On February 1, Mark Zuckerberg told investors that 2023 would be Meta's "year of efficiency". Six weeks later, Meta announced about 10,000 more job cuts, on top of the 11,000 announced in November. It was not alone. Amazon, Microsoft, Alphabet, Salesforce and hundreds of smaller companies cut staff in the first months of the year. By the end of August, Layoffs.fyi had counted 224,503 tech layoffs in 2023, already more than its total for the whole of 2022. January alone accounted for nearly 90,000.

The layoffs have slowed since then, but the people who remain are working in a different atmosphere. Three other numbers describe it:

  • Gallup's State of the Global Workplace 2023 (June): 23% of employees worldwide are engaged, a record high, but 59% are "quiet quitting" (not engaged) and 18% are "loud quitting" (actively disengaged). Gallup estimates the cost of low engagement at 8.8 trillion dollars.
  • Microsoft's Work Trend Index 2023 (May): 49% of people say they are worried AI will replace their jobs, while 70% would delegate as much work as possible to AI to lighten their workload.
  • A field experiment with 758 BCG consultants (published this month) found that GPT-4 helped them complete 12.2% more tasks, 25.1% faster and at higher quality, on tasks within the AI's capabilities. On a task outside them, consultants using AI were less likely to get the right answer.

Teams are smaller, more anxious, being asked to do more, and being handed powerful new tools whose effects are uneven. This post looks at what the research says about leading in those conditions.

1. What layoffs do to the people who stay

Research on "layoff survivors" goes back to the 1980s, when Joel Brockner and colleagues began studying how employees who keep their jobs react to colleagues losing theirs. The findings have been consistent.

Effect What it looks like Why it happens
Guilt and grief Survivors feel bad about colleagues who left, especially close ones Relationships and identity are disrupted
Job insecurity People keep their heads down, avoid risks, update their CVs If it happened once, it can happen again
Reduced commitment Less discretionary effort, more focus on self-protection The psychological contract feels broken
Workload strain The same work spread across fewer people Cuts often outpace changes to scope
Talent drain The strongest performers leave voluntarily They have the most options

The psychological contract, a concept developed by Denise Rousseau, describes the unwritten beliefs employees hold about what they owe their employer and what they will get in return: security, development, fair treatment. Layoffs, especially at profitable companies, tend to break that contract. People do not stop working when that happens. They recalibrate: they give what is required and keep the rest for themselves. That is quiet quitting, seen from the inside.

2. Fairness is the strongest lever you have

Brockner's research found that survivors' reactions depend less on the fact of the layoffs than on how they were carried out. This is the domain of organisational justice, which researchers usually divide into four parts:

Type The question employees ask What good looks like
Distributive Were outcomes fair? Cuts shared sensibly, not just at the bottom; leaders accept some of the cost
Procedural Was the process fair? Consistent criteria, applied without favouritism, with a way to raise concerns
Interpersonal Were people treated with dignity? Personal conversations, not mass emails or cut-off badges
Informational Were we told the truth, in time? Clear explanations of why, what happens next and what is not yet known

A meta-analysis led by Jason Colquitt (2001) showed that each type predicts different outcomes. Procedural justice is closely tied to trust in the organisation and its leaders. Interpersonal and informational justice are tied to how people feel about their direct manager.

That second point matters. Most managers did not decide the layoffs. But they control how the aftermath is handled in their team, and that is what their people will remember.

3. Rebuilding trust: what managers can control

Trust is commonly broken down into three components: ability (can you do what you say?), benevolence (do you care about my interests?) and integrity (do you follow principles I accept?). Layoffs damage all three. They suggest that leaders did not plan well, that they do not care about people, or that their earlier promises were empty.

You cannot rebuild trust with a speech. You rebuild it with consistent behaviour over months.

Tell the truth, including about what you do not know. "I don't know whether there will be more cuts, and I will tell you as soon as I do" is more credible than reassurance that later proves false. One false reassurance can undo months of trust.

Reset the work, not just the headcount. If the team shrank by 20%, the scope has to shrink too. Make an explicit list of what the team will stop doing, and get it agreed above you. Asking fewer people to do the same work is a recipe for burnout and departures.

Talk about the people who left. Pretending nothing happened adds to survivors' guilt. Acknowledge the loss, thank the people who left for their work, and help them with references and introductions if you can.

Hold one-on-ones that are about the person. Ask how people are doing, what worries them, what they need. Listen more than you talk.

Be visibly fair from now on. Who gets the interesting projects, the promotions, the flexibility? Survivors watch closely for signs that decisions are political.

Do not overreach on loyalty. Do not ask people to show more commitment than they believe the organisation has shown them. It reads as hypocrisy.

4. Engagement after quiet quitting

"Quiet quitting" spread through social media in 2022 as a term for doing the job as described and nothing more. Gallup's data suggest it describes the majority of workers, and that it is mostly a response to how people are managed. In the 2023 report, Gallup asked the "quiet quitters" what they would change about their workplace. The most common answers were about engagement and culture (recognition, better relationships with managers, clearer goals, more autonomy), followed by pay and benefits, and then wellbeing.

The research on motivation gives a clear picture of what drives sustained effort. Self-determination theory, developed by Edward Deci and Richard Ryan, identifies three basic needs: autonomy (having a say in how you work), competence (feeling effective and growing) and relatedness (feeling connected to others). Layoffs typically hit all three: decisions are made from the top, people lose colleagues they relied on, and the scramble to cover gaps leaves little room for learning.

In practice, for a team recovering from cuts:

  • Autonomy: involve the team in deciding what to stop doing and how to divide the remaining work.
  • Competence: give people time to learn the systems and work they inherited, and recognise when they do it well.
  • Relatedness: rebuild rituals that were lost, re-form working relationships, and make time for the team to talk about more than tasks.

