The AI Switch Framework Change Coach skill for Claude
You describe a change that just won't stick, and the skill designs it from three angles at once: the Rider, who needs direction, the Elephant, who needs feeling, and the Path, which has to make the chosen behaviour easier to carry out. That's Switch, the framework by Chip and Dan Heath. First it gathers your situation, then the three tracks, then a 30-day plan with exactly three actions. Not a list of twenty resolutions you never get round to.
the rules above come from the zip on this page · SKILL.md is 12,738 bytes
Situation in, three tracks out
reg. S.001This is a shortened conditional sketch based on the second example in SKILL.md. The literal input does not establish a cause, a feasible target or a meeting schedule. Phase 1 gathers context before advice; proposed actions remain hypotheses until the baseline, quality, workload and constraints are understood.
What the AI Switch Framework Change Coach is
The AI Switch Framework Change Coach is a free skill from our skill library. A skill is an instruction file, SKILL.md, that gives an AI assistant a fixed way of working for a task. No software to install, no subscription, no connection to your systems: a 2,063-word text file that tells Claude how to design a change according to Switch, what order the phases come in, which nine techniques it uses in total, three per track, and what it never does under any circumstance. You download the zip at the top of this page, put it in your Claude environment, and from that moment Claude thinks in Rider, Elephant and Path as soon as you say something needs to change.
Switch offers three angles for exploring a change: direction, motivation and environment. They are questions to investigate together, not an exhaustive explanation of resistance or proof that a change fails after three weeks. The skill keeps all three in view without diagnosing people from a short request.
This skill is written for anyone who needs to get something moving without having authority over it, or who has the authority and still no result: business owners pivoting their company, leaders who need to change culture, teams stuck in old patterns, and people who want to change a personal habit. Together with the rest of the skill library, we make it available for free from our AI and automation service, with no account and no sales email afterwards.
The file asks for context before advice. It covers five topics and uses information already supplied instead of repeating questions. Three months is an illustrative horizon; the actual deadline is agreed with the user. The prompt does not guarantee the quality or outcome of an AI response.
Why change can stall
The Heath brothers use Jonathan Haidt’s Rider and Elephant metaphor, with the Path representing the environment. The Rider stands for planning and analysis; the six-tonne Elephant illustrates emotion and motivation. It is an organising metaphor, not a literal model of the brain, a clinical diagnosis or a measurement of why a particular person has not acted.
Explore whether direction is clear. The first track asks about observable behaviour, useful bright spots and a concrete destination. Ambiguity can make action harder, but it does not establish the cause of a person’s delay. Script the Critical Moves turns an abstract aim into a proposed behaviour that can be discussed and tested.
Explore motivation alongside the plan. Feelings, energy and identity may matter, but tiredness or fear must not be assumed. The techniques invite a meaningful connection, a small first step and opportunities to learn. A first success within a week is something to assess, not a promised result.
Examine the environment. Physical or digital arrangements, routines and voluntary peer support may make a chosen behaviour easier. They do not guarantee success, and reliance on willpower does not by itself establish failure. Any dashboard, meeting or other cue should fit the actual situation.
The workflow considers Rider, Elephant and Path together. This is a coverage rule for designing a plan, not proof of why most organisations or changes fail. The analysis remains provisional and should make other constraints, missing facts and uncertainty visible.
In diff form, with lines pointing to the examples in the skill: what goes out and what comes back in its place.
Rider, Elephant, Path
reg. S.002Phase 2 explores each track through questions, a provisional working hypothesis and three to five candidate actions. These are options, not simultaneous commitments. There are nine techniques in total, three per track. Phase 3 selects one final action per track; considering all three does not guarantee that a change succeeds.
What's actually in the SKILL.md
A skill is only as good as its instructions, so we simply describe them here. At the top is the frontmatter: the name ai-switch-framework-change-coach, version 1.0.0, MIT licence, and a compact description. The full original trigger list remains in the opening body text. Alongside the term switch framework and the names Chip Heath and Dan Heath, it's mostly made up of ordinary sentences: I need to change, the team won't come along, I can't get moving, why doesn't change stick, I've been trying this for months, and the sharpest of them all, I know what I should do but I don't do it. Those phrases are matching instructions, not a guarantee that a particular assistant will activate the skill.
The file credits the Rider and Elephant metaphor to Jonathan Haidt and describes the Heath brothers’ framework around the Rider, Elephant and Path. It invites investigation of direction, motivation and environment without claiming these explain almost all resistance or delay. The metaphor is not a diagnosis of a person.
