Action Mapping: Cathy Moore's Model With a Worked Example

Build, launch and sell quality courses in minutes

Start Free

Action mapping, Cathy Moore's method for designing workplace training, starts where many course requests leave off: with what should change at work. Picture a hypothetical case: the operations manager of a residential solar and battery installer asks for "a course for our installers on how solar systems work, inverter types, battery basics and good customer service". Meanwhile about 1 in 3 jobs ends with the customer ringing the support line within 30 days, often because nobody connected the monitoring app or said what happens in a blackout. Nothing in the request says what should change on site, or how anyone would know it had.

I'm an instructional designer at Coursebox, where we make instructional design software that turns a brief into a course, so I'll mention our product where I tested it. In this guide I set out what action mapping is in Moore's own words, where it came from and the step she says people skip: asking why staff aren't already doing what the job needs. Then I map that solar request from a business goal down to a one-page handover card, compare action mapping with ADDIE and SAM, and say plainly what the evidence does and does not show. Along the way there are two real cases, a product test and two free templates.

What is action mapping?

Action mapping is Cathy Moore's method for designing workplace training: start from a measurable business goal, list what people must do on the job to reach it, find out why they aren't doing it, and add only the practice and information those actions need.

View the full-size diagram

Moore's own definition is short. On her overview of action mapping, she writes that "Action mapping is a streamlined process to design training in the business world" and describes it as "a mashup of performance consulting and backward design, with a focus on real-world behaviors rather than assessment questions." Its aims, in her words, are to "Commit to measurably improving the performance of the business", to "Identify the best solution to the performance problem" and, "When training is necessary", to "create realistic practice activities, not information presentations".

The name comes from the question it keeps asking. In her original 2008 post, Moore explains: "I call it action mapping because it helps you change what people do, not just what they know."

The process is often summarised as four questions, from her one-page summary of action mapping. Here they are at a glance, with what each one produces in the solar example.

Five points set the scope of the model:

• It is a needs analysis, not only a content outline. Moore is blunt about this: "Action mapping includes needs analysis; it's not just a way to organize content." The second question has two halves, and the second half decides whether any training gets built.
• It designs activities, not courses. Her action mapping FAQs say it "helps you design activities, not courses", and she has written that the activities can run in a workshop, online or in whatever format you need.
• It is for business training. Moore first presented it as an e-learning model "Because I was designing elearning at the time", but writes that "it works for all types of business training". She also writes that it "isn't intended for use in academia".
• It is done with the client and the experts. "No designer should create a design in isolation." The client owns the goal and the subject matter expert knows the job.
• The map is a working diagram, not a mind map. "Everything on the map has to justify its existence by linking directly to the business goal." Moore also warns that "It's usually not a super-simple diagram."

A Canadian project shows what changes when a designer switches from planning learning outcomes to mapping job actions, and what can still go wrong at the build.

Case study: Insurance adjudicator training in Canada

Saskatchewan Polytechnic's Center for Continuing Education worked with an organisation that administers services for Saskatchewan's health system to build four self-paced online micro-credentials for adjudicators at insurance companies across Canada. Adjudicators administer group life, extended health care, dental and disability income plans. The designer's usual method, built around learning outcomes and aligned assessments, did not suit the subject matter expert, who thought of learning as closing the gap between current and desired work performance.

• What they did: Partway through the project the designer switched to action mapping, choosing it over SAM because it fitted the design phase of their existing model and suited a two-person team. He and the expert built four action maps in weekly or fortnightly video sessions. One map held a single performance goal, six job behaviours, eight practice opportunities and thirteen information items. Its goal was "Build trusting and professional relationships with plan members", one behaviour was "Build trust with a plan member during the initial phone call", and its practice was "Have curious conversations with plan members".
• Evidence and result: In Benoit's 2025 case report in the International Journal of Advanced Corporate Learning, the designer judged the switch against three criteria he set: ease of learning, ease of resuming after breaks and focus on performance. He judged all three met. The expert learned the method in "a brief ten-minute training session", and "no further explanation was needed when developing subsequent Action Maps, even with gaps of up to two weeks between meetings."
• Limitation: It is one designer's account, judged against his own criteria, with no data on learners, adjudicator performance or any business measure. The goal quoted has no measure or date. The paper's four steps are the author's summary and leave out the question of why people aren't already doing each behaviour. The built lesson also drifted from the map: in its practice step, "Learners are not provided with an opportunity to practice with one another" and answer an agree or disagree question about attitudes instead.
• Lesson: In this case, the expert learned the mapping method quickly and the designer found it easier to keep the work focused on performance. A map does not build the course, though. Check the finished lesson against the map, action by action, because practice can turn back into a question about attitudes or facts during the build.

