AI TOOLS
In short: Most people type "create the app for me" straight into Claude Code; starting in planning mode with a written development plan produces better apps with less time and fewer tokens.
The output of Claude Code is defined by your input, and your input lives in the development plan. Prompted with a one-liner, the tool will happily build an application you are only partly happy with; prompted through planning mode, it builds against an agreed specification, as this demo of a "celebrity friend" web app shows.
Planning mode, activated by typing a forward slash and choosing plan, makes Claude Code write a development plan before touching any code. Your application is going to be as good as your plan is, so approving the plan first saves time and tokens compared with iterating on finished code. The demo's first prompt: create a product specification for the app, use the AskUserQuestion tool to clarify requirements, output a PRD with the full feature list.
The AskUserQuestion tool, described in the Claude Code documentation's best-practices section, makes Claude interview you instead of assuming your requirements. In the demo it asked about the product's core concept, target audience, and intended business model, surfacing things you might not have considered yet: technical implementation, UI/UX, edge cases, trade-offs. If you have no answer to a question, ask Claude to propose motivated recommendations and pick one.
Claude Code has a context window, a limit on how much text can be held in one session, and when that limit is reached a new context window starts. Without a file tracking which features are built and which remain, Claude will not know where to continue in the fresh window. Asking Claude to create and maintain a process.md file at plan-approval time gives every future session a reliable to-do list.
| Workflow step | What it prevents |
|---|---|
| Enable planning mode before coding | Building an app you are unhappy with, then burning tokens on rework |
| Prompt with the AskUserQuestion tool | Claude guessing at audience, business model, and edge cases |
| Create a process.md file | Losing track of progress when the context window resets |
| Test every feature right after building it | Discovering broken features only at the end of the build |
Ask Claude Code to write a test for every single feature it creates and to run that test immediately after the feature is built. In the demo, one message approved the plan, requested the process.md file, and required per-feature tests, and Claude confirmed with "ready to execute".
Ani Björkström, a Stockholm-based tech consultant, summarizes the workflow in four rules: plan first, let Claude ask the questions, track progress in a markdown file, and test each feature as it lands.
Type a forward slash in the Claude Code prompt and choose plan. Claude then drafts a development plan and asks "Would you like to proceed?" before writing code.
No. The demo prompts were drafted in the Claude chatbot and pasted into Claude Code; ChatGPT or any other LLM works the same way.
When the context window fills up and resets, Claude Code loses its picture of which features exist. The process file is what lets it resume exactly where it stopped.
0:00 Hi, my name is Annie. I'm a tech consultant based in Stockholm. In this video, I will show you how to create better apps using less time, less tokens with cloud code. I'm also giving away a free week of cloud code access. Stay until the end of this video to know more. We will go into file explorer. We'll create a folder that will be called celebrity friend. And then from here, we will right click, open in terminal. Then you can see that this folder will open in terminal and from there we will write cloud. This will ask if we want to be able to read, edit and execute files and we will click on yes. Okay.
0:45 So right now cloud is initiated and it will have access to all the files in this catalog. Now most people in this stage will write create celebrity web application for me and we let cloud to do the job and once the application is ready you will realize that there are different things that you are not happy about. So that is why instead of jumping into developing the code we will create a development plan. So the reason we want to create a development plan is because we want cloud code to follow the features specified in the plan because in the end of the day your application is going to be as good as your plan is or in other words the output is defined by your input and input is specified in your development plan.
1:35 And in order to create a development plan, we will create forward slash and here we will choose plan. And plan will enable planning mode. And the idea is that first we will create a plan for our application. And once the plan is fine, we will ask cloud to create the actual application. In this way, we will save a lot of time and tokens. So in here we will enter our prompt. I created my prompt using cloud chatbot and of course you can use chat GPT or any other LLM of your choice. This is the first prompt that I will use. Create product specification for my app. Use ask user tool to clarify requirements in best practice section inside cloud code documentation.
2:25 Cloud is mentioning ask user question tool. So basically if you include this tool in your prompt cloud will ask you about things you might not have considered yet including technical implementation UI UX edge cases and trades offs. So that is why it is important to include this tool in your prompt. Just copy past this prompt into cloud code. So it says create production specification for app and here we need to specify the app and our app will be web celebrity friend use ask user tool to clarify requirements and then output will be prd which is product description file with full spec and feature list.
3:13 So we will click enter in here. Cloud is saying I will help you create PRD for celebrity friend application. Let me start by understanding your vision for this product. And the first question will be what is the core concept of celebrity friend? Is it an IE chatbot that simulates conversation with celebrities, a social platform connecting fans with celebrities or something else? Our product is something else than the options that cloud mentioned. So I have to specify what it is. And now this is the second question. Who is the target audience for this product? We have Gen Z, general consumers. I want to go with option two. Next one is what is the intended business model. Cloud will basically continue asking many more questions to understand what are your vision for the product and if at some point you don't have an answer for the question.
4:04 You can just ask cloud to come up with recommendations and motivate recommendations so that you can choose one of those recommendations and in the end it will create a development plan for you. Okay, we have a result. You can see the plan that cloud created. And if we read through the plan and come down in the end of the plan, we have a question from cloud. Would you like to process? So in this stage, we can of course read the plan understand if this looks fine for us. If something isn't looking good, then we can just iterate. I will go to step four that is saying type here to tell cloud what to change. So basically it's not that I want to change something but I want to give instructions.
4:50 So in the next step I want to approve the development plan and at the same time I want cloud to create a process markup file and the next step is build the app according to this plan. Then I also want to ask cloud to create a process MD or markup file. The reason is because cloud code has a context window which means that there is a limit on how many text can be sent into the cloud code. And once the limit for context window is met, a new context window is created. And if we do not have a process file that is controlling all the created features and all the features to be created then cloud will not know where to continue in the next context window.
5:39 So that is why it is very important to include a process markup file. And also the third thing that I want to mention at this point is that I want cloud to create a test for every single feature it creates. And I want cloud to test the feature after the feature was created. So I will copy a prompt from cloud chatbot and I will paste it in here and then I will click on enter button. Okay, we got a new answer from cloud and we can see the plan. We can see the project structure and it asks cloud has written up a plan and is ready to execute. Would you like to process? And at this point we will click on yes button. Now to summarize workflow first remember to use planning mode and create a development plan.
6:30 Next ask cloud to use user testing tools to ask you a lot of questions. Second ask cloud to create process MD file to keep in track on the process. And last but not least, ask cloud to test every single feature after developing the feature. Now, if you want to try cloud code for free, comment cloud code and I will choose one commenter randomly and give away one week free cloud code Access.
Want this working inside your finance team?