involved? Technicians, dispatch, and the gym or office where the machine lives. What happens today when a machine breaks? Why do you think they need a second visit? And what does good look like?

I answer like I’m the client. Here’s where our hunch comes in: technicians turn up without the right part because the fault report is vague. And good looks like fixed on the first visit, most of the time. Every answer is context. That context is what makes the app fit your client, not just look like an app.

Recap and what’s next

So: client first, the problem with a number on it, and let EAI ask. Next, we watch that conversation turn into a working app.