Where action mapping came from

Moore started out in the early 1980s as a technical trainer and moved into corporate e-learning in 2001. In her account of how action mapping began, she credits marketing for the core idea: "good marketers create business goals to justify the expense of a project, and they measure (in a million ways!) whether people actually change their behavior." In an excerpt from her book she says she developed it "because I felt we needed a structured way to combat the school model."

The model has changed since 2008, and the dates help when you read older material about it:

• May 2008: Moore publishes the original post, which she now labels "the original blog post about action mapping, published in May 2008." The steps were in an embedded slideshow, not in the text.
• May 2013: she adds a flowchart to the second step. In her words, "I deepened step 2 to include the use of a flowchart that helps identify lean, in-the-workflow solutions, such as job aids or process improvements."
• 2014: her keynote handout for Training magazine's Online Learning Conference becomes the one-page summary often taught as "the four steps".
• 2017: she publishes Map It: The hands-on guide to strategic training design under her own imprint, Montesa Press, in 418 pages. I found only one edition.
• Now: her action mapping workflow at a glance runs to 15 steps, from writing the goal to releasing and evaluating the activities.

The dates explain one correction. A version of action mapping that runs goal, actions, practice and information as four content-design steps leaves out the barrier check that Moore built into step 2. She added the flowchart because "People were skipping this aspect of step 2". The four questions are her summary; the 15-step workflow is the process she teaches now.

Which of Moore's sources to read

• Read this one first: Map It (2017). It is Moore's full treatment of the process, with worked stories such as the hospital example in its third chapter.
• If you have ten minutes: the action mapping workflow at a glance. It shows the current 15 steps on one page, including where training drops out.
• Before your first client meeting: her flowchart post, Is training really the answer? Ask the flowchart, which explains the barrier check that the rest of the map depends on.

A worked example: a solar installer's customer handover

Here is that example in full, which I'll map from here on. Its figures are illustrative. A residential solar and battery installation company in an Australian city runs about 12 install crews, each led by an installation lead. Leads are licensed and experienced at installing, and some were promoted recently. Being licensed to install a system does not mean someone is confident explaining it to a householder, so the course cannot assume that skill.

The operations manager has asked for a course on how solar systems work, inverter types, battery basics and customer service. The company already measures a problem the request does not mention. About 1 in 3 jobs produces a call to the support line within 30 days of installation. The support log shows four main reasons:

• "My app shows nothing": the monitoring app was never connected to the home Wi-Fi.
• "Is my system working?": the customer was never shown how to read the inverter or the app.
• "My power went out in a blackout and the solar didn't help": nobody said that a system without a battery switches off in a blackout.
• "I haven't got my paperwork": the commissioning sheet and photos were not uploaded, so the office cannot finish the paperwork.

The intended outcome is that installation leads complete the customer handover before they leave site, and that fewer customers need to call. The course does not teach wiring, electrical safety or standards, and it is not a licence or a certification.

This example suits action mapping for three reasons. The request arrives as content, the company already tracks a measure the handover drives, and some of the causes are plainly not about training. Those are the conditions in which the barrier check has real work to do.

Action mapping steps, worked through one course

Moore's one-page summary asks four questions, and the second has two halves: what people need to do, and why they aren't doing it. I give each half its own heading, so the five headings that follow are her four questions with the second one split in two. Each gets the same three parts: what it looks like in the solar course, how to check it, and a common slip.

