Safe On-Ramps for Designers and BAs to Contribute to Code

Pillar
People & Organization
Company
Accedo
Audience
Executive, Hiring Manager, Product Leader, Design Leader
Date
2026
PLAYBOOK BRIEF
Capability
Team capability, AI enablement, design leadership
Overview
I set a department ambition for designers and business analysts to contribute to code by the end of the fiscal year, then built the support around it: weekly AI office hours, GitHub fundamentals, sandboxes, technical buddies, starter tasks, access guardrails, quality checks, and protected learning time.
Evidence & limits
Evidence: Weekly office hours, live examples, GitHub practice, sandboxes, technical buddies, starter tasks, guardrails, and protected time formed the support around the ambition. The notes show more sharing and experimentation, not completion of it. Trade-offs: I kept the first contributions narrow and paired them with technical support and review. The model was designed to limit risk, but it used time from me and stronger technical partners, and protected learning time had to be treated as real capacity. Limits and failure modes: A first task that is too broad, no sandbox, unsafe access, no reviewer, no technical support, or learning quietly pushed into evenings and weekends. What this proves: I do not yet know how many people completed the path, whether the fiscal-year ambition was met, or what reached production. Those are the results I would need before presenting this as a completed capability change.
The weekly AI office hours
By the end of the fiscal year, designers and business analysts should be able to contribute to code.
The ambition immediately created practical questions. Which task was bounded enough to start with? Where could someone experiment without touching production? Who would help with GitHub? What access did they need? Who checked the work? When were they expected to learn?
A narrow first contribution
I was not trying to turn every designer or business analyst into a full-stack engineer. I wanted them to influence more of product delivery through bounded contributions.
The first task had to be low-risk and bounded. That meant a protected environment, a narrow starter task, someone technical to help, and a review boundary before the work moved closer to production.
What I brought each week
I ran AI office hours every week. In each session, I brought something I had built or was still tinkering with.
I brought unfinished work to encourage other people to bring their own questions, experiments, and ideas into the same space.
Around those sessions, I put together:
introductory learning and tool setup
GitHub fundamentals
contributions to a project in the company GitHub organisation
protected sandboxes for practice
safe starter tasks
technical buddy support
access guardrails
quality checks before work moved forward
protected learning time rather than after-hours expectations
discussions where people across the organisation could share what they were trying
The notes record more sharing and experimentation around these activities. They do not confirm that every person reached the code-contribution ambition.
The support load
This required close involvement from me. I ran the weekly space, brought examples, taught fundamentals, and kept the next contribution small enough to support. It was not a self-serve course.
Sandboxes and starter tasks kept early contributions bounded. Technical buddies supplied practical support and used time from stronger technical partners.
Protected learning time consumed working-week capacity. It could not be treated as free.
A reader exercise: one on-ramp card
Pick one role and one real, low-risk contribution. Write down:
the task
the sandbox or protected environment
the access required
the technical buddy
the quality check
the protected practice slot
the boundary that stops the work moving closer to production
If one field is missing, resolve it before asking the person to begin.
I still do not know how many people completed the path, whether the fiscal-year ambition was met, or what effect the contributions had on shipped product work. Those are the results I would need before presenting this as a completed capability change.

