Updated November 2026

AI-DLC and Spec-Driven Development for Product Owners and Business Analysts

Class Duration

7 hours of live training delivered in 1 day.

Student Prerequisites

Experience as a product owner, business analyst, product manager, or project manager on a software delivery team
Familiarity with Agile delivery, backlogs, and user stories
Basic experience with an AI chat assistant is helpful
No coding experience required

Target Audience

Product owners, business analysts, product managers, and project managers whose teams are adopting AI coding agents, AI-DLC, or spec-driven development. The course complements two beta certifications for these roles: AWS Certified AI Business Strategist (AIB-C01) and GitHub Certified: Agentic AI Business Professional (GH-700). Teams can continue with AI-Driven Development Life Cycle (AI-DLC) or Spec-Driven Development with AI Coding Agents.

Description

When AI agents draft requirements, designs, and code, the people who own intent and approvals decide what gets built. This one-day course prepares product owners and business analysts for that role, with no coding required. It covers where product roles sit in AI-DLC, from Ideation through Mob Elaboration and approval gates. Students practice writing EARS requirements and acceptance criteria, and reviewing agent-generated specs and plans. The course also shows Kiro specs and GitHub Spec Kit from a product seat, plus traceability and spec drift.

Learning Outcomes

Explain how AI-DLC changes product owner and analyst work.
Describe AI-DLC phases, intents, units of work, and bolts.
Decide at approval gates whether to approve or request changes.
Take part in Mob Elaboration with the delivery team.
Write EARS requirements and acceptance criteria.
Turn user stories into traceable spec requirements.
Review agent-generated requirements, designs, and task plans.
Work in Kiro specs and Spec Kit projects without writing code.
Trace requirements to tasks, code, and tests.
Recognize and resolve spec drift.
Refine a backlog with AI.
Define metrics that show whether AI-driven delivery works.

Training Materials

Comprehensive courseware is distributed online at the start of class. All students receive a downloadable MP4 recording of the training.

Software Requirements

A web browser and read access to the team's repository host, such as GitHub, for viewing specs and pull requests. The instructor demonstrates Kiro and a GitHub Spec Kit project, and participants with access can follow along. No coding or local development setup is required.

Training Topics

AI-Driven Delivery for Product Roles

When AI drafts the work
AI-assisted, AI-autonomous, and AI-driven
Specs as the statement of intent
Spec-first, spec-anchored, and spec-as-source
What stays the product role's job

Where Product Owners and Analysts Sit in AI-DLC

The AI-DLC phases
Intents vs. epics
Units of work and bolts
Ideation and intent capture
Feasibility and scope definition
Inception and requirements analysis
User stories and delivery planning
Answering stage questions
Approve or Request Changes
Answer modes

Mob Elaboration

Preparing an intent
Starting from a PRD or vision
Validating AI questions as a team
Resolving open questions in the room
The unit dependency graph
Reviewing the bolt plan
Handing off to Construction

Writing Requirements Agents Can Execute

Separating what from how
Testable functional requirements
Measurable success criteria
EARS requirement patterns
Event-driven WHEN requirements
State-driven WHILE requirements
Unwanted behavior with IF and THEN
Acceptance criteria as WHEN/THEN scenarios
User stories with traceable IDs
Constraints and open questions
Vague and oversized specs

Reviewing Agent-Generated Artifacts

Reading AI-drafted requirements critically
Reviewing user stories and domain models
Reviewing designs for behavior and rules
Checking task plans for coverage
Spotting scope creep
Clarification questions and checklists
Weighing reviewer-agent findings
Requesting precise changes

Kiro Specs and GitHub Spec Kit Without Writing Code

Kiro requirements.md, design.md, and tasks.md
Requirements-first and design-first workflows
When a Quick Spec is enough
Bugfix specs
The Spec Kit constitution
/speckit.specify and /speckit.clarify
/speckit.checklist for requirements quality
Answering [NEEDS CLARIFICATION] markers
Reviewing spec changes in pull requests

Tracing requirements to tasks and tests
How specs and code drift
Catching drift as a product owner
Living specs and change proposals
Splitting and sizing stories with AI
Finding gaps and duplicates
Intents on a Scrum or SAFe backlog

Metrics and Governance

Gate wait time and rework
Request Changes rate
Open questions and convergence passes
The audit trail of approvals
Workflow profiles and depth levels
Regulated changes
Confidential data in prompts
Human accountability for requirements