1. Set a measurable business goal

The goal is the result the organisation needs, written so that everyone can see whether it happened. Moore's workflow gives a pattern: "A measure we already use will increase/decrease X% by date as people in a specific group DO something". Her one-page summary adds what the measure will not be: "make sure all employees complete the training". As she puts it, "No one but L&D cares about that."

• In the solar course: "Customer calls to the support line in the first 30 days after installation fall from about 1 in 3 jobs to 1 in 6 by 30 June, as installation leads complete the customer handover before leaving site."
• How to check it: the measure already exists in the support log, it has a baseline and a date, and it names who has to act. If the operations manager could not report it in July, it is not yet a goal.
• Common slip: writing a learning goal, such as "installers will understand solar systems", or a training measure, such as completion. Neither tells the business whether anything changed.

A goal like this is the kind of result Kirkpatrick puts at level 4, which our guide to the Kirkpatrick model covers in full. Moore herself calls putting the goal first a public commitment "to improving our organization's performance". Writing it does not claim the course will reach it alone. In her post on "awareness" requests, she writes that "By identifying a measurable change, you aren't claiming that your project will single-handedly achieve this change."

2. List what people need to do on the job

Actions are the observable things people must do to reach the goal. Moore's test is that they are "actions that a guy with a clipboard could observe and check off", and she warns against "verbs that take place only on a test", such as define or identify. In her FAQs she notes that "each behavior or action that you write could be called a performance objective."

• In the solar course: four actions, all at the end of an install. Connect the monitoring app to the home Wi-Fi and show the customer today's generation on their own phone. Show the customer the shutdown procedure using the shutdown label at the inverter and switchboard, then have them point to each switch in order. Explain what the system does in a blackout and what to expect on the first bill, without promising a dollar saving. Upload the commissioning sheet and photos in the job app before leaving.
• How to check it: could the operations manager stand on site and tick each one off? "Explains the system well" fails that test; "has the customer point to each switch in order" passes.
• Common slip: listing results or goals as actions. Moore's example from a reader's map was "minimize the use of premium labor", which she said "sounds more like a goal than a discrete, observable action."

The safety content in action 2 is the company's own procedure and label. The course tells leads to follow the label and the manufacturer's procedure on site; it does not teach electrical rules.

3. Ask why they aren't doing it

This is the step that decides whether to build training at all. For each action, Moore's web flowchart, Will training help?, asks "Why aren't they doing that ONE thing?" and sends you down four branches: "Environment (culture and tools)", "Knowledge", "Skills" and "Motivation". Her instruction is to "Try the environment branch first." On the one-page summary she is more direct: "Don't skip the flowchart. If you assume that every action requires training, I guarantee you'll waste time designing stuff you don't need."

• In the solar course: talking to two leads and the support team turns up the causes in this table. Two of the four actions need no training at all.
• How to check it: every action has a named cause, and every cause has a fix that matches it. A scheduling problem gets a scheduling fix, not a module.
• Common slip: assuming every gap is a knowledge gap. Moore argues that "Few performance problems are caused purely by lack of knowledge or skills" and that "Training is rarely the best or only solution to a problem." Those are her arguments from experience, not findings, but they are the reason the flowchart exists.

Moore's own teaching story for this step is fictional. In an excerpt from the third chapter of Map It, a designer finds that hospital staff recap needles "because the sharps container isn't handy", and the fix is to mount containers beside every bed. Moore notes that the story "was super-simplified". Two real hospitals tried that exact fix.

That is the San Francisco result reported by Makofsky and Cone in 1993: a real but modest drop. In West Virginia, Edmond and colleagues found that observed nurses recapped 93.9% of needles before bedside disposal units and education, and 94% after, with "no significant difference". Neither study was an action-mapping project. They show something useful about the barrier check, though. It gives you a hypothesis about the fix, and only the measure tells you whether the fix worked. In the solar course, that means watching the 30-day call reasons for "my app shows nothing" after the handover slot is booked, rather than assuming the scheduling change did its job.

