Rapid prototyping cycle
1
Understand

Capture the problem, not a spec document

2
Prototype

AI / v0 / Figma — something clickable

3
Review

Walk through it with the client

4
Approve

Explicit yes/no with remarks

5
Build

Implement exactly what was approved

6
Deploy

Ship and record the version

7
Feedback

Changes become new prototype versions

↻ Repeat

Understand the workflow

Talk to the user. Watch how they work today. Capture the problem, not a 40-page spec.

In the app: create a Feature with a Problem Statement.

In fact, your prototype becomes the specification. You don't write what you'll build separately — you show it, get sign-off, then build it.
Traditional SDLC (enterprise scale)

Each phase produces formal documents and sign-offs before the next begins. Click a step to see typical deliverables.

1
Requirements gathering

Interviews, BRD/FRD, use cases

2
Feasibility study

Cost, risk, and technical viability

3
High-level design (HLD)

Architecture, modules, stack

4
Wireframes

Low-fidelity layouts and flows

5
Mockups

High-fidelity static screens

6
Interactive prototype

Clickable validation before code

7
SDS / detailed design

LLD, schema, APIs, data model

8
Implementation

Coding, reviews, unit tests

9
Integration & system test

SIT, regression, performance

10
UAT

Business test scripts and sign-off

11
Deployment

Release plan, migration, go-live

12
Maintenance

Support, patches, enhancements

Requirements gathering

Business analysts interview stakeholders, document current-state pain points, and capture functional and non-functional requirements in structured templates.

BRDFRDUse casesUser storiesTraceability matrix

Often weeks of workshops before anyone sees a screen.

Design artifact ladder (where confusion happens)
Wireframe Structure
Mockup Look & feel
Prototype Behaviour
SDS / LLD Written spec
Production code Shipped app

In enterprise SDLC these are separate deliverables, often owned by different people, reviewed in different meetings. Drift between them is common.

Built for large teams, long timelines, and regulatory overhead. Wireframes, mockups, prototypes, and SDS each get their own review cycle. For a small team doing rapid AI-assisted delivery, that stack is too heavy — our cycle collapses design artifacts into one approved prototype link.

We deliberately do not use

User stories Epics Story points Sprint ceremonies BRD / FRD documents Formal UAT sign-off packs Jira-style workflows Change control boards Traceability matrices

What we use instead

  • Prototype URL — the living spec clients can click
  • Approval record — who said yes, when, and any remarks
  • Version history — v1, v2, v3 never overwritten
  • Change requests — scope changes logged, not sneaked in
  • Deployment ledger — which approved version shipped

Side-by-side

  • Requirements doc → Prototype link
  • Sign-off meeting pack → Approval + remarks
  • Change request form → CR in tracker
  • Release notes doc → Deployment record
  • "Was this in scope?" debate → Audit trail