Project Process
We are an engineering agency specializing in AI, delivering design and discovery-driven product development.
This document covers the project-level engagement process. For how individual features flow through our pipeline, see Feature Lifecycle.
Project Phases Overview
Every project flows through three major phases:
Discovery → Design → Engineering
| Phase | Purpose | Deliverables |
|---|---|---|
| Discovery | Find business value, scope features, budget | Proposal, feature breakdown, timeline, estimates |
| Design | Translate requirements into buildable specs | Wireframes, UI designs, prototypes, design system |
| Engineering | Build, test, ship | Working software, documentation, deployed product |
Phase 1: Discovery
Goal: Find the business value and create the artifacts needed to recommend the best possible approach given the client's timeline, deadlines, and budget.
Discovery is where projects succeed or fail. Skip it, and you build the wrong thing. Rush it, and you miss critical requirements. Do it well, and everything downstream flows smoothly.
What Discovery Answers
- What problem are we solving? — The core business value
- Who are we solving it for? — Users, stakeholders, customers
- What does success look like? — Measurable outcomes
- What constraints exist? — Budget, timeline, technical, regulatory
- What's the scope? — Features in, features out
Discovery Activities
| Activity | Owner | Output |
|---|---|---|
| Stakeholder interviews | PM | Problem statement, success criteria |
| User research | PM / Design | User personas, pain points |
| Technical assessment | Engineering | Feasibility notes, risk areas |
| Competitive analysis | PM | Market positioning, differentiators |
| Feature mapping | PM | Feature list with priorities |
| Budget alignment | PM | Scope recommendations |
Discovery Deliverables
1. Problem Statement Clear articulation of what we're solving and why it matters.
2. Feature Breakdown Every feature identified, prioritized, and sized. This becomes the input to our Feature Lifecycle.
3. Proposal & Recommendations Based on the client's:
- Budget — What can we build with the resources available?
- Timeline — What's realistic given deadlines?
- Value — What delivers the most impact first?
We recommend an approach:
| Constraint | Our Response |
|---|---|
| Fixed budget, flexible scope | Prioritize highest-value features, cut low-value ones |
| Fixed deadline, flexible scope | Time-box phases, ship MVP first |
| Fixed scope, flexible timeline | Plan for full delivery, phase releases |
| Fixed everything | Identify risks, recommend tradeoffs |
4. Timeline & Milestones Clear phases with dates:
- Discovery complete: [date]
- Design complete: [date]
- Engineering sprints: [dates]
- Release milestones: [dates]
5. Budget Breakdown Transparent costing:
- Discovery & Design hours
- Engineering hours by feature
- QA and deployment
- Contingency buffer
Feature Breakdown Process
This is the critical bridge between Discovery and our Feature Lifecycle. Every feature we identify must:
1. Be Independently Deliverable
A feature should be shippable on its own. If it can't stand alone, it's not a feature—it's a task within a feature.
2. Have Clear Business Value
Every feature answers: "What does the client/user get when this ships?"
3. Be Estimable
If we can't estimate it, we don't understand it well enough. Go deeper.
4. Map to the Feature Lifecycle
Each feature will flow through:
Backlog → Todo → Discovery → Design → Engineering → QA → Done
See Feature Lifecycle for stage details.
Feature Template
For each feature, document:
## Feature: [Name]
**Business Value:** What problem does this solve?
**Users:** Who benefits?
**Priority:** Must-have / Should-have / Nice-to-have
**Size:** S / M / L / XL
**Dependencies:** What must exist first?
**Risks:** What could go wrong?
**Acceptance Criteria:** How do we know it's done?
Prioritization Framework
We use a value-vs-effort matrix:
| Low Effort | High Effort | |
|---|---|---|
| High Value | Do First | Plan Carefully |
| Low Value | Quick Wins | Don't Do |
Priority order:
- High value, low effort (quick wins with impact)
- High value, high effort (core features, plan well)
- Low value, low effort (nice-to-haves if time allows)
- Low value, high effort (cut these)
Phase 2: Design
Goal: Translate discovered requirements into buildable, approved designs.
Design doesn't start until Discovery deliverables are complete. No exceptions.
Design Activities
| Stage | Work | Exit Criteria |
|---|---|---|
| Information Architecture | Site maps, user flows | Flows approved by client |
| Wireframes | Low-fidelity layouts | Structure approved |
| UI Design | High-fidelity mockups | Visual direction approved |
| Prototypes | Interactive demos | Interactions approved |
| Design System | Components, patterns | Documented and organized |
Design Deliverables
- Approved designs for all features in scope
- Design system / component library
- Interaction specifications
- Asset exports ready for engineering
Design ↔ Feature Lifecycle
During Design phase, individual features are flowing through:
- Ready for Design
- Design In Progress
- Design In Review
- Ready for Engineering
A feature isn't "designed" until it reaches Ready for Engineering with designs attached and approved.
Phase 3: Engineering
Goal: Build what was designed, test it, ship it.
Engineering doesn't start on a feature until it's in Ready for Engineering state with:
- Designs attached
- Acceptance criteria defined
- Edge cases documented
- Dependencies identified
Engineering Activities
| Stage | Work | Exit Criteria |
|---|---|---|
| Implementation | Build per designs | Code complete |
| Code Review | PR review, feedback | PR approved |
| Testing | Unit, integration, E2E | Tests passing |
| Staging | Deploy to test environment | Deployed and verified |
| QA | Functional testing | No critical bugs |
| Production | Deploy to live | Released |
Engineering ↔ Feature Lifecycle
During Engineering phase, individual features are flowing through:
- Ready for Engineering
- Engineering In Progress
- Engineering In Review
- Ready for QA
- Done
Engagement Models
Based on Discovery findings, we recommend one of these engagement structures:
Plan A: Standard
For mid-size products (1 main feature, 2 sub-features)
| Phase | Duration |
|---|---|
| Design & Discovery | 2 months |
| Engineering | 2 months |
| Total | 4 months |
Plan B: Accelerated
For focused MVPs or tight deadlines
| Phase | Duration |
|---|---|
| Design & Discovery | 1 month |
| Engineering | 1 month |
| Total | 2 months |
Plan C: Large-Scale
For complex products (1+ main features, 2+ sub-features)
| Phase | Duration |
|---|---|
| Design & Discovery | 2 months |
| Engineering | 4 months |
| Total | 6 months |
Client Collaboration
During Discovery
- Kick-off meeting
- Stakeholder interviews
- Weekly check-ins
- Discovery review & sign-off
During Design
- Design reviews (per feature)
- Feedback cycles
- Design approval gates
During Engineering
- Sprint demos (bi-weekly)
- Staging access for testing
- Release sign-offs
Project Success Criteria
A project is successful when:
- Business value delivered — The problem is solved
- On budget — Costs aligned with estimates
- On timeline — Deadlines met
- Quality shipped — No critical bugs, good UX
- Client satisfied — Would work with us again
Quick Reference
Discovery Checklist
- Stakeholder interviews complete
- Problem statement documented
- Features identified and prioritized
- Budget aligned with scope
- Timeline realistic for deadline
- Proposal delivered and approved
- Features broken down for lifecycle
Phase Gates
| From | To | Requirements |
|---|---|---|
| Discovery | Design | Proposal approved, features documented |
| Design | Engineering | Designs approved, acceptance criteria defined |
| Engineering | Release | QA passed, client sign-off received |
Related Documents
- Feature Lifecycle — How individual features flow through our pipeline