Build with the AI Architect
- Step 1 · Open your development environment
- Step 2 · Open the AI Architect
- Step 3 · Say what you want
- Step 4 · See what it knows
- Step 5 · Confirm part by part
- Step 6 · Build the first part
- Step 7 · Build the next part, and see how it fixes a mistake
- Step 8 · See what was made
- Step 9 · Open what it made
- Step 10 · Revert all
Tell the AI Architect what you want. It works in small steps: it first reads Flow's guide and your workspace, writes down what it knows, and proposes the plan part by part: what it will make, in which order, and what stays for you. Only when you say ‘Build it’ does it build, part by part, with the same tools an AI app uses over MCP and never with more rights than you. You see every step live; a mistake Flow finds, it fixes itself; access it proposes instead of granting. Before it changes anything Flow records a version, so ‘Revert all’ puts everything back in one go.
Say what you want
Write in your own words what you want to happen and send it. The Architect first reads, once per conversation, Flow's guide and what is in your workspace; after that it works in small steps. Every step shows as a card. It proposes the plan part by part: the data first. In this phase it changes nothing.
Confirm part by part
Is part 1 right? Click Agreed, next part (or say what should change). Then it proposes part 2. Access (here: read access for the front desk) it never grants itself; it proposes it. An idea you did not ask for stands apart as a suggestion: only what you take along is built.
Build the next part, and see how it fixes a mistake
Part 1 is built. Click Build part 2. Flow checks every saved workflow. The first version searched on a column that does not exist: that card is orange, with the problem next to it. The Architect reads it, changes exactly that and saves again: now it is green. Its proposal for the front desk is there too.
This guide was carried out step by step on 22 Oct 2026 by an automated test on Flow (version 3e563176).





