Remarkably, it also states when it should not trigger, which is rare for an instruction file. Not when someone is already in the middle of execution and just needs tactical help. Not for a one-off decision. And not for a conflict in a specific conversation. For that last case, the file points to a negotiation technique, and in our library that's the Negotiator skill.
The method itself is fixed in three phases, in this exact order, because they reinforce one another. Phase 1 is outlining the situation with a maximum of five sharp questions: what change, as concretely as possible in behaviour; who needs to change; what's the current situation and what's already been tried; what does success look like in three months in visible behaviour; and what deadline you're working towards. Then comes a brief reflection, without judgement and without advice. Phase 2 is the three-track analysis shown above in FIG.02. Phase 3 is the 30-day plan.
The writing chapter has six rules: British English unless the user explicitly writes in another language; no lists of more than five items; explain jargon; avoid empty management language such as synergy and alignment; address the reader as you; and be direct without being bossy. The English download does not instruct the assistant to answer in Dutch.
At the end are the sources the skill is allowed to cite and the five things it never does. Those five refusals get their own figure further down this page, because they say more about the quality of the advice than the techniques themselves.
Three actions, thirty days
reg. S.003Phase 3 selects exactly three final actions, one per track, from the Phase 2 options. It adds a check-in within seven days and a measurable indicator. Three is the plan format, not a claim that a fixed number of actions overrides necessary safety, quality or other obligations.
The theory the skill rests on
The SKILL.md has a separate chapter listing the sources the skill is allowed to cite, and that's deliberate: it lets you check them yourself and decide whether you agree.
The main source is Switch itself. Switch: How to Change Things When Change Is Hard, by Chip Heath and Dan Heath, published in 2010 by Crown Business. The book isn't a theoretical work but a collection of cases where change did succeed, organised along the three tracks. That also explains the shape of the skill: not a model to fill in, but a series of questions that force you to come up with something concrete for each track.
The metaphor comes from Jonathan Haidt. The Happiness Hypothesis, published in 2006 by Basic Books, introduced the Rider on the Elephant as an image for the relationship between the rational and emotional parts of the mind. The skill explicitly credits the metaphor to Haidt rather than the Heath brothers, which is more accurate than what most summaries online say. The Heath brothers' addition is the Path: the environment as a third lever alongside reason and feeling.
The skill may cite the related titles, not parrot them. Four books are listed that you don't need to read but that it's allowed to mention: Made to Stick from 2007, Decisive from 2013 and The Power of Moments from 2017, all three by the Heath brothers, plus Thinking, Fast and Slow by Daniel Kahneman from 2011. If you want to apply the first of those four directly, there's a separate Made to Stick Coach skill that makes messages sticky, which comes in handy precisely for the Elephant track.
Even without installing the skill, you can take apart any stuck change with these three questions: is the direction concrete enough, does anyone feel anything about it, and does the environment make the new behaviour easier than the old one. If you want to look at the same question through a different model, the Fogg Behaviour Model Coach skill splits behaviour into motivation, ability and trigger, and the Atomic Habits Coach skill works the Path track out into a system of habits.
What the skill never does
reg. S.004The SKILL.md closes with five things the skill never does, and all five are about the quality of the advice rather than safety. Which makes sense: for a change coach, bad advice is the real risk. In conversation those rules play out like this: each rule is a request you might make, together with the response the skill gives according to its own instructions.
Installing in Claude Code, Claude.ai or Codex
The zip contains a folder, ai-switch-framework-change-coach, with the SKILL.md inside it. Installing is simply a matter of putting that file in the right place, and that place differs per environment. SKILL.md has been an open standard since December 2025, so the same skill also works in Codex, Cursor and Gemini CLI. So you're not downloading a Claude file but a working instruction any modern AI assistant can read.
- Unzip it into
~/.claude/skills/(or.claude/skills/in your project). - Claude then recognises the skill automatically as soon as you start talking about change, habits, or a team that isn't keeping up.
- Start Claude Code with claude, then enter
/ai-switch-framework-change-coachinside the session.
- Go to Customize and then Skills.
- Upload the zip there as a skill.
- Or paste the contents of SKILL.md into the project instructions of a Project.
- Open
AGENTS.mdin your repo. - Paste the contents of SKILL.md into it, or place SKILL.md alongside it as a separate file and refer to it from
AGENTS.md. - Codex reads that along with every session.
