How I Use Rovo Dev to Make Decisions for Me

[#1]

From information chaos to structured decision-making: a developer's journey through the AI-enhanced SDLC

The Challenge: Information Overload in Service Integration

Have you ever felt like you spend more time searching for relevant information than actually building a solution? You aren’t alone: according to the Atlassian State of Developer Experience Survey 2025, finding information is the #1 developer productivity killer. And boy, I used to feel that pain.

Over the past year, I made it a personal mission to extract the most of the tools at my disposal for a maximum productivity boost. The difference has been significant: less time lost in information chaos, more time spent on the tasks that really matter.

In this article, I walk you through a specific example where my team was tasked with integrating onto a new software platform, sharing the practical steps I took with my available tools, so you can apply these learnings to your own workflow. If you’re a developer, a team lead, or somebody looking for strategies to cut through the noise, you’ll find some insights to regain your focus and speed up your results.

The Old Way: A Decision Marathon

In the pre-AI tools era, this integration decision would have been a classic case of:

• 🕵️ Day 1: Hunting down service documentation across multiple platforms

• 📧 Day 2: Email chains with different teams trying to understand ownership and integration patterns

• 📝 Day 3: Drafting decision documents from scratch

• 🔄 Day 4: Multiple review cycles and revisions

The Rovo AI-Enhanced Way: From Chaos to Clarity in Hours

Step 1: Service Discovery with Compass

My journey started with Compass - our software catalog that's like having a map for your tech stack. A service query revealed:

• ✅ The application I needed to integrate with

• ✅ Source repository location

• ✅ Relevant documentation links

• ✅ The owning team's Slack channel

Step 2: AI-Powered Research with Rovo Search

Instead of manually digging through internal wikis and codebases, I turned to Rovo Search. With my company's unique context already baked in, I received:

• 🔍 Specific integration pattern descriptions

• 📋 Payload examples tailored to our architecture

• 🎯 Code snippets relevant to our tech stack

Step 3: Collaborative Brainstorming in Slack

Armed with solid background knowledge, I initiated a focused Slack conversation. While the team quickly identified viable options, here's where the old process would have hit a wall: Slack discussions are great for brainstorming but terrible for decision documentation.

Step 4: AI-Enhanced Decision Documentation

This is where the magic happened. Rovo transformed our Slack ramblings into a structured decision document by:

• 📋 Using our standardized Confluence decision templates

• 🎯 Extracting key options and pros/cons from the conversation

• 📝 Creating a professional first draft in minutes

Step 5: Code-First RFCs with Rovo Dev

Once we identified the high-level solution, Rovo Dev took our decision document and:

• 🔗 Identified repository references

• 📋 Explained the code structure in natural language

• 🔍 Found the appropriate integration points

• 📚 Generated pseudo-code illustrating key architectural changes

• 💻 Implemented working code fragments that illustrate the integration

Step 6: Speed Up Implementation with Rovo Dev CLI

With the main tasks already identified, the Rovo Dev CLI was a fantastic companion to flesh out the first implementation details.

• A natural language prompt in the context of the prior spec would get me a working prototype

• Forget traditional searches: a question in plain English would scan the whole codebase for me and find what area carried out a specific functionality

• As a developer, this fit seamlessly in my terminal (and my IDE’s terminal) in my regular workflow

The Results: What We Achieved

Before AI-Powered Tools: days of manual search and drafting

With AI-Powered Tools: a few hours of focused collaboration

What We Delivered:

• 📋 A comprehensive decision document

• 🔗 Links to all relevant documentation

• 💻 Actual code references and examples

• 👥 Clear ownership and next steps

• 🎯 Structured decision framework for future reference

• 💻 A quick first prototype

The Bigger Picture: Why This Matters

This process is transforming how we make engineering decisions. When developers can focus on the actual problem-solving instead of information hunting, we unlock:

• 🚀 Faster delivery cycles

• 🧠 Better decision quality

• 👥 Improved team collaboration

• 📚 Knowledge preservation

• 🔄 Faster onboarding for new team members

The Atlassian Advantage

What makes this possible is Atlassian's integrated approach to developer experience:

• Compass is my one stop shop and single source of truth for the software catalog of the whole company, allowing me to find owners, specifications and documentation in no time

• Rovo Search has replaced any other search tool with context-aware information from my entire organization, answering questions that no external search tool could satisfy

• Rovo Dev generates code that actually works with my codebase, considering the interfaces and standards that apply to my organization

• Confluence contains the standardized templates that our company has defined for quick collaboration, and exposes them to our AI tools, so making quick, standardized decisions has never been this fast

• Slack starts our real-time collaboration quickly capturing our first solutions and ideas

• Rovo agents with access to the information above build comprehensive decision-making documents, getting a professional document draft ready in seconds

• Rovo Dev CLI is my assistant during practical implementation, finding key code modules faster than a traditional CTRL + F scan, explaining new code areas in natural language, and creating working solutions as requested

Looking Forward: The Future of AI-Enhanced Development

This experience isn't unique to me. Across Atlassian, teams are discovering that AI isn't replacing developers - it's amplifying their capabilities. We're moving from spending most of our time on information gathering to spending it on actual problem-solving.

[Josh]

Hi @Enrique Serrano Valle . The content of your post is quite interesting.

I think the article title may be unintentionally misleading. You're using Rovo to *document* decisions and provide context around them, not make decisions on your behalf. What you describe in the article seems very much aligned with human in the loop, rather than Rovo deciding things autonomously.