Documentation Production Pipeline · Domino's NextGen Pulse

📋 Documentation Production Methodology

Every feature release at Domino's follows a structured 10-stage documentation production workflow. For the Predictive Inventory feature, this began with JIRA intake and a developer demo to gather firsthand context before any drafting began — and ended with a validated ES localization and L10N handoff for global distribution. The pipeline includes two review gates: an SME accuracy review after the first draft, and a peer editorial review before final publication. Click any pipeline stage below to expand the detail view.

⚡

10-Stage Documentation Pipeline

Select a stage to view inputs, actions, and outputs — click all stages to trace the full workflow

JIRA Intake

JIRAConfluence

→

Developer Demo

Zoom / TeamsFigma

→

Jenkins Testing

JenkinsNextGen POS

→

Figma Parsing

FigmaSnagit

→

Release Note — First Draft

ConfluenceGPE Template

→

SME Accuracy Review

Zoom / TeamsConfluence

→

Editorial Review

ConfluenceTrack Changes

→

Inline Manual Walkthrough

Inline ManualEN

→

Confluence Update

ConfluenceSnagit

→

ES Localization + L10N Handoff

Inline ManualL10N

JIRA Intake — Ticket Review & Scope Definition

PULSE-2847 · Epic INV-24 · Predictive Inventory Phase 1

📥 Inputs Received

📋JIRA ticket PULSE-2847 with acceptance criteria and linked Epic INV-24

📄Engineering spec (Confluence): data model for recommendation engine — 4 inputs defined (sales velocity, logistics lead time, on-hand count, waste log)

📐Product PRD: feature rationale, rollout scope, manager override requirements

🎨Figma design file link: 3 frames — dashboard state, discrepancy state, manager override modal

⚡ Actions Taken

🔍Identified 3 new UI states requiring documentation coverage

🔍Identified 2 new workflows: discrepancy alert trigger and manager override confirmation

🔍Flagged 4 potential user questions from acceptance criteria (override authority, draft save behavior, recommendation accuracy, code rotation)

📝Created documentation scope brief and deliverables list

📤 Outputs Produced

✅Documentation scope brief: 3 deliverables mapped (Release Note, Inline Manual walkthrough, Confluence page update)

✅8 test scenarios defined for Jenkins validation phase

JIRAConfluenceEpic INV-24

Developer Demo — Cross-Functional Collaboration & Context Gathering

Engineering · Product · Design · Live feature walkthrough · Questions and open items resolved

📥 Inputs Received

📋JIRA scope brief and open questions from Stage 01 intake

🖥️Live developer demo of the Predictive Inventory feature in a dev environment — walked through all three UI states: baseline dashboard, discrepancy alert trigger, manager override modal

👥Cross-functional attendees: engineering lead, product manager, and UX designer

⚡ Actions Taken

🔍Observed all 3 UI states live — confirmed discrepancy trigger behavior and manager override flow matched JIRA spec

❓Asked targeted questions: exact threshold logic for discrepancy trigger, override code rotation frequency, whether save-as-draft preserves discrepancy alerts, and rollout timing relative to EN/ES

📝Captured 6 context items not documented in JIRA: column positioning in table, exact UI copy for alert and modal, draft save behavior confirmed, manager code scope (not user-specific), recommendation update cadence, and EN/ES requirement confirmed for Inline Manual only (not Release Note)

🔄Aligned on 1 scope change: override code applies globally (not per-store) — updated documentation scope brief accordingly

📤 Outputs Produced

✅Updated scope brief with 6 context items and 1 scope correction

✅8 Jenkins test scenarios confirmed and finalized

✅Preliminary Behavioral Changes outline drafted from demo notes

Zoom / TeamsFigmaJIRA

📋 Insight — Undocumented Context Captured Six context items essential to documentation accuracy were surface during this demo that were absent from the JIRA ticket: exact UI copy strings, draft-save behavior, manager code scope (global, not per-store), recommendation update cadence, column positioning, and the EN/ES split requirement. Without live demo attendance, these gaps would have propagated into the published documentation.

