Dependability Framework

Project Overview

Client: Fidelity Investments, Unified Developer Experience organization
Industry: SaaS, Developer Tools
Timeline: 9 months (Oct 2023-July 2024)
My Role: Lead Product Designer

One month after joining the Fidelity Unified Developer Experience team, I was invited by the Dependability Framework working group to lead its strategic and visual design. This initiative, started by technology excellence experts, needed a transition from a conceptual prototype to a fully-funded, official product that solved core developer pain points on building quality software.

From our chats and our research a few things stood out to us

Over four weeks, we interviewed 5 users (tech leads and senior devs), 3 product stakeholders and audited internal tools. Research focused specially on pinpointing where current solutions fail and aligning on team goals. We prioritized three core problems to solve:

• 70+ duplicate standards existed...

  creating confusion as teams couldn't find a single source of truth.

• Copy-paste was the only option...

  due to poor templates, forcing teams to waste time reformatting duplicated pages.

• Squads maintained their own versions...

  leading to conflicting answers for the same question across the organization.

• Engineers constantly switched contexts...

  needing to leave their IDEs to hunt for information in documentation platforms...

• Gaps were found too late...

  often discovered during system alerts or in production, turning launches into firefights.

• Feedback was withheld...

  because it was public, so authors rarely learned when guidance was unclear or outdated.

• Documents were never reviewed...

  with no review process, outdated information lingered and misled teams.

We needed to bring reliability into the developer workflow: low-friction, self-serve, and team-owned from day one.

The product team’s initial goal was to build an assessment tool to "shift left" on reliability. However, we realized that building an assessment on this unstable foundation would create a hollow product—it might generate a score, but it would lack the authoritative guidance needed to drive real change and earn developer trust.

This discovery led to a strategic pivot. We aligned stakeholders on a new, more effective approach: to first build a robust, well-maintained knowledge base. This wasn't just about fixing content; it was about creating a versatile platform that would serve as the core for a wider ecosystem of tools.

By solving this foundational problem first, we ensured our solution would:

• Establish Trust: Create a single, authoritative source with clear ownership and maintenance.

• Secure Buy-in: Co-create the platform with SMEs and cross-org leaders.

• Integrate Seamlessly: Embed guidance into existing workflows (Backstage, Jira, CI/CD).

• Enable Innovation: Lay the groundwork for future tools like IDE plugins and automated audits.

We moved from building a single tool to building a scalable system for reliability.

#1. Transformed Unstructured Tribal Knowledge into a Scalable Content System

We moved from scattered documents to a structured content architecture designed for clarity and reuse. This involved two parallel strategies:

• A Modular Content Structure.
  We architected a content system (Module → Assembly) where each Module is typed as a Policy, Standard, Procedure, or Guide. Assemblies are modules combined to form complete guides on larger subjects. This turned ambiguous documents into atomic, referable units of knowledge.

• Contextual Delivery Over Passive Reading. 
  Instead of learning through passive reading, we integrated guidance directly into the tools developers use daily
  
  

✦ Modular, Atomic, Referable Content MECE knowledge centre that provides clear learning path and contextual guidance into assessment workflows, providing clear explanations for every step.

✦ Integrated Guidance Drives Self-Service A built-in support system within tools that explains each step, enabling developers to solve problems independently without needing to find an expert.

#2. For Our Experts: Making Great Content Easy to Create

Everyone agreed that reliable content was the key to fixing our reliability problem. But no one was talking about the people writing it—our subject matter experts. The process was frustratingly inconsistent, and it showed in the docs. I saw a chance to help, using design to bring clarity and calm to how we create knowledge.

• Gave them a content design system to lean on

  • Page level: Templates as the simplest starting points (Concept, Guide, Reference) so they never faced a blank page.

  • Block level: Reusable components like Preconditions, Steps, and Code blocks—each with deep links so their work could be woven precisely into tools like Jira.

  • Word level: A practical writing guide that encouraged a strong, helpful voice using a simple Do/Verify style.

• Shifted the focus from writing to helping

  Introduced lightweight prompts and principles that helped authors step into the reader’s shoes—like “What does ‘done’ look like?”—making it natural to write with purpose.
  
  

We turned a scattered, stressful process into a smooth and repeatable practice. Our pilot authors told us they saved 40% time on first drafts and, more importantly, felt confident their work would actually help people.

Content that speaks like a helpful teammate. Clear, actionable, and written for users to get things done—not to showcase expertise.

A library of thoughtful, reusable content blocks. Making it faster to write, and ensuring everything we published felt like it came from one trusted voice.

#3. Transformed Unstructured Tribal Knowledge into a Scalable Content System

To minimise context switches, we moved the experience from a separate application to a native component of the internal Developer Portal. This integration included:

• Jira integration that turns assessments outcome to tracked tasks, ensuring immediate action.

• A structured Doc Management Platform with version control and trust signals, transforming static pages into measurable, living knowledge.

• High discoverability through a portal index, reinforced by a guided workflow. Learning content is deeply linked to assessments, creating a tight 'Learn → Assess → Apply' cycle.
  
  

Faster, more intuitive discovery ~80% first-click success rate to find correct playbooks from the homepage.

A clear homepage experience Two sibling teams adopted the same integration pattern, proving its scalability.

I helped Dependability Framework evolved from a stalled prototype into a fully-funded, strategic platform.

By championing an entrepreneurial approach that considered the entire product lifecycle, I established a scalable foundation for the product, now a core part of the developer workflow.

✧ 60% faster time for developers to find and apply guidance.
✧ Strong adoption with ~80% first-click success from the developer portal.
✧ Cross-platform reuse of the design pattern by two other internal teams.

🎯 A Strategic Win
The project's success secured dedicated funding and headcount for a full build-out. More importantly, the structured knowledge base we created became the foundational corpus for future AI-powered developer tools, extending the impact far beyond the initial product.

“You didn’t just clean up docs—you gave us a compass.”
— SRE Lead

The project's most significant achievement was establishing a new, sustainable model for enhancing software dependability.

◈ A Sustainable System, Not Another Tool
We designed and secured funding for the first platform that manages the entire content lifecycle—from expert contribution to developer usage—ensuring long-term maintainability and scalability.

◈ Tangible Efficiency Gains
This system consolidated 70+ standards, cutting the "find and apply" time for developers by 55% (from 2.7 hours to 1.2 hours) in pilot tests.

◈ Validated Investment
The proven approach led to the approval of a $400K budget for full development, covering compute resources and two full-time engineers.