If you need a fuller method for sizing the gap between current and expected performance, our guide to the performance gap covers it. If the problem is not yet chosen and you need to survey a whole team first, the training needs analysis questionnaire is the place to start.

4. Design realistic practice activities

For each action that needs training, design an activity in which learners do the thing, or decide as if they were doing it. Moore's summary sets two conditions: a character, often "you", "faces a specific, realistic challenge", and "The feedback shows the consequences of the learner's choice." Her workflow gives a contrast. The activity is "A customer wants to use their widget at high altitudes. You have to ask them questions to identify the best widget." It is not "At which altitude should a customer use the AltaWidget?"

• In the solar course: three activities, two for action 3 and one for action 2. In the first, an older homeowner asks during the walk-through, "So if the power goes off, I'm still fine, right?" The system has no battery. In the second, the customer asks, "How much will this knock off my bill?" In the third, the learner puts the steps on a photo of a shutdown label in order, then chooses how to check the customer understood.
• How to check it: each activity opens with the customer's question or the task, the learner answers in the role, and every option shows what happens next before any explanation.
• Common slip: presenting the information first and quizzing it afterwards. Moore's advice is to "Resist the temptation to present and then quiz."

Here is the first activity in full. After each consequence, a short note from a senior installation lead explains why the plain answer works.

These are scenario-based activities, and our guide to scenario-based learning covers how to design them, including the case data learners need and how much guidance to give. For choices whose effects build on each other, the branching scenario examples show how to map the branches.

5. Add only the information the practice needs

The last layer is the information. Moore's instruction is to "identify the information that people absolutely must have for each activity. Also decide whether the info needs to be memorized or can be looked up in a job aid." In her workflow, learners "pull information as they need it during the activity", and she adds: "Don't assume everyone is equally ignorant."

• In the solar course: a one-page handover card with the blackout answer for systems with and without a battery, the company's rule never to quote a dollar saving, and the support line number. The site's shutdown label. There is no section on how solar works, inverter types, battery basics or customer service theory.
• How to check it: for every piece of content, name the activity that needs it. If you cannot, cut it.
• Common slip: keeping the original brief's topics "for context". The request listed how solar works, inverter types, battery basics and customer service; no practice activity needs a section on any of them.

Because the card is used on the job, during the conversation, it is a job aid rather than course content. Our guide to the five moments of need helps decide which kind of support fits the moment a person needs it, which is often the decision this step leaves you with.

How the parts of an action map relate

An action map is not a sequence of four lessons. It is a diagram with the goal at the centre, the actions linked to it, the practice activities linked to the actions they build, and the information linked to the activities that need it. Moore's rule is that every item has to link to the business goal. An item with no line back to the goal does not belong on the map.

Three features of that structure shape how you use it:

• The barrier check sits between the actions and everything else. Each action goes through it on its own, so one map can send some actions to training and others to a schedule, a tool or a job aid. In the solar map, actions 1 and 4 stop at the check.
• Practice is the centre of any training that remains. Information is attached to practice, never the other way round, which is why the map ends with a card rather than a module.
• The order is a design order, not a learner's order. You write the goal first and the information last, but learners may meet the handover card before the first scenario. Moore's workflow also loops: you prototype one activity, get it approved and revise before producing the rest.

Seen this way, the map is also a record of decisions. A client who later asks "where is the module on inverter types?" can see the line that was never drawn: no action needed it.

How to run an action mapping project

Moore's workflow at a glance lists 15 steps. Grouped for a project plan, with her step numbers in brackets, they run like this:

