How I build
AI-assisted production
- I architected Figma’s learning system, 150+ videos across 20+ programs and 50 learning paths, using AI-assisted production.
Revenue & GTM Enablement · Learning Experience Design · AI Readiness
Better listeners. Better advocates. Better advisors. Better people. That makes them better sellers.
I’ve spent a decade leading revenue, technical, and partner enablement at UKG, Fastly, Figma, and Dropbox. My approach combines technology and humanity to help people become their best selves.
Brandon EdlingMaster Instructional Designer
Association for Talent Development
New Jersey (NYC Metro)
Open to enablement leadership roles · Remote or hybrid
Pick a result to replay it.
Fastly · competency-based SE onboarding
Figma · curriculum and program redesign
Figma · same redesign
Dropbox · structured partner onboarding
Lower is better
Choose a program to see what it was and what it did.
UKG · 2026
Fictional company · 2026
UKG · 2026
Fastly · 2024–26
Figma · 2022–24
Figma · 2022–24
Dropbox · Figma · 2016–24
Dropbox · 2016–22
Career
You can only delegate well what you could evaluate yourself. That shapes how I build with AI, and how I teach people to use it.
How I build
What learners use
Getting people ready
I work in ADDIE, run as a loop, and I make the calls.
Establish the business problem, the target outcome, and how success will be measured.
With learners and their managers: what people actually do today, and where they struggle.
Scale, and a way to compare across roles or regions.
Sales results, support tickets, QA scores, or LMS completions show where the gap is real and how big it is.
Reviewing current materials and observing work in context fills in what people say versus what they do.
Choose a step to see my call, the result, and where AI helps.
My call
AI doesn’t replace talking to learners and SMEs, so I use it to prepare better interviews and synthesize them, not to skip them.
The result
A training specification covering audience, outcomes, and content.
Where AI helps Claude
I feed Claude the messy inputs (stakeholder interview notes, call transcripts, survey exports, existing decks) and have it draft the spec, pressure-test it against the standard analysis questions, and flag gaps and conflicting stakeholder expectations. I also have it analyze performance or CRM data to sharpen the business problem.
My call
I ask for a few structurally different options rather than one answer. Once I approve the storyboard, it becomes the source of truth downstream.
The result
An approved storyboard, saved as a markdown or JSON file: the source of truth for everything downstream.
Where AI helps Claude
I have Claude draft objectives written as observable behaviors, propose sequencing and instructional strategies with the tradeoffs spelled out, write assessment items mapped to each objective, and produce storyboards.
My call
A CLAUDE.md file in each project states the audience, objectives, tone, and style rules. I build reusable skills, one per asset type, so output stays consistent across projects.
The result
Scripts, facilitator guides, job aids, quiz banks, and HTML or interactive activities, consistent across projects.
Where AI helps ClaudeClaude Code
I put the spec and storyboard in a project folder and direct Claude Code to generate the assets from the storyboard, batch-produce role-specific versions and localizations, and run the consistency checks I define: every objective has an assessment, every storyboard frame has an asset.
My call
Prioritization is my call: what people need now, what can wait, and what they can actually absorb while still doing their jobs.
The result
A launch that’s ready to go: comms, talking points, FAQs, and LMS-ready content.
Where AI helps ClaudeClaude Code
I use Claude to draft launch comms, manager talking points, FAQs, and support content. Claude Code handles the repetitive work: formatting content for LMS import, generating enrollment or reminder lists, and preparing rollout checklists.
My call
Claude can analyze the data, but only my learners and the business metrics can tell me whether behavior actually changed.
The result
Evidence against Kirkpatrick-style levels, and the next set of gaps.
Where AI helps ClaudeClaude Code
I run evaluation across every phase, not only at the end. I have Claude check each phase’s output against the one before it (does the build match the storyboard, do the assessments match the objectives) and Claude Code turn survey results, LMS exports, and assessment data into a summary against Kirkpatrick-style levels.
“Brandon is a refreshingly smart and pragmatic professional, with a natural sense for what a user wants and needs.”
“His work is never overly technical or inaccessible — it’s clear, engaging, and designed for real people.”
“Every one-on-one session with Brandon is like a masterclass in professional growth. … Brandon’s guidance has been instrumental in my journey toward becoming a more capable and resilient professional.”
“As a leader Brandon approaches his team with empathy and compassion first. He cares deeply for his people and their success.”
I reply within two business days.
brandon@edlings.org