Stop Prompting.
The simplest (free) AI upgrade for HR and managers.
Welcome to FullStack HR, and an extra welcome to the 102 people who have signed up since last week.
If you haven’t yet subscribed, join the 11500+ smart, curious HR folks by subscribing here:
Understanding Claude’s potential and limits in HR.
On 15 July I’m joining Kinfolk for a conversation about the current stage of AI Adoption and where teams are building internally vs evaluating external solutions at the moment.
Most HR and People Operations leaders I work with are deep into the pilot stage now. The enthusiasm is real but many organisations are still working on what a sustainable AI setup looks like for an HR team, outside of individual productivity.
Building with Claude for individual work makes sense. Drafting policies, summarising documents, writing job descriptions. But when teams try to extend that into their operational layer, and start working on the employee lifecycle requests, the data flows, the governance, the accountability, the gap between “this works for me” and “this works for my team at scale” becomes clear very fast.
We’ll be working through the build vs. buy decision with the infrastructure context it deserves and understanding what are good use cases for internal builds with tools like Claude (or Copilot) and where it’s worth exploring a third-party solution.
Wednesday 15 July · 5pm BST / 9am PT / 12pm ET
Happy Saturday,
You open up ChatGPT, Copilot or Claude, the cursor blinks, and you freeze. Not because you cannot type. Because youdo not know what to type. So they write something thin like “write feedback to an employee who misses deadlines,” get back something generic and a bit corporate, and quietly conclude that AI is overrated.
It is not overrated. The starting habit is just wrong. And after a few hundred workshops, this is the single fastest fix I know.
Let’s get to it.
Most prompt advice is too complicated
If you search for how to prompt, you drown in frameworks. Personas, delimiters, role assignments, “act as a world-class expert in,” temperature settings, twelve-part mega-prompts. (I know, I know, I’m guilty of contributing to it all…)
HR people and managers do not need to be mega advanced in prompt engineering. (Never liked the term.) You have a job to do and four minutes between meetings. You need a better starting move, not a certification.
The good news is that the highest-leverage upgrade is also the simplest, and you can teach it to an entire team in about ninety seconds.
The problem is (usually) missing context
When you write to the model as per above, it’s usually vague or way too general. But the model is not vague because it is dumb. It is vague because you were vague.
When you ask AI to write feedback about missed deadlines, it has no idea whether this is a first slip or a six-month pattern, whether the person is overloaded or careless, whether you want warm or firm, or what the deadline miss cost. So it does the only thing it can, it writes a safe, average, slightly soulless draft that fits every situation and therefore fits yours badly.
You did not give it enough to work with, so it guessed. The fix is not a cleverer prompt, the fix is to stop asking AI to answer before it has the context, and instead make it pull the context out of you first.
The 3-question prompt
This is the whole trick. Paste this on top of almost anything.
Before you answer, ask me three clarifying questions that would help you
give a better, more practical answer.That is it. You flip the model from answering to interviewing. It asks you three sharp questions, you spend thirty seconds answering them, and the output is on a completely different level, because now it is built on your real situation instead of a generic one. If you want to be even more proficient, I can highly recommend using a tool such as Wispr, which lets you talk to your model(s) super-easily.
Save it as a text snippet, a skill, a keyboard shortcut, a pinned note. Whatever means you will reuse it. The power is not in the words, it is in the habit.
Five examples from the HR and manager day
Let me show it on five things you probably do in the upcomoing weeks.
1. Manager feedback. You have someone slipping on deadlines, and you want to say something useful, not just annoyed.
I need to write feedback to an employee who is missing deadlines.
Before you draft it, ask me three clarifying questions.
The model might come back with, is this a repeated pattern or a recent thing, what has it cost the team or the customer, and do you want the tone supportive, direct or formal. You answer in a sentence each, and now the draft sounds like a real manager who knows the situation rather than a policy handbook.
2. Job ad. You have a tired old ad and want it to actually attract the right people.
Help me improve this job ad. Before you rewrite it, ask me three
questions about the role, the audience, and what we want to signal.
Now it asks who you are really trying to reach and what would make a strong candidate choose you, instead of just polishing the same corporate filler.
3. Onboarding. A new hire starts Monday and you want their first week to not be chaos.
Help me plan a first week for a new hire. Before you build the plan,
ask me three questions about the role, the team, and what a great
first week would look like for us.
The questions force you to actually decide what good looks like, which is most of the work anyway.
4. Policy communication. You have to explain a new policy to managers without it landing as legal mush.
Help me explain this policy to managers in plain English. Before you
draft the explanation, ask me three questions about the audience, the
sensitivity, and the behavior we want to see change.
It stops guessing the reading level and the politics, and you get something managers will actually read and act on.
5. Meeting prep. You have a tense topic and thirty minutes to handle it well.
Help me plan a 30-minute team discussion about workload. Before you
suggest an agenda, ask me three questions.
Now it asks what outcome you need and what tension is already in the room, and the agenda is built for your team rather than a textbook team.
Five examples, one habit. Notice you did not learn five prompts. You learned one move and pointed it at five tasks.
When not to use it
Two honest cautions, because practical does not mean reckless.
First, the model asking you questions does not change what you are allowed to feed it. Do not break any IT rules here (but question them if they restrict you).
If you need to, strip it down, anonymize it, talk about the situation in general terms, or use the enterprise tool your IT and HR have cleared. The three-question habit works just as well on an anonymized version.
Second, this helps you think and draft. It does not make the decision. A termination, a discrimination complaint, a legal call, a performance rating that affects someone’s pay, those stay with a human who owns the outcome and signs their name to it. Use AI to prepare the thinking, not to outsource the responsibility.
If you teach your team one thing about AI this quarter, make it this. Before you ask AI for an answer, ask it what it needs to know.
That is the entire lesson. It needs no training budget, no new license, and no framework. A manager can demonstrate it live in a standup in two minutes, and the people who freeze at the blinking cursor suddenly have a way in.