Jenkins Environment Testing — Hands-On Validation

Cloud-native staging · 8 test scenarios · 0 post-deployment defects

📥 Inputs Received

🔧Access to cloud-native Jenkins staging environment with Predictive Inventory feature enabled

📋8 test scenarios defined in intake phase: recommendation population, discrepancy trigger conditions, modal behavior, submit blocking logic, manager code validation, draft save, EN/ES display

⚡ Actions Taken

✅Executed all 8 scenarios end-to-end in staging environment

⚠️Surfaced 2 defects pre-deployment: (1) discrepancy alert not clearing on quantity correction; (2) Submit button blocked even when all discrepancies resolved

🔄Re-tested both scenarios post-fix; confirmed resolution before proceeding

📝Documented exact UI copy strings from live environment to inform Release Note accuracy

📤 Outputs Produced

✅2 defects filed in JIRA, resolved pre-deployment

✅8/8 test scenarios validated in final build

✅Confirmed UI copy: column label, alert text, modal CTA strings

JenkinsNextGen POSJIRAQA/UAT

⚠️ QA Finding — Pre-Deployment Defects Two defects were identified and filed during this testing phase — neither reached production. Defect 1: Discrepancy alert persisted on a row after the operator corrected the quantity to match the AI recommendation; the alert should have cleared automatically. Defect 2: The Submit button remained blocked even after all discrepancy alerts were resolved, preventing order completion. Both were reproduced, filed in JIRA, fixed by engineering, and re-validated in staging before the release was cleared. The 0 post-deployment defect record reflects this pre-deployment gate — not the absence of issues found.

Figma Design Parsing — UI Copy & Screenshot Planning

3 design frames reviewed · UI copy extracted · Screenshot capture points identified

📥 Inputs Received

🎨Final approved Figma file (3 frames): dashboard with AI Rec. Qty column, discrepancy alert state, manager override modal

✅Confirmed final build matches Figma designs (validated in Jenkins step)

⚡ Actions Taken

🔍Extracted all UI-visible copy: column header "AI Rec. Qty," discrepancy callout "Qty differs from AI recommendation," modal title "Manager Override Required," CTA "Confirm & Submit"

📸Identified 3 screenshot capture points for Confluence documentation

📝Cross-referenced Figma annotations for EN/ES string requirements

📤 Outputs Produced

✅UI copy inventory: 8 strings documented with context

✅Screenshot plan: 3 captures (dashboard, alert row, modal)

✅EN/ES localization flag confirmed for Inline Manual

FigmaSnagitEN/ES

GPE Release Note — First Draft

GPE Confluence template · Overview · Behavioral Changes · UI Changes · 8 steps

📥 Inputs Received

📋GPE Release Note Confluence template (Name, Overview, Attachments, Behavioral Changes, UI Changes)

✅Validated UI copy strings from Jenkins testing

✅UI copy inventory and screenshot plan from Figma parsing

⚡ Actions Taken

✍️Authored Overview (2 sentences): feature name, location, behavioral impact summary

✍️Authored Behavioral Changes: context, workflow impact, concept and usage explanation, 4 anticipated user FAQ items, read-only column note

✍️Authored UI Changes: 8 numbered steps covering full workflow including discrepancy and manager override scenarios, 2 notes (submit button behavior, draft save)

🔗Added JIRA Epic, Solution doc, and Figma links in Attachments section

📤 Outputs Produced

✅GPE Release Note — first draft complete, shared with SME team for accuracy review

✅All 3 GPE template sections populated per standard

ConfluenceGPE TemplateJIRA INV-24

SME Accuracy Review — Team Collaboration & Iterative Editing

First-draft review with Engineering, Product, and Design · Accuracy confirmed · Final edits incorporated

