Prompts for running training and course-build projects

Turn a one-paragraph training request into a one-page charter you can actually take to a sponsor.

You are an experienced project manager who also understands instructional design. I'm starting a training/course-build project. Here's what I know so far: [paste the request, business problem, or intake form].

Draft a one-page project charter that includes:
1. A problem statement written in business terms, not training terms
2. 3-5 measurable learning and business objectives
3. In-scope and explicitly out-of-scope items
4. Key stakeholders and their role (sponsor, SME, approver, learner rep)
5. High-level milestones mapped to ADDIE or SAM phases
6. Top 3 risks and an assumption list

Flag anywhere the input is too vague to charter responsibly, and list the clarifying questions I should ask before this goes to a sponsor.

Works well with: Claude, ChatGPT

Generate a starter RAID log built for training projects, not a generic software-project template.

Act as a project manager who has run multiple training/curriculum rollouts. Given this project summary: [paste scope, timeline, audience size, delivery modality (ILT/vILT/eLearning/blended), and any known constraints], generate a RAID log with 5-8 entries per category specific to training projects.

Think specifically about: SME availability, content approval cycles, LMS/tooling readiness, translation or localization needs, facilitator certification timelines, and learner release time from their managers.

For each entry give: description, likelihood/impact (for risks), an owner, and a suggested mitigation or tracking action.

Works well with: Claude, ChatGPT

Convert a vague stakeholder request or SME interview into measurable, Bloom's-aligned objectives.

I'm going to paste a raw stakeholder request or SME interview notes for a training program: [paste text].

Extract and draft 4-6 terminal learning objectives written with measurable verbs aligned to Bloom's taxonomy (specify the cognitive level for each). For each objective, note:
- The business outcome it supports
- The level of the audience it assumes (novice/intermediate/advanced)
- One assessment method that would prove mastery

Call out any objectives that are really just "topics" restated, and rewrite them to be observable and measurable.

Works well with: Claude, ChatGPT

Turn messy project notes into an exec-ready status update, in the tone execs actually want.

Write a steering committee status update for a training/learning program based on these raw notes: [paste sprint notes, standup notes, or a task tracker export].

Structure it as:
- Overall RAG status with a one-line justification
- 3 key accomplishments since the last update
- 2-3 upcoming milestones with dates
- Risks/blockers that need an executive decision (be specific about what decision is needed and by when)
- Budget/resourcing status, if mentioned in the notes

Keep it to one page, written for people who don't care about instructional design methodology — they care about dates, cost, and risk.

Works well with: Claude, ChatGPT

Generate role-specific TNA interview guides that separate real skill gaps from process or motivation problems.

I'm conducting a training needs analysis for [describe the initiative — new system rollout, compliance requirement, performance gap, etc.].

Generate three separate interview guides — one for executive sponsors, one for people managers of the target audience, and one for a sample of the target audience/learners themselves.

Each guide should have 8-10 questions that distinguish a genuine skill/knowledge gap from a motivation, process, or environmental barrier (use Gilbert's Behavior Engineering Model or a similar performance-analysis framework as the lens), plus 2 questions that surface how success will be measured after training.

Works well with: Claude, ChatGPT

Generate assessment items that actually measure your stated objectives instead of testing trivia.

Here are my terminal learning objectives: [paste objectives with Bloom's levels noted].

For each objective, write 2 assessment items — choose the format that matches the cognitive level (multiple choice, scenario-based, or performance-based) rather than defaulting to multiple choice. For each item, include:
- The correct answer
- One plausible-but-wrong distractor, explained (why a learner might pick it)
- Whether this item belongs in a knowledge check, a formative assessment, or a summative/certification assessment

Works well with: Claude, ChatGPT

Turn a high-level course outline into a scene-by-scene storyboard skeleton to hand to a developer.

Convert this course outline into a storyboard skeleton: [paste module/topic outline].

For each screen or scene, include:
- A working title
- The on-screen text or narration summary
- The visual treatment (talking head, animation, screen capture, static graphic, interaction)
- The interaction type if any (click-to-reveal, drag-and-drop, branching scenario, knowledge check)
- An estimated runtime

Flag any section of the outline that's too dense for one screen and should be split, and any section too thin to justify its own screen.

Works well with: Claude, ChatGPT

Generate a RACI chart specific to instructional design roles instead of a generic PM template.

Build a RACI chart for a training/course-build project with these roles involved: [list roles present — e.g., project sponsor, SME, instructional designer, project manager, e-learning developer, graphic designer, voice talent, LMS admin, QA reviewer, legal/compliance reviewer].

Map RACI across these deliverables/phases: needs analysis, design document/storyboard approval, content development, media production, QA/UAT, LMS upload and configuration, pilot delivery, launch, and post-launch evaluation.

Flag any deliverable where more than one role is marked Accountable — that's usually a sign of an unclear decision-maker.

Works well with: Claude, ChatGPT

Build a power/interest stakeholder map tuned to the specific politics of training programs.

Help me build a stakeholder map for this training initiative: [describe the initiative and who you know is involved].

Place each stakeholder on a power/interest grid and explain the placement in one sentence. Then, for the high-power stakeholders, suggest a specific engagement approach appropriate for training projects — for example, how to get an executive sponsor to actively promote the training instead of just approving the budget, or how to handle a people-manager who resents losing staff time to training.

