Owner: Project Manager (PM)
Contributors: Head of Delivery (HoD) — technical oversight + escalation ·
AI Engineer / Developer — build and deployment ·
Account Lead — client communication + health monitoring ·
BA — scope validation (pre-build)
Department: D4 — Delivery
Entity: PrivateAI
Status: v0.9 · Reviewing
Version: 0.9 | Last updated: 2026-07-18
What this workflow does: Governs the full lifecycle of a fixed-scope AI Project
build — from technical scoping and environment setup through iterative development
sprints, UAT, client acceptance, and production deployment. Covers both standard
AI Project builds and AIc Compliance deployments where the build phase applies.
What triggers it: CS-01c (AI Project Client Onboarding) complete and CS-02
kickoff scheduled — confirmed scope document (from BA scoping call) and technical
summary in hand. Payment confirmed (Gate G2).
What it produces: A deployed, client-accepted AI solution in production —
with a completed Scope and Acceptance Document, UAT sign-off, production deployment
confirmation, and DEL-07 knowledge capture entry.
Why it matters: AI Project delivery is the primary revenue realisation mechanism
for fixed-scope work. Scope discipline, sprint cadence, and a clean UAT and acceptance
process protect margin and client satisfaction simultaneously. A project that delivers
to scope, on time, and with a clean handover is the foundation for a referral or
renewal.
| Role |
Department |
Responsibility |
CEO? |
| PM |
D4 |
Owns the delivery timeline · manages Jira board · coordinates sprints · runs client status updates · manages UAT process |
No |
| HoD |
D4 |
Technical oversight and design review · signs off architecture before build starts · escalation authority |
No |
| AI Engineer / Developer |
D4 |
Builds all solution components · unit tests · deploys to staging and production |
No |
| Account Lead |
D4 |
Client-facing communication · health monitoring · escalation flags |
No |
| BA |
D4/D6 |
Validates scope pre-build · reviews acceptance criteria · supports UAT |
No |
Phase 1 — Technical Scoping & Setup
↓ HoD architecture sign-off
Phase 2 — Sprint Development (iterative)
↓ Sprint reviews with client
Phase 3 — UAT & Client Acceptance
↓ Signed Scope and Acceptance Document
Phase 4 — Production Deployment & Handover
↓ DEL-07 knowledge capture
Phase 5 — Post-Delivery (→ CS-04 triggered)
Step 1 — Review scope and technical summary
- Who: PM + HoD + AI Engineer
- Action: Review the signed scope document and technical summary from the BA
scoping call. Confirm all acceptance criteria are clear and measurable. Flag any
ambiguities to the BA before build starts — never start building against an
unclear acceptance criterion.
- Tool: Google Workspace (scope doc) · Jira
- Output: Confirmed understanding of scope, acceptance criteria, and technical
approach
- Decision: Scope clear → proceed to Step 2 · Ambiguity found → BA resolves
within 24h before proceeding
Step 2 — Technical architecture design
- Who: HoD + AI Engineer
- Action: Design the technical architecture for the solution — components,
integrations, data flows, infrastructure requirements, and security considerations.
HoD reviews and signs off on the architecture before any build work begins.
- Tool: Google Workspace (architecture doc) · draw.io or equivalent
- Output: Approved technical architecture document
- Decision: Architecture approved by HoD → proceed to Step 3 ·
Revision needed → revise and resubmit
Step 3 — Environment setup and access provisioning
- Who: AI Engineer + PM (coordinates client access)
- Action: Set up development and staging environments. Request all required
client access, API keys, credentials, and data samples from the client via PM.
Log all access provisioned in Jira.
- Tool: AWS · n8n · client's platforms · Jira
- Output: Development and staging environments ready · all client access
confirmed and logged
- Decision: Access provided within 5 business days → proceed ·
Access delayed → PM escalates to Account Lead → client SLA clock still runs
Step 4 — Build Jira project board
- Who: PM
- Action: Build the full sprint board in Jira — epics mapped to scope sections,
user stories per acceptance criterion, sprint plan with dates, and definition of
done per story. Share board access with HoD and Account Lead.
- Tool: Jira
- Output: Live Jira board with full sprint plan
- SLA: Board live before first sprint starts
Step 5 — Sprint planning
- Who: PM + AI Engineer (+ HoD for complex sprints)
- Action: At the start of each sprint (typically 1–2 weeks), confirm the sprint
goal, assign stories, confirm dependencies and blockers, and agree the demo scope.
- Tool: Jira · Google Meet
- Output: Sprint goal confirmed · stories assigned · Jira board updated
Step 6 — Build sprint deliverables
- Who: AI Engineer
- Action: Build all solution components assigned to the sprint. Unit test each
component before marking complete in Jira. Flag any scope ambiguity or technical
blocker to PM same day.
