Turn Any Vague Training Request into Learning Objectives with Claude
You get a Slack message that reads: “Can you put something together on feedback for the managers? They really need it.” No follow up, no specifics, no sense of what the actual problem is. The L&D person opens a doc and immediately starts trying to guess.
I’ve been on both sides of that conversation. As a stakeholder I’ve written the message, and as the L&D person I’ve spent two hours writing back with clarifying questions only to wait three days for the answer. Both halves of that loop are slow.
There’s a Claude prompt I keep saved for this exact moment. It gets me out of the guessing phase in about thirty seconds.
What you’re building
A prompt that takes any vague training request and turns it into three to five SMART learning objectives, each tagged with the right Bloom’s Taxonomy level and one assessment idea. Save it once. Paste it whenever a request lands. The objectives become the spine of the design conversation that follows.
Why it matters
Learning objectives are where most training programs either win or fail. Activities follow from them. Assessments follow from them. The stakeholder conversation gets easier when you can say “by the end of this session managers will use the SBI framework in three workplace scenarios” instead of “we’ll cover feedback better.”
Writing good objectives is harder than people give it credit for. It takes knowing Bloom’s levels, knowing your action verbs, and knowing how to connect a behavior to something you can actually observe. Most stakeholders making requests don’t know any of that, and most L&D pros can do it but the request usually arrives at 4pm on a Thursday.
Claude is genuinely good at this. The mechanical part runs faster on the machine, which leaves you free to do the actual thinking.
The build
Open Claude at claude.ai. A free account works. Sonnet does noticeably better than the base tier on this kind of structured writing.
Paste this in:
You are an instructional designer helping me write clear, measurable
learning objectives for a training program.
Stakeholder request: [paste the exact words they used]
Please do the following:
- Write 3 to 5 SMART learning objectives based on this request.
- For each objective, identify the Bloom's Taxonomy level
(remember, understand, apply, analyze, evaluate, create).
- For each objective, suggest one measurable assessment idea
(quiz question, observation rubric, scenario exercise, etc.).
- If the request is too vague to write good objectives, ask one
clarifying question before proceeding.
Format as a numbered list. Use action verbs. Write from the
learner's perspective.
A few notes that change the output quality.
Paste the stakeholder message raw. Don’t summarize it or clean it up first. The mess is the data. If they wrote “they really need it” leave that in. Claude reads tone too.
Add context after the first pass. “These are mid level managers in a Japanese business unit. Indirect communication is the norm.” Then ask Claude to revise. That second pass is where the objectives start fitting your actual people.
Save the prompt somewhere you’ll find it again. Mine lives in a Notion doc titled “Prompts I actually use.” The whole point is that next time it takes two minutes instead of two hours.
A real example
Input pasted in: “Our managers need to be better at giving feedback.”
Claude returned:
1. Participants will identify two situations from the past month where
structured feedback would have helped and explain why it wasn't given.
(Bloom's: Understand)
Assessment: Reflection exercise submitted before session 1.
2. Participants will apply the SBI framework to write feedback responses
for three workplace scenarios.
(Bloom's: Apply)
Assessment: Scenario responses scored with a shared rubric.
3. Participants will evaluate a recorded feedback exchange and identify
what made it effective using the SBI criteria.
(Bloom's: Evaluate)
Assessment: Small group debrief with annotated transcript.
4. Participants will design a feedback plan for one direct report,
including two specific moments to use SBI in the next 30 days.
(Bloom's: Create)
Assessment: Peer review of plans in the final session.
That gives you four anchors for the next stakeholder conversation. You can hand them this and say “tell me which of these matter most to you” and suddenly you’re designing something specific instead of guessing.
Where this breaks
Claude doesn’t know your culture, your prior programs, or your internal language unless you tell it. Raw quotes help more than polished summaries.
Compliance training still needs a human subject matter expert. These objectives are starting points, not legal signoff.
Bloom’s levels are recommendations. You know your learners. Move things up or down accordingly.
What to do next
Hand Claude the finished objectives and ask it to draft a ninety minute workshop agenda mapped to each one. That’s the next thirty minutes of your day saved.
Sources:
- Anderson and Krathwohl’s revised Bloom’s Taxonomy, the version most modern instructional designers use.
- SMART criteria, Doran 1981, the original framing behind measurable objectives.
- Claude, Anthropic’s chatbot used in this tutorial.
Part of the Build with AI series on leadhuman.ai.
Jay Vergara is an L&D strategist and cross-cultural communication specialist based in Tokyo. He is a partner at Peak Potential Consulting and writes about leadership, learning, and building with AI at leadhuman.ai and on LinkedIn.
You might also like
Build a 1:1 Prep Prompt with Claude
A Claude prompt that turns a 90 second brain dump into a focused 1:1 agenda. Three questions, one action item, one human opener.
How to Use Claude to Build a Personal Knowledge System in 30 Minutes
Most people use AI like a search engine. I turned Claude into a personal knowledge management system that remembers my context, my career goals, and my writing voice. Here's the architecture behind it.
How I Built a Council of AI Advisors (And Why You Should Too)
What if instead of one AI chat for everything, you had a personal advisory council? Each specialist trained on your context, your voice, and your goals. Here's how I built mine using Claude and why it changed how I work.