📥 Inputs Received

📄First draft of GPE Release Note shared via Confluence comment thread

👥Review attendees: engineering lead, product manager — same cross-functional team as developer demo

⚡ Actions Taken

🔍Walked the team through all three sections of the Release Note in a live review session

✅Engineering confirmed: data inputs (sales velocity, on-hand count, logistics lead time, waste log) are accurate and complete

✅Product confirmed: manager override scope (global code, not user-specific) and draft save behavior are correctly described

🔄Incorporated 2 editorial clarifications from engineering: recommendation update cadence (nightly, not real-time) and minimum data threshold before recommendations generate (3 weeks of sales history)

📝Added nightly cadence note to Behavioral Changes section; added 3-week threshold to FAQ ("What if I'm a new store?")

📤 Outputs Produced

✅SME-approved Release Note with 2 accuracy updates incorporated

✅New FAQ item added: behavior for stores with fewer than 3 weeks of sales data

Zoom / TeamsConfluenceComments

📋 Insight — Accuracy Updates from SME Pass Two accuracy corrections surfaced in the SME review that were not in the JIRA spec: recommendation update cadence is nightly (not real-time), and recommendations only generate after 3 weeks of sales history. Both were incorporated into the Release Note and triggered a new FAQ item for new stores. This review gate is the mechanism behind the 0 post-deployment defect record.

Editorial Review — Peer Writer Quality Pass

Second writer review · Style, clarity, and template compliance · Final edits before publication

📥 Inputs Received

📄SME-approved Release Note draft passed to peer technical writer for editorial review

📋Domino's documentation style standards and GPE template compliance checklist

⚡ Actions Taken

✍️Peer writer reviewed for: style consistency, sentence clarity, step numbering integrity, and GPE template field compliance

📝Received tracked-change feedback: 3 phrasing adjustments for clarity, 1 note re-sequenced for better placement, 1 step split into two for cleaner task separation

🔄Reviewed all feedback, accepted 4 of 5 changes — one phrasing suggestion declined with rationale (operator-level language preserved over corporate register)

✅Final version approved by editorial reviewer; cleared for publication

📤 Outputs Produced

✅Publication-ready Release Note with editorial sign-off

✅Step count: 8 steps (1 step split from original 7-step draft)

ConfluenceTrack ChangesStyle Guide

Inline Manual Walkthrough — In-App Guidance Creation (EN)

3-step walkthrough · English · Tooltip + hotspot + modal guidance types · QA validated in staging

📥 Inputs Received

📋Final approved Release Note — used as primary content source for guidance copy

🎨Figma UI states for trigger point targeting

⚡ Actions Taken

⚙️Built 3-step EN walkthrough in Inline Manual: Step 1 (orientation hotspot on AI Rec. Qty column), Step 2 (discrepancy callout explanation), Step 3 (manager override modal guidance)

✍️Authored guidance copy applying operator-level language — not IT framing

✅QA tested walkthrough trigger conditions in staging environment

📤 Outputs Produced

✅Live 3-step EN walkthrough in Inline Manual; trigger accuracy confirmed in staging

🔜EN content sourced for ES localization in Stage 10

Inline ManualJenkins (QA)

Confluence Update — Documentation Deployment

New child page · 3 new screenshots · 2 existing pages updated · 1 page archived

📥 Inputs Received

📸3 final screenshots captured from production build: dashboard view, discrepancy alert row, manager override modal

📄Final approved Release Note and Inline Manual walkthrough (Stages 07–08)

⚡ Actions Taken

📄Created new child page under Inventory Management space: "Predictive Inventory — AI-Recommended Order Quantity"

📸Embedded 3 annotated screenshots with callout overlays for key UI elements

🔗Updated "Weekly Order Submission" parent page with cross-reference to new feature

🗄️Archived legacy "Manual Inventory Ordering — Pre-Automation" page (superseded)