1. Write the goal with the client (step 1). Moore's estimate from experience is that "Often, one two-hour meeting can cover the first two steps of action mapping." Bring the subject matter expert.
2. List the actions and choose the important ones (step 2). Use observable verbs, then pick the actions with the most effect on the goal.
3. Ask what makes each action difficult, and brainstorm every change that would help (steps 3 and 4). Her prompt is "Brainstorm all changes, not just training", with examples such as "Streamline the procedure" and "Create a how-to guide (job aid)".
4. Decide whether training is part of the solution (step 5). If practice will not help, "You'll just make the changes identified in step 4."
5. Brainstorm activity ideas, then prototype one and get it approved (steps 6 to 8). One finished activity shows the client what the rest will look like before you write them.
6. Outline all the solutions, get the outline approved and produce the activities in batches (steps 9 to 13). The outline includes the non-training fixes, not only the training.
7. Release the activities (step 14). Moore's note on this step is "Not necessarily as a course!" Activities can go into a workshop, a team meeting or a short online set.
8. Evaluate and improve (step 15). "Once all solutions are in place, look for progress toward your goal." She suggests "Robert Brinkerhoff's Success Case Method to identify what's working and what isn't", which our guide to Brinkerhoff's success case method explains.

The action map for the solar handover course

Here is the finished map in table form. The two non-training actions stay on it, because they are two of the four actions the goal depends on.

The measure is the support log's 30-day call rate per lead, monthly. For the actions themselves, the operations manager observes or reviews the job app for each lead's next five handovers. In Kirkpatrick's terms those observations are level 3 and the call rate is level 4.

A national health service team used a full day of action mapping to plan three e-learning programmes. What it could report afterwards shows why the measure has to be written at the start.

Case study: Information skills e-learning for health staff in England

The national NHS England Knowledge and Library Services team built three e-learning programmes for the health workforce in England, on literature searching, critical appraisal and knowledge mobilisation, hosted on the national e-learning platform for health and care staff.

• What they did: "The design process was led by the learning technologist and started with a full day of action mapping (Moore, 2024) and creating audience personas to inform the overall programme and each module." The map then informed the assessment, the questions, the scripts and the marketing, and four modules used video scenarios to show the techniques in action.
• Evidence and result: In the team's 2025 report in Health Information and Libraries Journal, "Between April 2023 and July 2024, 24,029 E-Learning sessions have been launched." Users who completed a module could rate it from 1 to 5, and "all the courses have received a majority rating of 4 or above." Some users reported trouble getting the e-learning to work on different browsers and devices.
• Limitation: Launches count sessions, not people, and only those who completed a module rated it. The report gives no measure of how well staff searched or appraised evidence afterwards, no business goal of the kind a map centres on, and no comparison group.
• Lesson: A mapping day can organise a national programme's design, assessment and scenarios, and the team still ended up reporting launches and star ratings. Write the measure you will check before the design day ends, using data the organisation already collects, or the map's goal has nothing to report against.

What Coursebox course requests leave out

To see how training requests arrive, I looked at course requests typed into Coursebox in the 90 days to 7 October 2026: one per person, from English-language browsers, with no uploaded file. Of 266 prompts, 104 were course requests, and 25 of those were for people in a work role. Two independent AI readers each classified all 104 against written definitions built from the layers of an action map, and I re-read every request either reader flagged for a goal, an action, practice, limited information or a non-training fix.

Only one of the 25 named a result the organisation needed: bringing a division to "a successful 1st stage audit in January 2027". It was also the only workplace request that asked for a checklist next to the course.

Between 17 and 19 of the 25 workplace requests described only topics, modules or content to cover, the lower number where both readers agreed. None of the 104 requests asked whether training was the right fix, or mentioned a cause training cannot fix, such as tools, workload or how people are paid.

This is a small sample of what people typed, not of the courses they finished, and the person typing may not be the one who holds the business goal. A brief is a request for content by design, so these counts do not mean anyone is doing their job badly. They do show where requests start: at the third and fourth layers of the map, with the goal and actions missing. A builder handed a topic list, whether a person or an AI, will build a content course. That is why an action map starts with a goal sentence, agreed with the client before anyone writes a topic list.

Action mapping vs ADDIE, SAM and criterion-referenced instruction

Action mapping is mostly an analysis and design method, so it sits inside a project process rather than replacing one. It shares its "is training the answer?" question with older performance-first models.