Goals that help rather than hurt

In a year of efficiency, the temptation is to add targets. Goal-setting research by Edwin Locke and Gary Latham, built on hundreds of studies, shows that specific, challenging goals do raise performance, but only under some conditions: people must be committed to them, have the ability and resources to reach them, and get feedback on progress. For complex work, learning goals ("figure out three ways to cut the deployment time") often work better than outcome targets ("cut deployment time by 40%"). Stretch targets imposed on an overloaded, distrustful team usually produce gaming, shortcuts and departures.

5. Introducing AI without breaking the team

Generative AI arrived in offices this year, often before any policy did. The first rigorous studies of its effects at work were published in 2023, and they are more nuanced than the hype.

Study Setting Main finding
Peng, Kalliamvakou, Cihon and Demirer (February) Developers writing an HTTP server in JavaScript, with or without GitHub Copilot The Copilot group finished 55.8% faster
Brynjolfsson, Li and Raymond (April) 5,179 customer support agents at a software company, with a GPT-based assistant introduced in stages 14% more issues resolved per hour on average; 34% for novice and low-skilled agents, little effect for the most experienced. Customer sentiment improved and attrition fell
Dell'Acqua, McFowland, Mollick and others (September) 758 BCG consultants with GPT-4 On tasks inside the AI's capabilities, 12.2% more tasks, 25.1% faster, over 40% higher quality. On a task outside them, AI users were 19 percentage points less likely to be correct

The BCG study's authors call this the "jagged technological frontier": AI is very good at some tasks and poor at others that look similar in difficulty, and the boundary is not obvious. People using AI on the wrong side of the frontier tended to trust it too much.

The motivational implications for teams are important:

  • AI can raise competence, especially for juniors. In the support study, the AI seemed to spread the tacit knowledge of top performers to newer colleagues. That is a powerful benefit for teams that lost experienced people in layoffs.
  • AI can also threaten competence and identity. People whose expertise is suddenly matched by a tool may feel devalued. Experienced staff in the support study gained little and may have seen their advantage shrink.
  • AI raises job security fears. Introducing AI right after layoffs will be read by many as preparation for the next round, whatever the intent. Microsoft found that managers were more likely to value AI for productivity than for headcount reduction. Say which you mean, and be honest.

A practical approach

  • [ ] Set clear rules on what data can go into which tools. Frame them as enabling safe use, not forbidding experimentation.
  • [ ] Run experiments in the open: let the team try tools on real tasks and share what works and what fails.
  • [ ] Map your team's frontier: where AI is reliable, where it needs checking, where it should not be used. Update the map as tools change.
  • [ ] Keep human review for anything customer-facing, high-stakes or factual.
  • [ ] Use time saved to reduce overload and invest in learning, not just to add more work. If every gain becomes a new target, people will stop reporting gains.
  • [ ] Talk honestly about what AI means for roles. Uncertainty is often worse than bad news.

6. The manager's own load

Most layoffs this year have included managers, and several companies have announced explicit plans to remove management layers. The managers who remain often have more direct reports, more cross-team work and their own fears about job security.

Gallup has long reported that managers account for about 70% of the variance in team engagement. If managers burn out, engagement falls across the board.

If you lead managers:

  • [ ] Keep spans of control manageable. A manager with 15 reports cannot hold real one-on-ones with each of them.
  • [ ] Give managers the information they need to answer their teams' questions honestly, early.
  • [ ] Train managers in difficult conversations. Many have never had to lead a team through layoffs.
  • [ ] Ask managers how they are doing, and mean it.

7. A six-month playbook

First month after cuts

  • [ ] Hold a team meeting to acknowledge what happened, explain what you know and what you do not, and answer questions honestly.
  • [ ] Hold one-on-ones with everyone within two weeks.
  • [ ] Agree on a list of work to stop, pause or slow down, and get it signed off.
  • [ ] Redistribute responsibilities explicitly, with handover time.

Months two and three

  • [ ] Restore the team's rituals: planning, retrospectives, social time.
  • [ ] Set a small number of clear goals, with the team's input.
  • [ ] Recognise specific contributions, especially from people who absorbed extra work.
  • [ ] Start an open AI experiment on a few well-understood tasks.

Months four to six

  • [ ] Review workload honestly. If people are still stretched, escalate.
  • [ ] Discuss development and career paths with everyone. Survivors need to see a future.
  • [ ] Measure: run a short engagement pulse or simply ask the team what has improved and what has not.
  • [ ] Share what the team learned about AI tools with the wider organisation.

Why this matters

  • The cost is in who stays. Layoffs are usually justified by savings, but the larger hidden cost is reduced effort and voluntary departures among those who remain.
  • Fair process is controllable. Managers rarely choose the cuts, but they control much of the fairness, honesty and dignity that determine how survivors react.
  • AI's effects are uneven. Productivity gains are real on some tasks, losses are real on others, and teams need to learn where the frontier lies.
  • Trust is the foundation. Without it, neither goals nor tools will produce the performance leaders expect.

Final Thoughts

"Efficiency" is a reasonable goal for a business. But a team is not a machine, and cutting it does not simply produce a smaller version of the same thing. It produces a different team: one with fewer people, more fear, less slack and new questions about whether the organisation can be trusted.

Generative AI adds a twist. The tools may genuinely help stretched teams do more and help junior people learn faster. They may also deepen fear and over-reliance if introduced carelessly. Which way it goes depends largely on whether people trust their leaders.

None of the research points to a clever trick. It points to old virtues done consistently: tell the truth, treat people with dignity, make fair decisions in the open, adjust the work to the people you have, and involve them in deciding how to use the new tools. That is how teams recover. And it is how they become the kind of team good people choose to stay in.

Sources