Kamille Rigsby Lee
kamillelxd / case-studies / ai-enablement.lxd
Case study · Concept
AI enablement · Financial services · Concept case study

Helping a global workforce work alongside AI agents.

A concept case study based on my AI enablement and onboarding work at a global payments company. It shows how I turn agentic AI and Copilot concepts into learning people can use on the job.

Day 1 with your AI agent

Three small wins before lunch

1 of 3 done · tap to check off
Recreated asset
Industry
Global financial services
Audience
Employees across regions, from AI-curious to power users
My role
Learning experience design; led designers and contractors
Format
AI onboarding, enablement paths, performance support
Focus
Agentic AI, Copilot, future of work
Note
Concept piece. No proprietary content or data.
AAnchor in Human Needs
Who am I designing for?

Everyone, at every level of AI confidence.

The audience spans regions, roles, and languages. Some people are excited about AI agents, some are skeptical, and many are quietly worried about getting it wrong. One course can't meet all of them, so I mapped where people are on the path to trust.

Recreated asset · the AI trust ladder · tap a stage
What they needPlain-language answers about what the tool is, what it can see, and why it's safe to try.
CClarify the Why
What problem am I solving, and why does it matter?

The tools were ready. People needed a way in.

AI rollouts rarely stall because of the technology. They stall on confidence, clarity, and trust. Feature tours teach where the buttons are. They don't teach judgment, and judgment is what makes AI safe and useful at work.

Build confidence, not only awarenessTeach judgment, not only featuresDesign for global audiences
CCreate Narrative & Context
How do I make this meaningful?

Learn it inside a real workday.

Instead of a tour of features, the learning follows one employee through a normal day: handing a task to an agent, reviewing what comes back, and deciding what stays human.

  1. DelegateAsk the agent to summarize overnight requests.
  2. ReviewCheck a drafted reply before it goes to a partner.
  3. DecideChoose what not to hand off, and why.
  4. SharePost a prompt that worked to the team channel.
EEnable Through Action
What will learners actually do?

Practice the judgment call, not just the click.

Scenarios put learners in the moment where AI skill matters most: deciding what to do with output they didn't write.

Sample scenario · choose what you'd do
Your agent drafted a summary of a partner's quarterly results from three internal reports. You need it for a 3 p.m. meeting.
Choose what you'd do.
SSupport the Ecosystem
How will learning transfer to the real world?

Help that shows up where the work happens.

Confidence fades fast without support. The learning extends into the flow of work with quick references, shared prompts, and people to ask.

Recreated asset · job aid

Before you trust the output

  • Where did this come from?
  • Do the key numbers match the source?
  • Is anything here confidential?
  • Would I sign my name to it?
Concept · playbook

Prompt playbook

Role-based prompts that worked, collected from real teams and easy to copy.

Concept · community

AI champions

Peers who help their teams get unstuck.

Concept · in the flow

Just-in-time tips

Two-minute videos for the moment someone needs them.

Concept · live

Office hours

A safe place to ask the questions people won't post publicly.

SScale Through Innovation
How can this solution grow and improve?

Human + AI: who did what.

I use AI in my own design process, with clear lines. AI speeds up the drafting. The decisions about people, tone, and what matters stay with me. I also led fellow designers and contractors on this work, coordinating contributions and owning delivery.

  • AIFirst-draft outlines from SME documents
  • KamilleLearner research and the trust ladder
  • AIScenario variations for different roles
  • KamilleDeciding what not to teach
  • AIDraft knowledge-check questions
  • KamilleMaking scenarios feel real and fair
  • AIPlain-language rewrites for global readers
  • KamilleAccessibility review and final sign-off
How I'd measure it

Confidence, correct use, and adoption.

Completion rates can't show whether someone trusts the tool or uses it well. I'd track a short confidence check before and after, how often people verify AI output in practice scenarios, and adoption over the first 90 days.

What I'd add next timePair new users with a champion in their first week. Peer support is the fastest way up the trust ladder.