I had been a business partner for about three months when a night-shift manager stopped me in the car park. He had just been told, by me, that he could not move someone off a forklift rota the way he wanted to. He was not angry. He was tired. “Can I ask you something,” he said. “Whose side are you actually on?”

I gave him the answer I had been trained to give: that HR is there to make sure things are fair and lawful for everyone. He nodded in the way people nod when they have heard a sentence before. Then he walked to his car.

What I believed at the time

I believed my job was to be the guardrail. Managers made decisions; I checked them against policy and legal risk, and I said yes, no, or “not like that.” I was good at it. I could find the clause. What I could not do was explain why that manager kept bringing me decisions that were already half-made and already in trouble.

For a while I thought the answer was more manager training on policy. We ran two sessions. Attendance was fine. Nothing changed.

What changed

The shift came from a much smaller habit. A colleague in finance mentioned that she never started a budget conversation with the numbers; she started with “What are you trying to protect?” I began asking managers the same thing before we talked about any process.

The forklift manager, it turned out, was trying to protect a newer driver from a colleague he suspected was pressuring her into swapping shifts. He had no evidence, only a feeling, and the rota change was the only lever he thought he had. The policy conversation had been about the wrong problem.

Once I knew what he was trying to protect, we had a real problem to solve together instead of a rule to argue about.

We handled it differently: a private check-in with the newer driver, clearer shift-swap approval, and a conversation with the other colleague that did not rely on a hunch. I am deliberately leaving out details that belong to them.

My answer now

If a manager asks me whose side I am on today, I say: the side of the decision being good: good for the person affected, defensible for the organization, and something the manager can explain in their own words. Sometimes that means I side with the manager against a process. Sometimes it means I slow a manager down when they want to move fast.

It is not a clever line. It works because it is true of how I now spend my time: less checking, more earlier questions.

What is still unresolved

This approach assumes managers will tell me what they are protecting. Some do not, especially when what they are protecting is themselves. I have not found a reliable way to get to that conversation without it feeling like an accusation. I also know that “neutral problem-solver” is not always the right stance; there are situations where HR has to act on a concern regardless of what a manager wants. I would like to hear how others hold both positions without confusing people about which one they are in.

Situation. Our onboarding checklist had grown to 41 items across HR, IT, finance and the hiring team. Everyone had added something after a bad week. Yet in their first fortnight, new hires still messaged me the same five questions: who approves leave, how expenses work, where the team’s real priorities were written down, who to ask about tooling, and when they would get feedback.

Action. I printed the checklist and sat with two recent hires and one hiring manager for forty minutes. For each item we asked two questions: who actually owns this? and would a new hire notice if it was skipped? Items owned by a system (account creation, payroll setup) moved to an internal task list with named owners. Items nobody could explain were deleted. The five recurring questions were added, answered rather than linked.

Observation. Twelve items survived. The checklist now fits on one screen and every item has a human owner. In the next four cohorts, I received far fewer repeat questions. I want to be careful here: I did not count messages before the change, so this is my honest impression plus two new hires telling me the first week felt “less like a scavenger hunt.”

Every item you add to a checklist is something you are promising to check. We had made 41 promises and kept about 20.

Lesson. A long checklist can hide the absence of ownership. Shortening it forced conversations about who was responsible for what. The IT handoff in particular had been unowned for months. The limit: our company is small and co-located twice a week. I do not know how well this translates to a large, fully remote organization.

Question for peers. If you have done something similar at scale, how did you stop the list from growing back?

Context

We are two HR people supporting clinics and outreach teams. The same questions about leave, travel reimbursement and scheduling arrive every week, and our written FAQs were out of date. We wanted to know whether an AI writing tool could help us refresh them faster.

Hypothesis

If we gave the tool our already-published policy documents, it could produce FAQ drafts that would take less time to finalize than writing from scratch. We agreed up front that no employee names, cases, medical information or anything not already public to staff would go into the tool.