- Tool: n8n · AWS · Claude · client platforms · Jira
- Output: Sprint stories marked complete with unit tests passed
- Decision: Blocker identified → PM escalates to HoD if technical ·
escalates to Account Lead if client-side
Step 7 — Sprint review and client demo
- Who: PM + AI Engineer + Account Lead + Client
- Action: At the end of each sprint, demo completed work to the client. Walk
through each deliverable against its acceptance criterion. Capture client
feedback. Any approved changes to scope → log as a change request in Jira
(separate from original scope). Any rejected items → return to the sprint backlog.
- Tool: Google Meet · Jira · staging environment
- Output: Sprint review notes · change requests logged · client feedback
captured
- Decision: Sprint accepted → next sprint planning ·
Scope change requested → PM assesses impact (time + cost) → Account Lead
presents to client → approved: add to backlog · rejected: decline politely ·
Fundamental scope change → HoD + Account Lead + client reset meeting
Step 8 — Build Tracker update (DEL-03-DOC-02)
- Who: PM
- Action: After each sprint review, update the Build Tracker with current
sprint completion percentage, hours consumed vs budget, change requests logged,
and overall project health. This feeds the client progress report in CS-03.
- Tool: Google Workspace (DEL-03-DOC-02)
- Output: Updated Build Tracker · data available for CS-03 progress report
Step 9 — UAT preparation
- Who: PM + BA + AI Engineer
- Action: Prepare the UAT environment and test cases aligned to the acceptance
criteria in the scope document. Share UAT test scripts with the client at least
5 business days before UAT begins. Brief the client on what UAT covers and how
to log issues.
- Tool: Staging environment · Google Workspace (UAT test scripts)
- Output: UAT-ready environment · client briefed · test scripts shared
Step 10 — UAT execution
- Who: Client (primary tester) + AI Engineer (issue resolution) +
PM (coordination)
- Action: Client runs UAT against the agreed test cases. Any issues logged by
the client are triaged by PM within 24 hours — in-scope bug (fix immediately) vs
out-of-scope change request (price and timeline impact presented). PM tracks all
UAT issues in Jira.
- Tool: Jira (UAT issue tracking) · staging environment · Google Meet
- Output: UAT issues triaged and resolved · or escalated as change requests
- Decision: Issue type assessment
- In-scope bug → AI Engineer fixes · re-test · close
- Out-of-scope change request → PM prices and presents → client accepts/declines
- Disputed scope item → BA reviews original scope doc → HoD final call
Step 11 — Client acceptance sign-off
- Who: PM + Account Lead + Client
- Action: Once all UAT issues are resolved or formally accepted, present the
Scope and Acceptance Document (SAL-01e-style) to the client for signature.
Document confirms: all deliverables received and accepted, any open items and
agreed resolution path, and formal project acceptance.
- Tool: Google Workspace (Scope and Acceptance Document) · Documeso or email
- Output: Signed Scope and Acceptance Document
- Decision: Client signs → proceed to Phase 4 ·
Client withholds sign-off → PM + Account Lead identify outstanding items →
resolve → resubmit · If unresolvable → HoD + Account Lead + CEO escalation
¶ Phase 4 — Production Deployment & Handover
Step 12 — Production deployment
- Who: AI Engineer + HoD (approves production go-live)
- Action: Deploy the accepted solution to the production environment. HoD
approves go-live before deployment. Run smoke tests post-deployment to confirm
production stability. Document all deployment steps and configuration in the
handover notes.
- Tool: AWS · n8n · client's production environment · Jira
- Output: Solution live in production · smoke tests passed · deployment
documented
- Decision: Smoke tests pass → go-live confirmed · Smoke tests fail →
rollback to staging · investigate · re-deploy
Step 13 — Client handover
- Who: PM + Account Lead + Client
- Action: Formal handover session with the client. Walk through: what was
built, how to use it, admin access and credentials, monitoring approach, and
support process going forward. Share all handover documentation.
- Tool: Google Meet · Google Workspace (handover notes + user guide)
- Output: DEL-03-DOC-04 (AI Project Handover Checklist) fully ticked ·
client confirmed receipt of all documentation
Step 14 — Trigger CS-04 and CS-06
- Who: PM (notifies CS team)
- Action: Notify Zimal (D3) that all deliverables are confirmed complete and
the handover checklist is fully ticked. This is the formal trigger for CS-04
(Post-Delivery Process — NPS survey) and subsequently CS-06 (Offboarding).
- Tool: Discord (internal notification) · Jira (status update)
- Output: CS-04 triggered
Step 15 — Knowledge base contribution
- Who: AI Engineer + PM + HoD
- Action: Within 5 business days of go-live, add a knowledge entry to
DEL-07-DOC-01 covering: solution architecture, key technical decisions,
reusable components or prompts, lessons learned, and any client-specific
context worth preserving.
