# 32 Claude Skills for AI Adoption at Work, From the First Pilot to the Habit 32 Claude skills for getting a team to use AI well: find where it helps, run a small pilot, teach judgement, support managers, make it safe to question, measure what changed and keep the habit, with nobody's use tracked. - Page: https://lantern.meet-polar-bear.com/skills/ai-adoption-pack - Download: https://lantern.meet-polar-bear.com/skills/ai-adoption-pack.zip - Source: https://github.com/polar-bear-org/claude-skills/tree/main/packs/ai-adoption-pack - Price: free ## Install in Claude Code ``` /plugin marketplace add polar-bear-org/claude-skills /plugin install ai-adoption-pack@polar-bear-skills ``` ## The 32 skills you get The set follows AI adoption in seven stages: find where it helps, start small, teach it, support managers, make it safe to question, measure what changed and sustain it. ### 1 · Find where it helps - **AI Adoption Brief** — use when: The instruction is "get everyone using AI" and nobody has said what anyone should do differently on Monday. Output: One observable work change per role, what stops as well as starts, scope, sponsor and decision owner, the first check with the people affected. - **AI Starting Point Survey** — use when: You are about to plan training for people you have never asked. Output: 10 to 15 anonymous questions, a minimum group size you set, a never-asked list, a team-level summary template. - **AI Use Case Map** — use when: People say "I do not see what I would use it for". Output: Recurring tasks role by role, where AI helps and where it does not, what is off limits, who checks, the first three use cases to try. - **AI Adoption Barrier Diagnosis** — use when: People are not using it and the first proposal is another training. Output: Competing explanations with evidence for and against, the smallest respectful check, a support matched to each barrier. ### 2 · Start small - **AI Pilot Plan** — use when: The choice is between everyone at once and a pilot that proves nothing. Output: The one question the pilot answers, a prediction written first, measures with definitions, stop and rollback conditions, the review rule. - **Team Involvement Plan** — use when: The rollout is "consulted" after the decision and people can tell. Output: Decided, open and unknown, who is affected, three concrete questions, channels with a private route, a ledger of what was heard. - **Adoption Capacity Check** — use when: People are told to learn AI on top of a full week. Output: Learning and practice time per role, current load, the gap, what can pause or move, options with trade-offs. ### 3 · Teach it - **AI Basics Curriculum** — use when: The AI training on offer is a tool tour and a prompt list that dates in a month. Output: A short path by role built on four judgements, a practice in each unit, what is left out and why, a refresh date. - **Hands-On AI Session Plan** — use when: The last AI session was a webinar where nobody touched the tool. Output: A timed plan where each person brings one real task, a demonstration, a practice round, a compare-and-check round, a no-screens fallback. - **AI Practice Tasks** — use when: People left the session keen and never opened the tool again. Output: Three to five tasks from the team's own work, a first attempt before any model answer, a common exception, a check of the work. - **AI Output Check** — use when: People were taught to prompt and nobody was taught to check. Output: A one-page check a person runs before using AI output for one type of work, with a stop condition and who signs. - **AI Job Aid** — use when: People need the answer in the middle of the task, not a module. Output: One card for one task: when to use AI, what to give it, what to check, what never goes in, who to ask. - **AI Train-the-Trainer Kit** — use when: One person cannot teach everyone and the internal trainers are a few weeks ahead of the room. Output: Delivery learning points, a teach-back plan, practice rounds that rise in difficulty, a readiness conversation with no score. ### 4 · Support managers - **Manager AI Kit** — use when: The manager never mentions AI, or mentions it only as a target. Output: What the manager says, shows, asks and removes, and what a manager never does. - **Team AI Agreement** — use when: Some people use AI for everything, some refuse, and nobody has said what is fine. Output: Agreed uses per task, a never-put-in list, what checking means, whose name is on the work, a review date. - **Manager Follow-Up Guide** — use when: People come back from the AI session to a full diary and a manager who never asks. Output: For one session: the conversation before, the first real task after, the conversation after, check-ins, a barrier log read in aggregate. - **Freed Time Agreement** — use when: The time AI saves quietly turns into more tasks and longer days. Output: A team conversation guide and a one-page record of what takes less effort, what stops, where freed time goes, a workload check date. ### 5 · Make it safe to question - **AI Rules Explainer** — use when: The policy is eleven pages long and people still do not know what they can paste. Output: Your organisation's own AI rules as can, cannot and ask first, with real task examples and who to ask. - **AI and Jobs Answer** — use when: "Will this replace my job?" is in the room at every session. Output: The questions people are asking, and for each what is decided, open and unknown, and who gives the answer. - **AI Concern Statement** — use when: Someone can see a problem and cannot say it without sounding against progress. Output: One person's concern in their own words: what I saw, what I think it means, what could follow, what I am asking for. - **Speak-Up Practice Round** — use when: The concern is written and the conversation with the sponsor is tomorrow. Output: A short rehearsal one turn at a time, comments on the choices made, a revised opening, the next step they choose. - **Team Challenge Invitation** — use when: The only people who speak are the ones who already agree. Output: An invitation to question the plan before it is final, questions sent ahead, spoken, written and private routes, a response commitment. - **Concern Response Log** — use when: Concerns were collected and nobody ever heard what happened to them. Output: For each concern: the substance checked, what changed, what did not and why, owners and dates, a shareable update. ### 6 · Measure what changed - **AI Value Measures** — use when: The only measure anyone