What we did

• Picked twelve frequently asked questions across three policies.
• For six, we wrote answers from scratch. For six, we started from an AI draft.
• Timed ourselves on each: from blank page (or draft) to “ready for a manager to read.”
• Asked a skeptical clinic manager to review all twelve without knowing which were which.

What we observed

The AI drafts looked finished on arrival. That was the trap. They used a tone our staff would find stiff, occasionally described practices from other jurisdictions, and twice stated something our policy did not say. Editing those drafts took roughly as long as writing from scratch for the complex questions, and noticeably less for the simple, structured ones.

The drafts were fastest where we needed help least, and slowest where judgment mattered most.

The clinic manager correctly identified four of the six AI-started answers, mostly, she said, because they “sounded like nobody in particular.” She preferred the edited versions of the simple ones.

Limits

Twelve questions and two writers is a tiny sample. We knew which answers were AI-started while editing, which may have changed how hard we looked. The tool, our prompts and our policies are specific to us; your results could differ in either direction.

What we will try next

We will use drafts only for structure on simple, procedural questions, and keep writing the answers that involve discretion ourselves. We also wrote down what we decided not to automate (anything touching individual circumstances) so the decision does not quietly erode when we are busy.

For six years I worked in a large retail HR team. There was an employee relations specialist, a reward team, a learning team and a very patient HR systems analyst. When something landed on my desk that was not mine, I knew exactly whose it was.

My first month as the only HR person at a sixty-person startup, I kept a list of every request that reached me. It had ninety-three lines. They ranged from an employment contract for a new country to a founder asking whether we “should do OKRs.”

The assumption I brought with me

I assumed that being useful meant being responsive. In my old team, a fast answer was a kindness, because the answer usually existed. Here, most answers did not exist yet. Being fast meant inventing policy in a Slack thread.

What shifted

The turning point was small. A manager apologized for “bothering me with something so minor”: a question about a notice period that I had been sitting on for a week because it was tangled up with a contract template I had not finished. He was not bothering me. I had simply never told anyone what I was working on, so everything looked equally ignored.

That week I wrote a one-page note: what I do now, what I will do next quarter, and what I don’t do yet. It included the things I was deliberately not building, like a formal performance cycle. I shared it in our all-hands and pinned it.

The note did not reduce the work. It reduced the number of people wondering whether their request had disappeared.

What I do differently

I say “not yet” more, and I attach a month to it. I ask leaders to choose between two things rather than asking whether they want both. And I update the note every quarter, including the items that slipped. People have been more forgiving of slipped dates they can see than of silence.

What remains unresolved

I still do not know how to handle requests that are small, urgent and emotionally heavy, like a manager who needs help today with a difficult conversation. They never fit on a quarterly note, and they are often the most important thing I do all week. I would value hearing how other team-of-one practitioners protect time for those without letting everything else slide.

Situation. Every month, forty production and engineering supervisors joined a one-hour HR update. We covered policy changes, upcoming training, safety reminders and benefits deadlines. Attendance was high because it was expected. I had started to suspect that very little of it reached the people on the floor.

Action. Before changing anything, I asked five supervisors from different shifts a simple question the day after the meeting: “What did you pass on to your team?” The honest answers ranged from “the benefits deadline” to “nothing yet.” None of them had notes; the slides were long.

We replaced the meeting with a one-page written brief sent on the same day each month. At the top are three lines labelled say this to your team, written in the words a supervisor would use at a shift start. Everything else is reference. We asked supervisors to reply with one question if anything was unclear.

Observation. Replies were few but specific, mostly about how a change applied to a particular shift pattern, which the old meeting never had time for. Two supervisors told me they now read the three lines aloud at the start-of-shift huddle. Others asked us not to lose live discussion entirely, so we kept a fifteen-minute optional Q&A. About a third join.

The meeting was not the communication. It was the ritual we used to avoid writing things down clearly.

Lesson. Writing “say this to your team” forced us to decide what mattered each month. The limit: this relies on supervisors reading email, which is not true for every team leader in every plant. For those, we still brief in person.