- Tool: Wiki.js (DEL-07-DOC-01)
- Output: Knowledge entry in Delivery Knowledge Base · feeds DEL-07
| Tool |
How it's used |
Agent-ready |
| Jira |
Sprint board, UAT tracking, build tracker, change request log, hours tracking |
Yes — confirmed |
| Google Workspace |
Architecture doc, build tracker (DOC-02), UAT scripts, handover notes, acceptance doc |
Yes — API + MCP |
| n8n |
Core automation platform for most AI Project builds |
Yes — native |
| AWS |
Infrastructure deployment — staging and production |
Yes — confirmed |
| Claude |
AI capabilities embedded in solution builds |
Yes |
| Google Meet |
Sprint reviews, UAT coordination, handover session |
Yes |
| Documeso |
Scope and Acceptance Document signature (or email alternative) |
Yes — API |
| Wiki.js |
Knowledge base contribution at project close |
Yes — native |
| CS-05-DOC-01 |
Client Health Scorecard — updated throughout delivery |
Yes — API + MCP |
Scope ambiguity at scoping review (Step 1)
| Condition |
Action |
Owner |
| Scope clear and agreed |
Proceed to Step 2 |
PM |
| Ambiguity found |
BA resolves within 24h |
BA → PM |
| Fundamental gap in scope |
Stop build · reset scoping with client |
HoD + Account Lead |
Sprint review outcome (Step 7)
| Outcome |
Action |
Owner |
| Sprint accepted |
Next sprint planning |
PM |
| Minor change requested |
Log change request · price and present |
PM → Account Lead |
| Fundamental scope change |
Reset meeting |
HoD + Account Lead + Client |
UAT issue triage (Step 10)
| Issue type |
Action |
Owner |
| In-scope bug |
Fix immediately · re-test |
AI Engineer |
| Out-of-scope change |
Price and present impact · client accepts/declines |
PM → Account Lead |
| Disputed scope item |
BA reviews scope doc · HoD final call |
BA → HoD |
Production go-live (Step 12)
| Condition |
Action |
Owner |
| Smoke tests pass |
Go-live confirmed |
HoD approves |
| Smoke tests fail |
Rollback · investigate · re-deploy |
AI Engineer + HoD |
| Scenario |
How to handle |
Escalation |
| Client access delayed beyond 5 business days |
PM escalates via Account Lead. Engagement hours still run from start date. Flag in health scorecard. |
Account Lead → HoD |
| Scope creep — client requests additions during build |
Log as formal change request. Never add to scope without PM pricing and client approval. |
PM → Account Lead |
| Client refuses UAT within agreed window |
PM flags as Amber health signal. Account Lead engages client directly. |
Account Lead → HoD if no response after 5 days |
| Client withholds acceptance sign-off |
PM and Account Lead identify outstanding items. Resolve or formally scope as punch list. |
HoD + Account Lead + CEO if unresolvable |
| Production deployment fails |
Rollback immediately. HoD leads investigation. Client informed by Account Lead within 1 hour. |
HoD → CEO for P1 classification |
| Project runs over budget or timeline |
PM flags to HoD and Account Lead immediately. Client communication plan agreed before external comms. |
HoD + Account Lead |
| Metric |
Type |
Target |
| Milestones delivered on schedule (%) |
Leading |
≥ 90% |
| Scope changes after sign-off (count) |
Lagging |
0 — all changes logged as formal requests |
| Client acceptance achieved on first UAT submission (%) |
Leading |
Baseline needed |
| Production deployment — smoke tests passed first attempt (%) |
Leading |
Baseline needed |
| Knowledge base entry filed within 5 days of go-live (%) |
Process |
100% |
| DEL-03-DOC-04 handover checklist complete before CS-04 trigger (%) |
Process |
100% |
| Step |
Current state |
Suggested automation |
Tool |
Effort |
| Step 8 — Build Tracker update |
Manual |
n8n weekly pull from Jira → auto-update Build Tracker (DEL-03-DOC-02) |
n8n + Jira |
Medium |
| Step 6 — Blocker detection |
Manual |
n8n monitors Jira for stories flagged as blocked >24h → alert PM and HoD |
n8n + Jira |
Low |
| Step 10 — UAT issue triage |
Manual |
n8n routes client-logged UAT issues to PM triage queue automatically |
n8n + Jira |
Medium |
| Step 14 — CS-04 trigger |
Manual (PM notifies Zimal) |
n8n: handover checklist complete in Jira → auto-notify Zimal via Discord |
n8n + Jira + Discord |
Low |
| Step 15 — Knowledge capture reminder |
Manual |
n8n: project closed in Jira → reminder to PM/AI Engineer to file DEL-07 entry within 5 days |
n8n + Wiki.js |
Low |
Upstream:
Downstream:
| Version |
Date |
Description |
Author |
| 0.9 |
2026-07-18 |
Initial draft — 5-phase delivery lifecycle (scoping, sprint development, UAT, production deployment, knowledge capture), Jira-centric sprint management, UAT triage decision framework, scope change governance. |
Hamza Tasneem (HoD) |
← Back to Delivery Hub · ← Back to Company Operating System