Works well with: Claude, ChatGPT

Design a Kirkpatrick-based evaluation plan tied to real business data, not just a smile sheet.

Design a training evaluation plan using the Kirkpatrick model for this program: [describe the program and its business objective].

For each level (Reaction, Learning, Behavior, Results), specify:
- What exactly will be measured
- The data collection method and timing
- Who owns collecting it
- Which existing business or HR data source could supply Level 3/4 data without building a new survey

Be explicit if a level can't be credibly measured given the constraints I've described, rather than forcing a metric that won't hold up.

Works well with: Claude, ChatGPT

Set expectations with a new subject matter expert before the review cycle starts, instead of discovering the gaps mid-project.

I'm about to onboard a subject matter expert (SME) onto a training project: [describe the project, the SME's role/background, and how much time they've been allocated per week or per phase].

Draft a one-page SME onboarding brief that sets expectations up front, including:
1. What we need from them at each phase (content review, technical accuracy check, scenario validation, etc.) and roughly how much time each ask will take
2. The format we need feedback in (tracked changes, a review form, a live workshop) and why that format, so they're not surprised later
3. Turnaround-time expectations for reviews, and what happens if a deadline slips
4. A short list of what they do NOT need to worry about (instructional design decisions, LMS mechanics, visual design) so they don't over-scope their own involvement

Flag anywhere the allocated time seems unrealistic for the asks listed, based on typical SME availability for a project like this.

Works well with: Claude, ChatGPT

Build an adoption-focused communication plan so the training doesn't launch to an empty room, grounded in a real change model instead of a generic comms calendar.

I'm rolling out mandatory or high-priority training to this audience: [describe audience size, why the training exists, and any known resistance — e.g., "these are experienced staff who see this as a step backward" or "managers are protective of their team's time"].

Using ADKAR (Awareness, Desire, Knowledge, Ability, Reinforcement) as the structure, build a change communication plan that includes:
1. A specific message and channel for each ADKAR stage, timed relative to launch (e.g., "3 weeks out: awareness email from the sponsor, not from L&D")
2. A distinct message for people managers, separate from the message for learners — managers need to know what to say when someone asks for release time
3. One reinforcement touchpoint scheduled 30-60 days after training ends, since most comms plans stop at launch and adoption dies right after
4. Where in this plan a people-manager is most likely to quietly undermine adoption, and a specific tactic to prevent it

Keep every message short enough to survive being forwarded without a summary.

Works well with: Claude, ChatGPT

Turn raw pilot feedback into a decision memo a sponsor can act on, instead of a summary that just restates the survey.

I ran a pilot of a training program and collected feedback: [paste raw survey results, facilitator observation notes, and/or completion/assessment data from the pilot cohort].

Write a go/no-go memo for the sponsor, structured as:
1. A one-line recommendation up front (launch as-is / launch with specific fixes / delay and re-pilot) — don't bury this at the end
2. The 3-5 findings that actually drove that recommendation, each tied to specific evidence from the data, not general impressions
3. For any "launch with fixes" items: what the fix is, who owns it, and whether it blocks the launch date or can happen post-launch
4. What the pilot data could NOT tell us (sample size limits, self-selection bias in who volunteered, anything the pilot audience doesn't represent about the full rollout audience) so the sponsor isn't over-trusting a small sample

If the feedback is too thin or too mixed to support a clear recommendation, say that explicitly instead of forcing a decision the data doesn't support.

Works well with: Claude, ChatGPT

Decide whether an aging course should be retired, refreshed, or left alone, using usage data instead of whoever complains loudest.

I need to decide what to do with an existing course: [paste what you have — LMS completion trends over time, last content update date, assessment pass rates, any recent complaints or manager feedback, and whether the underlying content/process it teaches has changed].

Recommend one of three outcomes — retire, refresh, or maintain as-is — and structure the memo as:
1. The recommendation and the single strongest piece of evidence behind it
2. A short table of the signals considered (usage trend, content staleness, pass rate, business relevance) and what each one indicates on its own
3. If "refresh": what specifically needs updating and a rough effort estimate (light edit vs. a partial rebuild vs. treat as a new project)
4. If "retire": what should replace it, if anything, and who needs to be notified before it disappears from the catalog

Be explicit about which signal is doing most of the work in this recommendation, so the decision doesn't look more data-driven than it actually is.

Works well with: Claude, ChatGPT

Generate a WCAG-aware review checklist tailored to your actual content and interaction types, before it goes to development.

Here's my storyboard or course outline: [paste it, including interaction types used — click-to-reveal, drag-and-drop, video, branching scenario, etc.].

Generate an accessibility review checklist specific to this content, organized by interaction type, covering at minimum:
- Alt text requirements for any images, icons, or graphics described
- Caption/transcript requirements for any video or audio
- Keyboard-navigability concerns for each interactive element (don't just say "make it keyboard accessible" — name the specific interaction and what to check)
- Color-contrast or color-only-meaning issues implied by the described visual treatment
- Timing issues (auto-advancing content, timed quizzes) that need a pause/extend option

For each item, note whether it's something the e-learning developer can fix directly or something that requires a design change first — that's usually the difference between a QA fix and a rework.

Works well with: Claude, ChatGPT