Moore's own comparison with ADDIE is easy to misread. In the Map It chapter excerpt, her fictional designer Tina follows a version of ADDIE in which "Analyze" means "The client says they need training; therefore, they need training", and "Evaluate" means "Ask the participants if they liked the training." Moore's comment is that Tina "didn't follow the model as it was intended to be used, but her interpretation is common." Her second designer, Anna, starts with "Evaluate: Identify how the success of the project will be measured." The contrast is between evaluation first and evaluation last in a common use of ADDIE, not between action mapping and ADDIE as designed. Our guide to the five steps of ADDIE explains how evaluation can feed every phase.

The practical difference shows in one line from the same chapter: "Tina: In a course, I mostly design information. Anna: In a course, I mostly design activities."

With SAM, the fit is close: Moore's workflow prototypes one activity and gets it approved before producing the rest, which is how SAM works too. Criterion-referenced instruction, from Robert Mager and Peter Pipe, also asks whether people already meet the standard before training them. Our overview of instructional design models places all four side by side.

When to use action mapping, and when not to

Use action mapping when there is a performance problem the organisation can measure, and people currently doing the job whom you can watch and talk to. Moore's FAQs set those conditions: you need "a performance problem (not a learning goal like certification)" and "a specific audience comprised of working adults who need to perform specific on-the-job tasks". Her overview adds that the method "assumes that you have people currently doing the job who you can talk to and observe."

Compliance training is the hardest fit, and Moore addresses it herself. In her post on mandatory training she calls the usual approach "sheep-dip training" and writes that "If there's really no problem, we shouldn't create a solution." She also accepts that the decision is sometimes out of the designer's hands: "If you're required by forces beyond your control to create something for all employees, you can at least break down the audience by major job roles". Then design activities in which people correctly apply the rules to decisions in their own role, with the rule shown in the feedback. The business goal may be weaker or a stand-in, but the map still cuts content that no role needs.

The academic limit is real, and an Australian reviewer has named it. Reviewing Map It for the Australian Journal of Adult Learning in 2023, a senior learning designer at RMIT's vocational college wrote that "the consistent corporate examples and business terminology risk excluding wider audiences", and that the book's questioning of whether a course needs to be made at all "could create an illusion of separation from academic settings that don't afford such a challenge." That matches Moore's own view that the method was not designed for education. It does not mean the practice-first layer is useless there, and some university designers have adapted it for course redesign.

Does action mapping work? What the evidence shows

I could find no controlled study of action mapping, and Moore says she knows of no research into the model. Her FAQs compare the question to asking whether ADDIE works, and continue: "I don't know of any research into the model and don't see how such research could be rigorously designed." She reports "lots of anecdotal reports of success", and adds that she has "also seen the process misapplied." The advantages on her overview page are introduced as what "Happy action mappers have told me".

I searched ERIC, PubMed, Crossref and OpenAlex for published uses of Moore's model. The uses I found are descriptive: the Canadian case judged by its designer, the NHS England report of launches and ratings, two papers on university course redesign and one conference abstract from Brazil. None measured the business goal its map was built around.

The Brazilian abstract has the biggest numbers, and it shows a common trap. The Open University of the Brazilian Unified Health System (UNA-SUS) ran five microlearning courses on the care of older people for primary care workers, and the abstract states that "It follows “action mapping” design". Its stated aim was to improve completion rates.

In the conference abstract by Savassi and colleagues, which is the whole publication, the authors conclude that the microlearning strategy promoted higher completion rates than traditional online distance education courses, but give no comparison figure. The courses combined action mapping with microlearning, so the result cannot be credited to either alone. And because the goal was completion, completion is all the programme could report. Moore's one-page summary warns against exactly that measure.

Practice activities have their own evidence base, which our scenario-based learning guide sets out. That is evidence for one layer of the map, not for action mapping as a whole. What a stakeholder can be promised is a measure you will check and a date you will check it, not a result.

