Why this series exists
I’ve been in software development since 2005. In that time I’ve written more BRDs, FRDs, Confluence pages, and JIRA tickets than I care to count. I’ve led teams that spent entire sprint planning sessions arguing about requirements that should have been written down weeks earlier. I’ve watched projects lose momentum not because the engineering was hard, but because the documentation overhead was eating everyone alive.
Nobody was writing those documents because they wanted to. They were writing them because skipping them was worse.
That was the state of things until I started seriously using Claude (Anthropic’s AI assistant) in my actual workflow. Not in a demo. Not to generate marketing copy. In the real daily work of a Technical Project Manager running projects across time zones.
What I found was not what I expected. Claude didn’t just speed things up. It changed the texture of the work. I could hand it a rough project brief and get back a structured BRD that needed editing, not starting over. I could paste in a functional document and have it extract JIRA tickets that were actually accurate. I could ask it to review a code snippet and it would catch the things I would have caught on a second pass — but I rarely had time for a second pass.
The documentation overhead that used to be a constant drag on every project largely disappeared.
That’s what this series is.
The running example: ClientHub
Every post in this series uses the same fictional project so you can see how each step builds on the last.
ClientHub is a client portal web app. Agencies and freelancers use it to manage projects; clients log in to approve deliverables and raise change requests; developers track assigned tasks. It has authentication, role-based access, a project dashboard, a deliverable approval workflow, comments, invoicing, and email notifications.
It’s a realistic mid-size project — complex enough to be useful as an example, simple enough that you can follow every decision without needing domain expertise.
Starting from nothing, this series takes ClientHub from an initial project idea all the way to a reviewed, AI-assisted codebase. Post 1 is the BRD. Post 10 is the code review. Everything in between is the real work.
Who this is for
This series is written for people who already know how software gets built and want to know where AI actually fits into that process — without the hype.
If you’re a developer trying to move faster on the parts of the job that slow you down. A tech lead trying to reduce the documentation and planning overhead on your team. A PM trying to write better requirements in less time and fewer revision cycles. A senior engineer trying to do thorough code reviews without sacrificing your afternoon.
Every prompt in this series is one I’ve actually used. Every output is representative of what Claude actually produces when you give it proper context.
The complete series
| # | Post | What it covers |
|---|---|---|
| 1 | How I Created a Full BRD in 20 Minutes Using Claude Code | The prompting workflow for generating a complete Business Requirements Document from a project brief. Covers structure, iteration, gap-checking, and the exact prompts used for ClientHub. |
| 2 | Writing a Functional Requirements Document with Claude Code | How to turn a BRD into a detailed FRD — with feature specs, user flows, business rules, and acceptance criteria. Claude generates the first draft; you refine the business logic. |
| 3 | Aligning Your BRD and Functional Document Using Claude | Using Claude to build a traceability matrix, find gaps between the BRD and FRD, and surface contradictions before they become development problems. |
| 4 | Getting Client Sign-Off: Finalize Design + Docs with Claude’s Help | How to use Claude to convert technical documentation into a client-readable sign-off package — and how to handle change requests after approval. |
| 5 | Auto-Generate Confluence Pages from Your BRD and Functional Docs | Generating properly structured Confluence wiki pages directly from the BRD and FRD, including how to push them live using the Atlassian MCP. |
| 6 | Create JIRA Epics, Stories, and Tasks with Claude — Zero Manual Work | Extracting a full JIRA hierarchy (Epics → Stories → Tasks) from the FRD, with acceptance criteria, story points, and dependencies. Includes the Atlassian MCP push workflow. |
| 7 | Write Complete Test Cases from Your Functional Docs Using Claude | Using Claude to generate a structured test case suite from FRD acceptance criteria — covering happy paths, negative paths, edge cases, and role-based access. |
| 8 | The CLAUDE.md File: Your Secret Weapon for AI-Assisted Coding | What goes in a CLAUDE.md file, where it lives, and why it’s the single highest-leverage thing you can do before starting an AI coding session. Includes the full ClientHub example. |
| 9 | Test Case Review with Claude: Find the Gaps Before QA Does | Using Claude to cross-check an existing test suite against the FRD, find the gaps that human reviewers consistently miss, and generate the missing tests. |
| 10 | AI-Powered Code Review with Claude: What to Check and How to Prompt | How to write code review prompts that actually work, what Claude reliably catches (and what it misses), and how AI review fits into a team’s existing PR process. |
Where to start
Start with Post 1 if you want to follow the full workflow from the beginning. Each post builds directly on the previous one, so reading in order gives you the most complete picture.
Jump to any post if you have a specific problem. Each post is written to stand alone — you don’t need to have read the previous ones to get value from it.
Start with Post 8 (CLAUDE.md) if you’re already using Claude Code and want the single change that will have the most immediate impact on your sessions.