📤 Outputs Produced

✅1 new Confluence page published with screenshots

✅2 existing pages updated; 1 archived

✅Documentation live before feature rollout

ConfluenceSnagitNextGen POS

ES Localization + L10N Handoff — Global Distribution

ES localization authored · L10N package prepared · Additional languages queued for localization team

📥 Inputs Received

📄Final EN Inline Manual walkthrough (Stage 08) — source content for ES localization authoring

📋L10N team handoff requirements: string export format, context notes, UI screenshot references per string

⚡ Actions Taken

🌐Authored ES localization of all 3 Inline Manual walkthrough steps directly — crafting operator-appropriate Spanish register, not a word-for-word conversion of EN copy

✅QA tested ES walkthrough in Inline Manual staging environment; confirmed display and trigger accuracy matched EN version

📦Prepared L10N handoff package: exported EN string file, attached Figma screenshots as visual context for each string, added context notes for UI-specific terms (e.g., "AI Rec. Qty" — retain abbreviation, do not adapt)

📤Submitted L10N package to localization team for additional target languages

📤 Outputs Produced

✅ES walkthrough live in Inline Manual — QA validated

✅L10N package delivered to localization team with context notes and visual references

✅Full pipeline complete — EN + ES live; additional languages in L10N queue

Inline ManualL10N PlatformJenkins (QA)Figma

📄

GPE Code Change Release Note

Completed per Domino's GPE Confluence template · Stage 05 output · PULSE-2847

PUBLISHED

📋

GPE Code Change Release Note

NextGen Pulse · Inventory Management · v1.0

GPE TEMPLATE

Name

Predictive Inventory — AI-Recommended Order Quantity Dashboard Column

PULSE-2847 · Epic INV-24 · Predictive Inventory Phase 1

✍️ Content Decision — Name Field

Title follows GPE convention: [Feature Name] — [What it adds]. "Dashboard Column" is the operative noun — it tells the reader exactly what kind of change this is before they open the document. Avoids vague names like "Predictive Inventory Update" which don't signal scope.

Overview

NextGen Pulse now includes an AI-recommended order quantity column in the Inventory Order dashboard, providing store teams with data-driven purchasing guidance based on sales velocity, current stock levels, logistics lead times, and waste history. Orders that deviate from AI recommendations require manager acknowledgment before submission.

✍️ Content Decision — Overview (2-sentence constraint)

GPE template specifies 1–2 sentences. Sentence 1 names the feature and its location — front-loading the "what" before the operator has to read further. Sentence 2 states the behavioral impact (the "so what") — specifically the constraint that will affect their workflow. Operators at this cognitive level need impact-first sequencing.

Attachments

JIRA Epic: INV-24 — Predictive Inventory Phase 1

Solution Document: PULSE-INV-24-SPEC-v3.1

Figma: Predictive Inventory UI States — Final Approved

Behavioral Changes

The Predictive Inventory feature adds an AI recommendation layer to the existing weekly ingredient ordering workflow. A new AI Rec. Qty column appears in the Inventory Order dashboard for every active order session. Recommendations are generated from four data inputs and update automatically as store data changes.

The AI Rec. Qty column is visible to all store roles. The Discrepancy Alert and Manager Override requirement are triggered only when an entered Order Qty differs from the AI Rec. Qty on any line item.

The recommendation engine analyzes: trailing 7-day sales velocity, current on-hand inventory count, inbound logistics lead time per SKU, and store-level waste log data. The output is a single recommended quantity per ingredient, displayed alongside the Order Qty input field. Store teams should use the AI Rec. Qty as the primary ordering reference, deviating only when store-specific context justifies it — for example, an upcoming catering event, a known supplier delay, or seasonal demand.

✍️ Content Decision — Behavioral Changes Sequencing

Context paragraph is placed before the constraint paragraph. This is intentional: operators need to understand what the feature does before encountering the manager override requirement — otherwise the first thing they register is a restriction. Sequence: what it does → who it affects → how the recommendations work → expected use.

Anticipated User Questions

What if the AI recommendation is wrong?