Limitations of action mapping

• No research tests the model. Moore knows of none, and the published uses I found report usage, ratings, completion or the designer's own judgement.
• It needs a performance problem and access to the job. Without a measure the organisation cares about, or people you can observe and talk to, the first two layers are guesswork.
• Some goals resist a single measure. Moore acknowledged in replies to readers that for a broad rollout across units "it's hard to set one measurable goal", and that soft skills "can be hard to quantify".
• Compliance and certification weaken the goal. When the requirement is exposure or a knowledge test, the map can still cut content by role, but the business goal may be a stand-in.
• It was not designed for academic settings. Moore says so, and the AJAL review points out that many academic and vocational designers cannot decline to build a course.
• It takes the client's and the expert's time. Moore's two-hour estimate for the heart of the map is her experience, not a benchmark, and she says no designer should map alone.
• A map does not guarantee the build. In the Canadian case, mapped practice became an agree or disagree question in the finished lesson, and an AI builder can drift in the same way.
• Measuring the goal takes data and time. The published projects reported launches, ratings or completion instead of the business measure, which is the easier path when nobody owns the data.

Using action mapping with an AI course builder

The course-request data shows that briefs usually arrive as topic lists. So I tested the opposite: a brief written from a finished action map, to see whether Coursebox would keep the practice first and leave the non-training fixes out. Coursebox is our product, so weigh my account with that in mind.

On 12 October 2026 I briefed Coursebox to build "The customer handover" for the solar installer. The brief gave the business goal for context only, the four actions with actions 1 and 4 marked "NOT TRAINING", the three practice activities in order with their options, a consequence for each blackout reply and the result the teach-back should show, the handover card's facts word for word and a placeholder support number. It asked for a short "Before you start", the two scenarios, the shutdown teach-back, an open-text debrief and a wrap-up as the very last section. It also said not to explain solar, inverters or batteries, and not to add recall questions.

The content dump I expected did not happen. The first outline had no section on how solar works, and the two non-training actions were not taught. It did merge the savings and shutdown activities and add a quiz, so I sent one revision asking for six sections in the briefed order, consequences before the senior lead's note and no quiz. The builder asked once whether the six sections should replace its own introduction and summary, and I said yes. The second outline had the six sections in order and no quizzes. On the rendered pages, the opening page, "Four handover actions", listed all four, marked the two the course practises and labelled actions 1 and 4 "Process change / Not training content" in a table. The wrap-up named the two process changes and showed the card with the placeholder number unchanged, with no savings figure and no standards. Both customer conversations were scenario blocks that asked "What do you say?".

The issue was on the blackout page. When I chose "Yes, the panels will keep you going.", the scenario replied with a verdict, "That sounds reassuring, but it is misleading", instead of the consequence in my brief, and moved to a new question. Under the block, visible while a learner is still choosing, the page text described the right answer and gave no consequences. So the page explained before it showed what happens, which is the opposite of a practice-first design. The brief asked for consequences followed by an explanation. Those consequences were missing, so this was the part to fix.

The scenario block's own replies cannot be edited, only regenerated, so I fixed the text under it in the page editor. I replaced it with "After you answer: what happens next with each reply", followed by the three consequences from the map and then the senior installation lead's note on why the plain answer works. I reloaded the page in the learner view and checked that the consequences and the note rendered and that the old text naming the right answer was gone. The text still sits on the same page as the choice, so a learner can read ahead.

I checked the outline and three rendered pages and ran one scenario; I did not run the course with installation leads. A clean outline shows the requested structure was followed. It does not show that support calls fall, which is the goal the course was mapped from, and which only the support log can show.

Putting action mapping to work

Action mapping answers the question many training requests skip: what should people do differently at work, and how will the organisation know? Its four layers are a goal, the actions that reach it, practice for the actions that need it and only the information that practice uses. The part that does the most work is the check between the actions and the practice, where each action is asked why people aren't already doing it. In the solar example, that check sent half the actions to a schedule and an app setting before a single page was written.

Use it when a request arrives as a list of topics and a measurable problem sits behind it. Adapt it for compliance by mapping job roles, and reach for a different model when the requirement is a knowledge test or an academic course you cannot decline. Do not promise that the course will move the goal. Promise the measure, the date you will check it and the fixes that are not training.

Your next step: take the next course request on your desk, write one goal sentence in Moore's pattern with a measure the organisation already uses, and ask the client which on-the-job actions would move it before you agree to any topic.