After that, using it is simple: describe the change you want to make and answer the phase 1 questions honestly, including the question of what you've already tried. Those failed attempts are especially useful, because they may reveal constraints, assumptions or approaches worth revisiting. If you're new to this kind of file, read what Claude skills are NL first and then how to install Claude skills NL. If you want to write one yourself, for example for the changes that keep recurring in your organisation, writing a SKILL.md NL explains how to go about it.
When to use it, and when not to
The trigger list includes personal habits, stalled teams, culture changes and business pivots. These situations may involve missing knowledge, resources, conflicting priorities or other constraints as well as difficulty taking action. The first phase investigates the circumstances instead of assuming a shared cause.
The skill is just as clear about when not to use it, and there are three situations. The first: you're already in the middle of execution and just need tactical help. Then targeted help may be more appropriate than a full change plan. The second: it's a one-off decision. A choice between two options isn't a change project, the file points to WRAP Decision Maker. Another tool for examining a proposed plan is the Pre-mortem Analyst skill, which lets a plan fail in advance to see where it breaks. The third: it's a conflict in a specific conversation. For that, the file points to Negotiator. A different conversation tool available in the library is the Nonviolent Communication Coach skill.
The five intake topics help make the situation concrete: desired behaviour, people involved, current state and earlier attempts, observable success and timing. Questions are not meant to establish that a user is uncomfortable or deficient. Use known facts, clarify important gaps and agree a realistic horizon.
Switch is a structure for designing and testing change, not a guarantee or a measured increase in success rates for this prompt. It remains possible that the proposed direction is unsuitable or the team has valid objections. Intake and review should leave room to change the goal or stop an inappropriate intervention.
Run it yourself, or have it run for you
reg. S.005This skill is the free do-it-yourself version of work we also deliver as a service. It stays complete and with no catches, but be aware what a skill actually is: it teaches your AI how to do something, while every new session starts empty. A skill is neither the engine nor the memory. In a 30-day change project you feel that straight away, because you have to explain what happened in week one all over again every session. Anyone who wants it differently has two follow-up steps: hand off the engine, or sort out the memory.
where you are now The skill: you're the engine You run the Switch Change Coach yourself in Claude, Codex or Cursor. Costs nothing, works today, and you keep it entirely in your own hands: no trial period, no gated parts. The limit is your own time: it only happens when you prompt, and you have to remember the check-in on day seven yourself.
have it prepared for you The AI worker: it's ready without you having to prompt Exactly this work, but as a service: an AI worker prepares recurring work without you having to sit behind Claude for it. Control stays with you, because the output stays a draft until a human gives approval. We deliver this through Mansotti, the company TheSEO trades under, which builds the Quote Worker, the Sales Worker and the Reporting Worker. For a change project, the Reporting Worker is the most useful of the three, because it produces the weekly and monthly reports that show you whether the new behaviour is actually happening.
everything from one source Jarvis: all your AI's work from the same company knowledge The skill teaches the AI, the brain is where the memory lives. Want all your AI's to work from the same company knowledge: that's Jarvis, the organisational brain. It connects ChatGPT, Claude, Codex and your people to the same projects, core knowledge and decisions, so your next AI session doesn't start from scratch. Jarvis selects context, delegates work and only stores what's been approved: client knowledge stays isolated and every step leaves an auditable trail. For a change running over 30 days, that's the difference between a plan and a project. What that means in practice, from the plans to your first week, you can read under Jarvis itself.
What Jarvis delivers in practice
reg. S.006Step 3 deserves more than a paragraph, because this is the difference between a smart chat and a system you can build on. Jarvis is the organisational brain: it remembers what your AI's need to know, distributes the work and keeps track of what happened. With change, that's the most sensitive point of all, because a 30-day project consists of sessions that need to know each other. Here's what a new session gets back.
We've been working this way ourselves for months. Every agent session, every task and every decision goes into the log and can be read back. A new session therefore doesn't start blank: it first retrieves the recorded decisions, the running projects and the latest changes, and carries on where the previous one stopped. For a 30-day project, that's precisely the part that usually goes wrong: week three, when nobody remembers any more what was agreed in week one. So not a promise, but the way we work every day.
See all four plans at jarvis/prijzen NL. Through the waiting list NL, you simply pass on your preferred plan, without obligation. That doesn't create an account, an order or a payment obligation. We discuss business customisation first.
The skills around it
reg. S.007Switch is one of several behaviour and change models in the library. Their emphases differ and may overlap; the right choice depends on the question and context.