has proposed is "hours saved", self-reported. Output: Four to six measures agreed before the pilot, each with a definition, a source and a baseline, one balancing measure for workload. - **AI Adoption Evidence Review** — use when: The dashboard says most people logged in and the sponsor reads that as adoption. Output: Use, usefulness and side effects kept apart at team level, counts beside percentages, what the evidence cannot show. - **AI Use Stories** — use when: The numbers are thin and the examples in circulation are demos. Output: A consent note, an interview guide, three to five anonymous accounts of real use, each marked as one case. - **AI Adoption Executive Update** — use when: Executives want one number for AI and the honest answer is more useful. Output: One page, bottom line first: what changed, the evidence, what it does not show, concerns raised, the decision asked. ### 7 · Sustain it - **AI Champions Network** — use when: Three enthusiasts answer everyone's questions in their own time and are about to stop. Output: A charter for a volunteer peer network: the role, protected time, how people join and leave, the route to the sponsor. - **Shared Prompt Library** — use when: Good prompts live in private chats and the shared folder holds forty untested ones. Output: The entry format, how entries are proposed, tested, updated and retired, where the library lives, the review rhythm. - **Local Adaptation Review** — use when: Teams are quietly changing the agreed way of working with AI. Output: One team's variation: what changed and why, the safeguard it must keep, a bounded test with a rollback, a versioned note. - **AI Rollout Decision** — use when: The pilot went well with volunteers and the sponsor wants everyone on it next month. Output: The pilot's evidence and its limits, how the next teams differ, options with trade-offs, pause rules, dissent on record. - **AI Habit Review** — use when: The launch was six months ago and half the team is back to the old way. Output: A quarterly review: the original need, current evidence, what changed in the tool, what to keep, simplify, refresh or retire. ## Workflows - **Find where it helps** (The instruction is “get everyone using AI” and nothing else.) AI Adoption Brief → AI Starting Point Survey → AI Use Case Map → AI Adoption Barrier Diagnosis - **Start with a pilot** (The choice is between everyone at once and a pilot that proves nothing.) AI Pilot Plan → Team Involvement Plan → Adoption Capacity Check - **Teach on real tasks** (The training on offer is a tool tour and a prompt list.) AI Basics Curriculum → Hands-On AI Session Plan → AI Practice Tasks → AI Output Check → AI Job Aid → AI Train-the-Trainer Kit - **Give managers a part** (People come back from the session to a manager who never asks.) Manager AI Kit → Team AI Agreement → Manager Follow-Up Guide → Freed Time Agreement - **Make it safe to question** (The rules are unclear, jobs are on people's minds and concerns go nowhere.) AI Rules Explainer → AI and Jobs Answer → AI Concern Statement → Speak-Up Practice Round → Team Challenge Invitation → Concern Response Log - **Measure what changed** (The sponsor wants hours saved and the dashboard only shows logins.) AI Value Measures → AI Adoption Evidence Review → AI Use Stories → AI Adoption Executive Update - **Keep the habit** (The pilot went well and the question is what happens next.) AI Champions Network → Shared Prompt Library → Local Adaptation Review → AI Rollout Decision → AI Habit Review ## The quality bar - One method, applied properly: each skill uses the real mechanics of its method (the questions, the levels, the order of the pass), not a generic “gather, analyse, recommend” - It tells you where it breaks: every skill says when not to use it and which skill in the guide fits instead - Same shape every time: a fixed output template, a short “done when” check, and a decision that a named person makes - No invented facts: no made-up adoption rates, savings, quotes or findings. Missing data becomes a bracketed placeholder or an open question, and no productivity result is promised - Team level only: nothing builds a list, score or ranking of who uses AI. Surveys are anonymous, results are reported by team, small groups are merged or suppressed, champions are volunteers and stories are told with consent - Rules and law stay with their owners: the skills explain rules your organisation already has. Every legal, data or employment point ends with “check with a qualified adviser” - The red line: Adoption is something people choose: Claude prepares the plan, the sessions and the evidence, a named person makes every call, nobody's use of AI is tracked, scored or ranked, and no saving is claimed that the organisation's own evidence does not show ## Questions people ask ### What is in the Claude Guide for AI Adoption pack? 32 Claude skills that follow AI adoption in seven stages: find where it helps, start small, teach it, support managers, make it safe to question, measure what changed and sustain it. Each skill keeps one method inside, with a fixed output template. ### How do I install these Claude skills? In Claude Code, two lines install all 32 at once. In Claude.ai or Claude Desktop, download the zip, open Settings, Skills, Add, and upload the skills you want. Either way it takes about five minutes. ### Do I need Claude Code to use them? No. The skills work in Claude.ai, Claude Desktop and Claude Code. In the app you upload them yourself; in Claude Code one command installs the whole pack. ### Do the skills work with ChatGPT or other assistants? They are built as Claude Skills, and installing them is a Claude feature. Each one is a plain Markdown file, so you can read it and reuse the method elsewhere, but it will not load and trigger by itself outside Claude. ### Does Claude track who is using AI? No. Adoption is something people choose. Claude prepares the plan, the sessions and the evidence, and a named person makes every call. Nobody's use of AI is tracked, scored or ranked, and no saving is claimed that the organisation's own evidence does not show. ### Is the pack free? Yes. It is free to use inside your company and in client work, and not for resale. There is no signup: the download and the install command are on this page. ## About Written by Polar Bear (https://meet-polar-bear.com/). Published on Lantern (https://lantern.meet-polar-bear.com/), which turns your notes, documents and call transcripts into social post drafts you review and publish.