Lean into AI: AI Is a New Language We’re All Learning
- Eric Olsen
- Jul 10
- 4 min read
"AI is a new language we're all learning." Dr. Ada Safak returned to the phrase more than once, and each time it seemed to land — a small permission slip for a room full of people still working out the grammar. In the fourth session of the Lean Into AI webinar series, Ada walked us through how Worthington Enterprises is building AI capability one persona at a time, and shared what has worked, what has stumbled, and what remains unfinished.
The series, hosted by Kelly Reo and The Ohio State University Center for Operational Excellence as part of the Future of People at Work initiative, explores where lean and AI meet. The problem Ada named is one many of us recognize: generic AI training that leaves people underprepared, slows adoption, and delays real business value. Her answer wasn't a bigger course catalog — it was a system.
Start with literacy, not tools
(Normal paragraph — bold "Venny Wong" on first mention; italicize the framework terms AI mindset, shared AI language, and AI skills) Worthington didn't begin with AI. It began with data literacy, a foundation Ada's manager, Venny Wong, had laid years earlier — teaching Ada that "data is a new language." Ada built AI literacy on top of it. The framework rests on three capabilities: an AI mindset (curiosity and honest awareness of limits), a shared AI language (a common vocabulary, since "agent" can mean five different things across a room), and AI skills (applying tools to real work). Skills don't stick, she observed, without the mindset and the language underneath them.
Match the person to the proficiency
From there, training is tailored by persona and proficiency. Worthington teaches the model as three tiers — workforce, managers, executives — while internally running a four-persona version through its McConnell Leadership Academy, complete with "trailblazers" and "crystals." Proficiency is a four-level ladder: starter → practitioner → creator → shaper. The workforce targets creator; managers and executives target practitioner, fluent users who guide and shape rather than build all day. Steve Pereira pressed on a lean tension here: leaders captivated by what's possible can flood teams with prototypes no one pulled for. Eric O named the same knot from the audience — "should versus can" — noting that "can" has grown enormous with AI. Defining expectations by role, Ada argued, keeps capability pointed at value.
Build community, not just courses
Much of Worthington's progress came from people, not platforms. An AI Champions Network, show-and-tell sessions, "Wednesday Whiz" Copilot tips, AI Spotlights, a dedicated AI community page, and twice-weekly office hours lowered the bar to participation. Ada described a recent show-and-tell where a champion, Clint, demonstrated turning a ten-hour task into ten minutes inside Excel — and three colleagues spontaneously raised their hands to share related solutions, something that "never happened before." Steve Pereira captured the conditions for that shift in the chat: "Curiosity + Capacity. People need space and safety to learn." To keep pace with fast-changing tools, the team even builds learning tracks inside the tools themselves, letting the environment stay current so they don't have to chase it.
Honest about the hard parts
Ada was candid that the smooth slides hide a bumpy road. Early Copilot adoption sat at 40 percent; today it is at least 95. People resisted — some wanted ChatGPT instead of Copilot, some ignored voluntary invites, some worried about hallucinations, and some raised environmental-impact concerns (the top-upvoted question of the session). A colleague near retirement simply didn't want to change how he worked, until he watched a five-hour task shrink to ten minutes at the desk next to him. Worthington's antidote to agent sprawl is a problem-first habit: ask "what's the problem?" before "what's the solution?" — supported by a scoping custom GPT called ScopeSensei. Sometimes the answer isn't AI at all. Or, as the training repeats, use AI as a starting point, not an ending.
What stayed with us is how ordinary the levers are: a shared language, clear role expectations, open office hours, and the patience to show value rather than mandate it. None of it is finished — formalizing the champions role, matching supply to real demand, and answering the environmental-impact question all remain open. That unfinished honesty is exactly the lean humble spirit this series keeps surfacing.
Continue the conversation
The Lean Into AI series continues in August. Recordings live in the Future of People at Work media library at fpwork.org — and if you have a story like Ada's, the invitation is open to share it. https://fpwork.org
Knowledge Map
Process Keywords
persona-based training, data literacy, AI literacy, proficiency ladder, problem-first scoping, in-tool learning tracks, AI champions network, office hours, show-and-tell, adoption measurement, train-the-trainer
Context Keywords
generic-training fatigue, low tool adoption, employee resistance, hallucination concerns, environmental-impact concerns, agent sprawl, keeping pace with changing tools, voluntary versus mandatory participation
Application Triggers
If your AI training feels generic and adoption is low → the persona-and-proficiency model offers a way to tailor expectations by role.
If new tools keep outpacing your curriculum → building learning tracks inside the tools may help you stay current.
If teams are generating overlapping AI solutions → a problem-first scoping habit (a Scope Sensei pattern) can refocus effort on value.
If people resist a mandated tool → peer-led show-and-tells and office hours can demonstrate value without pressure.
If environmental-impact concerns are a source of resistance on your team → this is a shared open question the community is still working on.
Related Continual Improvement Themes
respect for people, develop people, problem-first thinking, pull versus push, organizational learning, sustainable adoption
This post was developed through the Lean Into AI webinar discussion and synthesized with Claude AI assistance. It represents ongoing work by the Future of People at Work initiative, a collaboration of Catalysis, Central Coast Lean, GBMP Consulting Group, Imagining Excellence, Lean Enterprise Institute, Shingo Institute, The Ohio State University Center for Operational Excellence, Toyota Production System Support Center (TSSC), and University of Kentucky Pigman College of Engineering.
Comments