The other change models
same question, different lensSwitch looks at reason, feeling and environment all at once. These two look at individual behaviour and at the organisation.
Fogg Behaviour Model CoachSplits behaviour into motivation, ability and trigger, and shows whether it's about wanting to or being able to.SKILL Kotter Change CoachChange that has to move through an entire organisation, with urgency, coalition and visible early wins.SKILLThe environment and the habit
the Path track worked outThese two related skills explore routines and environmental support in more detail.
Atomic Habits CoachBuilding habits via cue, craving, response and reward: exactly what Build Habits is about.SKILL EAST Framework Behaviour CoachEasier, more attractive, more social and better timed: nudges for other people's behaviour.SKILLWhat the skill itself refers you to
the three non-triggersFor three situations, the skill itself says Switch isn't the right tool. This is what's ready instead.
NegotiatorFor the conflict in a specific conversation, which the file explicitly points to.SKILL Pre-mortem AnalystFor the one-off decision: let your plan fail in advance to see where it breaks.SKILL Made to Stick CoachFrom the same authors, for the message that needs to reach the Elephant.SKILLLooking further
the contextWhere this skill comes from, and what else there is.
The whole skill library100 free skills in one place, sorted by topic.HUB AI and automationThe service behind it: from individual skills to working automation in your business.SRV AI trainingIf your team wants to learn to set up this kind of work itself.SRV Knowledge baseSearchable articles on SEO, AI and online visibility.DOCFrequently asked questions
What does the AI Switch Framework Change Coach skill cost?
Nothing. It's free, the frontmatter of the file states the MIT licence, and you don't need to create an account or leave an email address. You download a 5.5 KB zip containing the folder ai-switch-framework-change-coach and a single file, SKILL.md, 2,063 words long. That's the complete skill. There's no paid version and no sales email follows.
Does this skill also work in Codex, Cursor or Gemini CLI?
Yes. SKILL.md has been an open standard since December 2025, so the same file also works in Codex, Cursor, Gemini CLI and other tools that follow the standard. In Codex you unzip it into .agents/skills/ in your project, or into ~/.agents/skills/ for all your projects; Codex has supported SKILL.md directly since the open standard of December 2025. Putting the contents of SKILL.md into your AGENTS.md still works too. The instructions are just readable text, so any assistant that accepts instruction files can work with it.
What are the Rider, the Elephant and the Path?
Rider, Elephant and Path are parts of an organising metaphor: planning and analysis, emotion and motivation, and the surrounding environment. The six-tonne Elephant is illustrative, not a brain measurement or clinical diagnosis. The Rider and Elephant come from Jonathan Haidt’s The Happiness Hypothesis; the Heath brothers use them in Switch alongside the Path.
How many phases does the skill go through?
Three phases, in order: clarify the situation through up to five questions while using facts already supplied; explore three tracks with provisional hypotheses and three to five candidate actions each; then select a 30-day plan with one Rider, one Elephant and one Path action. The plan includes a review within seven days and an agreed measurable indicator. Candidate actions are not all simultaneous commitments.
What techniques are in the three tracks?
Nine, three per track. For the Rider: Find the Bright Spots, Script the Critical Moves and Point to the Destination. For the Elephant: Find the Feeling, Shrink the Change and Grow Your People. For the Path: Tweak the Environment, Build Habits and Rally the Herd. It builds the concrete actions per track on these nine, and picks a few per situation rather than ticking all of them off.
What does the skill never do?
The five rules are to gather the situation before advice, reserve Switch terminology for change questions, consider Rider with Elephant and Path, select three final actions rather than turning the analysis options into a long commitment list, and avoid moralising about what someone should want. These are workflow rules, not proof of why most changes fail.
When isn't this skill the right choice?
Three situations are named in which the skill shouldn't trigger. If you're already in the middle of execution and just need tactical help. If it's purely a one-off decision. And if it's about a conflict in a specific conversation, because then a conversation technique helps you more than a change plan. The Negotiator skill from the same library is meant for that last case.
Change inside, visibility outside
A change inside your business only helps if there's something to show for it on the outside. If the flow of new business is drying up, the real work isn't more urgency in the team but being found by the people looking for you. The free SEO scan shows in a few seconds where your site stands. If this kind of thinking works for you, take a look at what else is ready in the library of 100 free skills we make available with no account needed. And if you want to talk further about what else AI can do for your business, from individual skills to full automation, we're happy to do that in a conversation.