How you approve your app and every change to it
The AI engineers build and check your app, but you make the call on what goes live. Here’s how the first release works, how a change goes from an idea to Test, and what Test and Live mean today.
When you finish the wizard, the AI engineers build your app from its Charter. Before your app is called live, two things happen: Orbulus checks the build for itself, and then you review it. The same idea runs through every change you make afterwards.
The first release
While the AI engineers work, your app shows as Building. When they report the build complete, it moves to Checking: Orbulus checks for itself that the deploy succeeded and that the deployed app serves your features. If those checks fail, the release is held, the AI engineers get the details, and your app shows Checks failed until it’s fixed.
When the checks pass, your app is Ready for review. You get a link to try it, the automated checks it passed, and a checklist built from your Charter, with a line for each feature you asked for.
- Approve and go live
- Available once you’ve ticked every line. Your app becomes Live and you get an email to say so
- Request changes
- Lines you leave unticked are sent as not met, with anything else you write. The AI engineers build again
So the last step before your app is called live is always yours.
Changing a live app
Once your app is live, you change it with change requests, from your app’s page. A change request is three things in your own words: a title, what should change and why, and how you’ll know it works.
A change request is discussed, built and tried on Test, then goes live once you approve it
Discuss
The Product Owner, one of the AI engineers, reads your request and asks anything it needs before anything is built. As you answer, it updates the request: it may sharpen the description or add acceptance criteria. It also shows what it checked, such as data you already have, who can use the feature, money, notifications, and legal and privacy. Discussing doesn’t count against your changes.
When the Product Owner has everything, you choose Implement now or Save for later. Implement now uses one of your changes for the month.
Build
The AI engineers then work through the change, and its page shows where they are:
- Plan the change against your app’s current code
- Check security and privacy, when the change needs it
- Write the code and its tests
- Run the checks, fixing anything that fails
- Review the new screens, when there are any
- The Product Owner checks the acceptance criteria
- Deploy to Test
If they hit a question only you can answer, the build pauses and shows Needs your answer. Answer it, and they carry on from where they stopped.
Try it on Test
When the change is on Test, open the Test site and try it against your acceptance criteria. Then choose:
- Approve and go live
- The change is merged into your app and added to its brief, so later changes build on it
- Reject
- Say what’s wrong. It goes back to the Product Owner with your note, and building it again doesn’t use another change
If the AI engineers can’t finish a change, it shows Failed. You can try again, which doesn’t use another change, or take it back to the discussion. You can also abandon a change at any point before it goes live, and nothing from it goes live.
Your app has one Test, so one change is built and tried at a time. If you start another, it waits its turn and shows Queued for Test.
Test and Live today
Your dashboard shows each app’s Test and Live side by side. Today every app has one website, and that website is its Test. When no change is being tried, Test runs the live version of your app. While a change is on Test, the website runs the change, and if you reject or abandon it, Test goes back to the live version. Live shows as Not promoted yet.
What stays with you
- Approving the first release, once every line on its checklist is ticked
- Approving or rejecting each change on Test
- Choosing when a change is built, and abandoning one you no longer want
The AI engineers do the building, checking and deploying in between.