Content Design · Foundational Systems

Cody Bay

Turning craft into systems that scale.

Meta · Microsoft · AP Companies

About

Hi. I’m a
word nerd.

Words and ideas are powerful. But in order to do their job, we need systems that make them operable at scale. I’m a senior content designer who specializes in bridging the gap between craft and scale, for both people and AI. Over five years at Meta, I led the operationalization of Facebook’s brand voice, embedding quality into the foundations so that it comes through consistently throughout the product.

Some of the things I like to build:

  • Frameworks that translate content strategy into measurable product heuristics
  • AI training infrastructure and pipelines that teach machines to write on-brand
  • Integrated tooling that gives designers real-time content quality analysis in their workflow
  • Content that just sings. Beneath the fancy frameworks, well-crafted words that hit right are at the core of my passion.

In my career traversing journalism to digital media to UX and AI, I’ve led the way in defining what “good” is, engaged people to take action, and influenced strategy up to executive level. No matter the medium, I make complexity understandable and messaging meaningful.

Selected work

Building an AI Brand Voice System

2026 Initiative lead Facebook

Brand voice quality is subjective by nature. As teams accelerated quickly into AI-native quality workflows and tooling, designers were independently building voice prompts from standards documentation that worked for humans, but left a lot of gaps for AI to fill in, leading to unpredictable and inconsistent results. Meanwhile there wasn’t a systematic way to validate whether AI-generated content really sounded like Facebook.

The challenge

Brand voice quality can’t scale through manual review, and it can’t be embedded in AI without a source of truth that machines can actually parse. As owner of the AI brand voice infrastructure, I set out to build a layered system that provided the validated tooling solution that designers immediately needed, while ensuring that AI content tools across teams and products could call from a trusted and unified foundation.

System architecture — four interconnected layers

01

Foundation

A universal brand voice prompt framework with a 100-point assessment matrix across three quality dimensions.

↓
02

Training data

A pipeline of real examples pulled from audits, crits and fix-a-thons, powered by a Claude skill automating scoring and curation.

↓
03

Tooling

AI-powered custom Figma plugin orchestrated with design system knowledge, delivering contextual, real-time content quality analysis and rewrite suggestions.

↓
04

Feedback flywheel

Fix-a-thon output feeds back into the golden set, improving AI training data with each cycle.

↻

Figma quality check plugin

Figma quality check plugin showing brand voice analysis and rewrite suggestions

Key insights

01/

Define the outer edges. AI needs to understand spectrums, not just pass/fail. Outcome: more nuanced scoring; reduced over-flagging of acceptable content.

02/

Add context awareness. Figma layer traversal and contextual input fields enabled the tool to understand the UI and product context. Outcome: more actionable analysis and feedback.

03/

Build the knowledge base as reusable infrastructure. A Figma plugin is one tool; a knowledge base powers many. Outcome: the same foundation now feeds multiple agentic and skill-based tools across teams.

Impact

V1
Pilot shipped, advanced to round 2 evals
300+
High-engagement strings updated in AI-powered voice fix-a-thons
100%
Experiment coverage via brand quality framework integration
↑
Scale model established for voice quality tooling across Meta apps

“Encodes quality into the systems that generate language.”

— Org leadership

The Brand Scorecard

2025 Sole creator Facebook

Facebook’s brand voice quality was subjective and difficult to scale. Designers evaluated content against persona traits using individual judgment, with no shared rubric or systematic evaluation. Brand review happened through a bottleneck, which meant most product updates shipped without any brand quality check at all. The team had been trying to solve the measurement problem for two years without success.

What I built

A Figma-based Brand Scorecard, enabling voice quality self-assessment translating our persona traits into concrete, checkable heuristics. Rather than another gated review process, I designed this as a tool CDs copy into their design files, bring to crits, and share with cross-functional partners. Iterative versions expanded the checks to include simplification principles and new tonal guardrails validated by UXR.

Outcome: Organic adoption and workflow integration. Canonical resource that became the source documentation for multiple AI training sets. Now, when we ask: “Does this sound like Facebook?”, we have an objective and quantifiable answer.

Impact

XFN
Scorecard data influenced team roadmap and content updates that moved metrics
4+
Use cases powered by one artifact: self-assessment, review, AI training, alignment

External Brand Audit

2025 End-to-end lead Facebook

We crafted Facebook’s brand persona to be durable: an authentic persona doesn’t shift whichever way the wind blows. But occasionally you need to pivot — in this case, leadership was calling for a more “low-key” vibe for young adult audiences — and it was our job to define what that meant. I built the evidence base that informed the brand team’s POV and gave leadership the confidence to commit.

Four phases

01/

Audit. Conducted a comprehensive analysis of young-adult-focused competitor apps, identified the tonal qualities that feel “low-key” and presented recommendations for how the Facebook app could adjust, including adding a missing “Playful” tonal register.

02/

Validate. Partnered with UXR on research validating the Playful tone hypothesis. Crafted content variants, delivered briefs, and translated findings into strategy and guidance, including establishing important tone guardrails for responsible execution.

03/

Align. Shaped executive narratives for design review, supported by a clear implementation plan. Successfully aligned core leads and Head of Facebook on the cross-functional persona shift.

04/

Implement. Shipped and operationalized guidance through peer education and training sessions. Integrated into Brand Scorecard heuristics and AI canonical documentation.

Outcome: A unified brand strategy adopted across marketing, visual design and content design under one umbrella, structured for scaled adoption.

External audit

External brand audit presentation showing competitor analysis

Impact

#1
Highest-ever attendance for CD org peer education sessions
+.27%
DAU increase from voice-optimized content updates (Professional Dashboard)
8/8
Voice-optimized notification variants met or exceeded performance thresholds
3.6→4.1%
CTR increase in voice-optimized footer strings (FB Monetization)

“Cody was an invaluable partner, demonstrating exceptional efficiency and a deep understanding of research needs, successfully translating research recommendations directly into strategy.”

— UX research partner

Let’s talk

What might we build together?

→ Email me → LinkedIn → Download resume