Most internal AI explanations confuse people. Either too technical ("It's a transformer-based language model trained on...") or too vague ("AI is going to change everything"). Here's the framing that actually produces understanding and adoption.
Frame AI as "a junior teammate who has read everything but knows nothing specific about your company." Their first job is to teach the AI about your work; in return, the AI absorbs the boring parts of their job. Practical, accurate, non-threatening, accurate enough to be useful.
"Think of Claude (or ChatGPT) as a brand-new junior teammate. Very smart. Has read most of the internet. Cannot type or speak — only writes. Knows nothing specific about us, our customers, our voice, our work. Our job is to teach it; in exchange, it absorbs the boring parts of our work — drafts, summaries, research, formatting."
Why this works:
"OK team, quick overview of why we're rolling out [Claude/ChatGPT]."
"Think of it as a brand-new junior teammate who has read everything but knows nothing specific about us. They can write well, summarize well, research well. They can't take meetings, can't make decisions for us, and won't know our customers or our voice unless we teach them."
"Here's what we're doing: we're building Projects loaded with our examples, our voice, our processes. Then any of us can use those Projects to draft proposals, draft emails, synthesize meetings, do research. Faster than we do it ourselves; quality holds because we teach it our standards."
"What we're NOT doing: replacing anyone. The work this absorbs is the drudgery part — drafting, formatting, summarizing. The judgment and relationship work stays with us, and frankly we'll have more time for it."
"The ask: use it. Try it on 2-3 things per week to start. Tell [AI lead] what worked and what didn't. We'll get good at this together."
Answer honestly. "Not in the foreseeable future. It changes what your job is — less drafting, more judgment. Most people I know who use it well report being more valuable, not less."
"We're using [enterprise tier] which doesn't train on our inputs. We have a policy about what kind of data goes where; I'll send it."
"It will be sometimes. Our job is to review its output before anything ships. Like reviewing a junior teammate's draft."
"We're protecting 2 hours per week for the next month. Calendar it now. The time it saves you compounds."
No. The function leader or COO. HR is for policy questions; rollout is about work.
No. One page max. Long FAQs feel defensive.
Yes — for people who joined later. Keep it under 10 minutes.
For ongoing skill development, yes. For the initial framing, no — they don't know your team.
Get fluent before rolling out. Use it daily for 30 days. You cannot lead what you don't use.