Recommendations can be overridden at any time. The manager override step confirms the deviation was intentional, not accidental — it does not prevent the order.

Who can authorize an override?

Any active manager override code is accepted. Override authority is not tied to a specific manager account.

Will overriding affect future recommendations?

Yes. Override history is one of the inputs to the recommendation engine. Consistent deviations in the same direction will be factored into future recommendation cycles.

What if I don't have a manager code during the order?

Save the order as a draft using Save for Later and retrieve the manager override code before returning to submit.

✍️ Content Decision — FAQ Ordering

"What if it's wrong?" is sequenced first because it addresses the primary operator anxiety — am I being locked into a system that can make bad recommendations? Answering this first before procedural questions reduces cognitive resistance to the feature. "Draft save" is placed last because it's an exception case (~15% of users); front-loading it would suggest the override process is harder than it is.

Note. The AI Rec. Qty column is read-only. Do not attempt to edit values in this column — all order quantity entries are made in the Order Qty column only. Recommendations reflect store-level data and may differ from regional defaults.

✍️ Content Decision — Read-Only Note Placement

This restriction note is placed at the section end, not at the beginning. Front-loading a constraint before the operator understands the workflow increases cognitive resistance and creates a negative first impression of the feature. By the time they reach this note, they understand why the column is read-only (it's AI-generated output, not a user input field).

User Interface Changes

The Inventory Order dashboard now displays a new AI Rec. Qty column between On Hand and Order Qty. Recommended quantities appear highlighted for each line item. Orders with any quantity deviations require manager acknowledgment before final submission.

1.From the NextGen Pulse home screen, navigate to Inventory → Weekly Order.

2.Locate the new AI Rec. Qty column (highlighted in blue) in the order table. Review recommended quantities for each ingredient line before entering your order.

3.Enter your quantities in the Order Qty column. If your quantity matches the AI recommendation, no additional action is required for that line.

4.If you enter a quantity that differs from the AI Rec. Qty, a Discrepancy Alert callout will appear on that row. Review the alert and confirm your entered quantity is intentional before continuing.

5.Continue entering quantities for all remaining items. Multiple Discrepancy Alerts can be active simultaneously across different ingredient lines.

6.When all quantities are entered, select Submit Order.

7.If any Discrepancy Alerts are active, a Manager Override modal will appear. Review the list of deviating line items and quantities displayed in the modal.

8.Enter the Manager Override Code in the provided field and select Confirm & Submit to complete order submission.

Note. The Submit Order button will not activate until all Order Qty fields are populated. Manager override codes rotate monthly and are distributed by your District Manager. To save an incomplete order, select Save for Later at any time.

✍️ Content Decision — Step 4 (Discrepancy Alert)

Step 4 uses "confirm your entered quantity is intentional" — not "the system has flagged an error." The phrasing positions the alert as a confirmation mechanism, not an accusation of a mistake. This matters at the operator level: store managers need to feel the system is assisting them, not auditing them. Small copy choices carry significant adoption friction implications at scale.

🎯

Inline Manual Walkthrough — In-App Guidance Simulation

Stage 08 output · 3-step contextual walkthrough · Simulated operator view · EN localization shown

INTERACTIVE

💡 About This Simulation

The simulation below represents the in-app guidance layer created in Inline Manual — the same tool used to deliver the context-sensitive modals, hotspots, and tooltips that drove the 89% support ticket reduction. Step through all 3 states to see how each guidance touchpoint is positioned relative to the operator's task.

Step 1 of 3 — AI Recommendation Column

NEXTGEN PULSE Store #4821 — Weekly Inventory Order

Weekly Inventory Order AI Recommendations Active

Ingredient	On Hand	AI Rec. Qty ✦	Unit
All-Purpose Flour	4	12	bags
Tomato Sauce	8	18	cases
Mozzarella	6	24	lbs
Pepperoni	10	20	lbs
Pizza Dough Balls	15	40	units