Emotional Intelligence and Relationship Building Questions

The core idea

Being on call means carrying the pager for a fixed stretch, often a week, and being the person who gets woken up when production breaks. It is a rotating duty rather than a job title: site reliability engineers, whose job is keeping production systems running, carry it, and on many teams so do the product engineers who wrote the service. An exhausted on-call engineer who feels unseen tends to either mask the fatigue and make more mistakes on the next page, or quietly disengage from the rotation altogether. Naming the exhaustion directly and pairing it with a real action protects both the person and the reliability of the systems they cover.

Why empathy is a reliability lever, not just a kindness

A burned-out on-call engineer is more likely to miss a signal, mis-triage a page, or take a shortcut on a runbook step, precisely because sustained sleep disruption degrades attention and judgment. Acknowledging that early, before it shows up as an incident, is a reliability intervention as much as a morale one.

Three phrases or actions for a follow-up 1:1 or Slack message

1. Name it plainly: "three pages in one week is a lot, I want to check in on how you're actually doing, not just whether the pages got resolved."
2. Offer something concrete, not just words: "I'm moving your on-call shift tomorrow to Jordan and picking up the pager myself tonight, so you can actually sleep."
3. Ask, don't assume, about pattern versus one bad week: "is this a rough stretch, or has this rotation been like this more often than it should be?"

When to escalate

If the pattern repeats across two or more rotations, if the same alert keeps paging, which points at the alert needing a fix rather than the person, or if the engineer mentions it's affecting sleep or life outside work beyond one bad week, I'd escalate to their manager and open a documentation trail, like an on-call load review, so the fix is structural, a schedule or alerting change, rather than one manager's private accommodation that quietly disappears.

Where this can go wrong

Empathy without a concrete offer, just saying "that sounds rough," can feel hollow to someone who's exhausted. Pair the acknowledgment with an actual, specific relief action whenever you can.

The core idea

The two techniques I reach for most in stakeholder meetings are paraphrasing, restating what someone said in your own words to confirm understanding, and labeling emotion, naming the feeling you're picking up on out loud so the person can confirm or correct it. I add open questions, ones that can't be answered yes or no, when I need someone to go further than their first answer.

Techniques with example phrasing

1. Paraphrasing: "So the concern is really about timeline risk, not the approach itself, is that fair?" This forces you to actually compress what someone said instead of nodding along, and it lets them correct a misread immediately instead of three weeks later.
2. Labeling: "It sounds like you're frustrated this came up again." Naming an emotion tentatively invites correction, "not frustrated, more just tired of re-explaining it," which often surfaces the real issue underneath the tone.
3. Open questions: "What would need to be true for you to feel good about this timeline?" instead of "Are you okay with this timeline?", which invites a one-word answer.

What happened

Marcus, a backend engineer on my team, went quiet in a planning meeting right after agreeing to a deadline. Afterward, one on one, I said "you went pretty quiet after we set that date, it sounded like something didn't sit right." He first said "it's fine," so I reflected his own word back next to what I had seen, without contradicting either: "okay, fine on the date, and also something shifted for you right after we set it. What was it?" He admitted the estimate didn't account for a migration dependency he hadn't mentioned, because he didn't want to seem like he was pushing back in front of the group. The two moves did different jobs. Labeling the reaction tentatively made it safe to say more. Setting his "fine" alongside the observation, rather than arguing with it, meant he never had to retract the word in order to keep talking, which is what usually stops people at "it's fine". In the meeting itself he had been asked straight out whether the date worked and had said yes, which is all a yes-or-no question in front of a group can reliably produce.

Where this can go wrong

Asserting an emotion as fact, "you're clearly upset," rather than offering it tentatively can feel like you're telling someone how they feel. Use "it sounds like" or "it seems like" so it stays easy for them to correct.

Rapport in the first three minutes isn't small talk, it's a set of moves that each relieve a specific tension the other person carries into the meeting: unfamiliarity, a power imbalance, the sense of being sold to, or the plain cost of the time they are giving you. Specificity is what makes them land, a generic phrase reads as scripted no matter how warm the tone.

Five phrases and their purpose

These work whether the first meeting is with a new engineering counterpart, a prospective account, or a business partner, the phrases don't need to change based on the other person's title.