Question for peers. How do you confirm that a message reached frontline teams, without adding a survey?

Context

We run a half-day workshop on giving feedback for newly promoted charge nurses. Evaluations were consistently positive. But when I shadowed shifts, I rarely saw the method used. The workshop was fine. The Monday after the workshop was the problem: short staffing, no time, and no one to ask “how would I say this?”

Hypothesis

If each participant had one short, practical follow-up soon after the workshop, focused on a real conversation they were about to have, they would be more likely to use the method at least once.

What we did

• With the second cohort (14 people), we booked a 20-minute one-to-one about two weeks after the workshop, at a time that fitted their roster.
• Each person brought one real upcoming conversation. We practised the opening two sentences only.
• Four weeks later, we asked both cohorts one question: “Have you used the method in a real conversation? Describe it briefly.”

What we observed

In the first cohort, which had no follow-up, a few people described a real use. In the second, most did, and their descriptions were more specific, often quoting the opening they had practised. Several said the follow-up was the first time they felt permission to adapt the method to their own words.

Practising two sentences for a real conversation did more than rehearsing a whole model for an imaginary one.

Limits

These are self-reports from two small cohorts, and the second cohort received more attention overall. We cannot tell whether the method, the practice, or simply being checked on made the difference. Rostering the follow-ups took real coordination time that we have not yet costed properly.

Next iteration

For the next cohort we will offer the follow-up as a small peer pair instead of one-to-one with L&D, to test whether the effect survives without us in the room, and to see if it is cheaper to sustain.

Situation. Our annual survey had 38 questions and a falling response rate in stores and warehouses. When I visited sites and asked why, the most common answer was not “no time.” It was some version of: “We filled it in last year. What happened?”

Action. For store teams, we stopped sending the long survey. Instead, every quarter, we ask three questions: What made your work harder this month? What is one thing your manager could change? Do you feel you can raise a problem without it being held against you? The first two are open text; the third is a simple scale. Store managers see their own results within a week and agree one visible change, which they post in the break room before the next round.

Observation. Over two quarters, participation in stores rose and the open-text answers became more specific, often about scheduling, equipment and handovers. Some changes were tiny, like moving the rota release day. Staff mentioned those changes in the next round, which never happened with the annual survey.

Three questions people answer honestly were worth more to us than thirty they skimmed.

Lesson. The visible change, not the survey, carried the message. The trade-off is real: we can no longer compare our results with external benchmarks, and head office misses the trend lines. We kept the long survey for office teams for now, which creates its own inconsistency.

Question for peers. If you have shortened a survey, how did you explain the loss of comparability to leadership?

I left my HR role to care for a parent. When I started applying again two years later, I had prepared a careful paragraph explaining the gap. In my first interview, the hiring manager read my CV and said, “Caring for family: that is hard work. What did it teach you?” I had not prepared for that question at all.

What I assumed

I assumed I would be behind, and that I would need to hide how far. I spent weeks reading about HR technology and new regulations. Some of that was useful. Much of it I could have learned in the job, and did.

What surprised me

The tools caught up in a month. The confidence took longer. For the first few months I double-checked everything and apologised too often. My manager eventually told me I was answering questions nobody had asked.

What I did not expect was that I would be better at one part of the job than before. Hotel operations staff bring me situations that are rarely just about work: a sick child, a visa worry, a family obligation abroad. I now listen longer before I reach for a policy. I recognise the specific tiredness of someone holding two responsibilities at once.

I expected to be behind on tools. I did not expect to be ahead on patience.

What I would tell someone returning

Name the break plainly. Say what you did and what you learned, in one or two sentences. Then talk about the work. The people worth working for will ask a real question about it; the ones who treat it as a flaw have told you something useful too.

What is unresolved

I was lucky with that first interviewer. I know other people returning after a break, especially after caregiving, do not get the same question. I would like to understand what HR teams can change in screening so that the question is asked by design, not by luck.