Worked example
On a first call with a new account contact, opening with the "I looked at" phrase (referencing a specific detail from their public case study) and following it with the "good outcome" question typically gets a fuller, more specific answer than "how are you today" ever does. They work for two different reasons, which is exactly why they pair well. The first earns attention because it is grounded in something real about them and proves you spent time before the call. The second is deliberately generic in its wording and works by a different mechanism entirely: it hands them control of what the next twenty minutes are for, which is what drains a power imbalance. Misreading the second as another specificity move is the common mistake, and it pushes people to over-engineer it into a leading version ("I imagine your goal here is to get the migration unblocked?"), which takes back the control the open version just gave away and converts an invitation into a guess they now have to correct.

Trade-offs and pitfalls
A rapport phrase that's clearly scripted or generic ("how about this weather?") can cost more trust than it builds; specificity, not the exact wording, is what makes it work. Spending the full three minutes on rapport with someone who's visibly rushed backfires, read the room and compress to one phrase if they're pressed for time.

When direct disagreement is culturally discouraged, real concerns show up as indirection, not silence: a qualified agreement ("that could work, maybe"), a question standing in for an objection ("have we considered timeline risk?"), or disproportionate quiet from specific people rather than the room overall. The fix isn't forcing directness, it's changing how you listen and how you ask so disagreement doesn't require someone to break a norm to voice it.

Adjusting how you listen

• Treat a qualified "yes" as incomplete information, not agreement, and follow up privately: "You said maybe, what's the maybe?"
• Watch for who goes quiet relative to their own usual participation, not just who's quiet overall, a normally vocal person going quiet on one specific topic is a stronger signal than a naturally quiet person's silence.
• Distinguish the type of misinterpretation before reacting: is the quiet indicating disagreement, uncertainty about whether it's their place to speak, or a language-comprehension gap in a fast meeting? Each needs a different follow-up (a private check-in, a direct invitation, or slowing down and restating).

Adjusting how you ask

• Replace yes/no questions with ones that presuppose a concern exists and ask what it is: "What's the biggest risk with this plan?" is easier to answer honestly than "Does anyone disagree?"
• Offer a written, asynchronous channel as a second path, some people who won't contradict someone live will write a direct concern in a follow-up document.
• In multi-country meetings, actively assign floor time by name across time zones rather than relying on whoever speaks first, otherwise the same few people in the convenient time zone dominate every call while others' input is muted by scheduling, not culture.

Ongoing practice
A recurring, low-stakes channel, an anonymous document or a rotating check-in, explicitly for concerns that didn't get raised live. When someone does surface a concern that way, visibly act on it (crediting them if they're comfortable), so the norm builds on evidence rather than a promise that speaking up costs nothing.

Worked example
On a multi-country account team, an engineer in one region kept saying "no problem, we'll manage" about an aggressive timeline. Reading the qualified tone as incomplete, the account lead followed up privately and learned "we'll manage" meant unpaid overtime, not that the timeline was actually fine. Raising the concern in the next call as the lead's own observation, rather than attributed to the engineer, got the deadline moved without anyone having to publicly contradict the plan.

Trade-offs and pitfalls
Over-applying this becomes its own problem, treating every "yes" as suspect erodes trust with people who genuinely mean it. Calibrate to the specific person over time rather than assuming an entire culture behaves one way, individuals vary more than any stereotype does.

The core idea

In remote meetings you lose most body-language signal, so the cues that remain, camera on or off, posture within the frame, silence, and vocal tone, carry more relative weight than they would in person, and each one has more than one plausible meaning. The mitigation is almost always to check rather than assume.

Cues, interpretations, and mitigations

1. Camera off: could mean disengagement, a bad connection, someone eating lunch, or low confidence about being on camera that day. Mitigation: don't call it out publicly; message the person directly and low-stakes, "no pressure to be on camera, just checking you're still with us," which surfaces the real reason without embarrassing anyone.
2. Slouched or turned-away posture when the camera is on: can signal boredom, fatigue, or simply being at a standing desk mid-task. Mitigation: address the group's energy rather than one person, take a two-minute pause or ask a direct question to the room, rather than assuming that one person is checked out.
3. Prolonged silence after a question: could mean agreement, confusion, or someone waiting for a natural gap that a video call makes harder to find, since there are no shared physical cues for turn-taking. Mitigation: explicitly hand the turn to someone, "what's your take on this," rather than waiting for silence to resolve itself.
4. Flat or clipped vocal tone: can mean frustration, distraction, or just a bad audio day. Mitigation: name it gently and check, "you sound a little flat today, everything okay, or just one of those days?"

Where this can go wrong

Treating any single cue as a confirmed diagnosis, assuming camera off equals disengaged, is the main failure mode. Use each cue as a prompt to ask, not a conclusion to act on.