AI Remakes Office Work: McKinsey to Cut 10% of Support Roles as Corporate Layoffs Mount

AI-driven job cuts are here. HR and Customer Support need a plan

McKinsey & Company plans to cut roughly 10% of non-client-facing roles over the next two years as AI tools automate work once handled by people. That follows a broader shift in 2025: Accenture restructured around AI and digital transformation and cut at least 11,000 roles, Salesforce removed about 4,000 customer service jobs after AI agents began handling nearly half of interactions, and Amazon trimmed around 14,000 corporate jobs tied to automation and efficiency pushes. Industry trackers also point to reductions at Microsoft, Intel, and Google as focus moves to AI development and automated systems.

Support, admin, and operations are bearing the brunt. The message is clear: repetitive, process-heavy tasks are getting absorbed by software, while companies reallocate headcount to AI-driven work and client-facing consulting.

Why this is happening

AI handles routine, high-volume, rules-based work cheaper and faster. Leaders see cost savings, higher responsiveness, and new revenue lines in AI consulting and data products. That pulls budget away from traditional support functions and into automation, analytics, and client delivery.

What this means for Customer Support leaders

• Tier-1 volume will keep shrinking as bots and self-serve improve containment and deflection.
• Human roles shift to complex cases, escalations, quality assurance, knowledge management, and coaching bots.
• Your scoreboard changes: containment rate, assisted vs. unassisted resolution, FCR on complex issues, CSAT by channel, and cost per contact.
• Team design matters: fewer generalists, more specialists in policy, workflows, and tooling.

What this means for HR leaders

• Workforce planning is now skills-first: inventory what people can do, map it to AI-era roles, and redeploy fast.
• Reskilling is a core product: short, job-relevant pathways with measurable outcomes, not generic learning sprees.
• Separation practices need an upgrade: fair packages, transparent criteria, and real placement support.
• Policy, ethics, and compliance must keep pace with automated decision-making and data usage.

Roles most exposed vs. roles rising

• Higher exposure: Tier-1 support, basic data entry, form processing, routine scheduling, standard reporting.
• Growing demand: Conversation designers, bot trainers, support QA analysts, knowledge engineers, workflow automation specialists, HRIS analysts, talent intelligence, change communications.

90-day action plan

• Weeks 1-2: Assess
  • Run a task audit: list top 20 repeatable tasks per team. Flag what can be automated in 3, 6, 12 months.
  • Pull baseline metrics: AHT, FCR, containment, CSAT, cost/contact, time-to-fill, internal mobility rate, skills inventory coverage.
  • Freeze net-new hiring for tasks likely to be automated within 6 months.
• Weeks 3-6: Redesign work
  • Stand up a small Automation Pod (Ops lead, QA, knowledge owner, analyst). Give it a clear backlog and weekly targets.
  • Shift agents handling repetitive tickets into QA, knowledge upkeep, and bot training rotations.
  • Create 3 skill pathways with badges: Support Ops (QA + knowledge), Automation Ops (workflows + prompts), Analytics (SQL + dashboards).
• Weeks 7-12: Implement and measure
  • Ship two automations per week (e.g., refunds, password resets, order status, password unlocks, address changes).
  • Set guardrails: human review for high-risk actions, clear escalation criteria, and bias checks on automated triage.
  • Publish a weekly scorecard: automation throughput, error rate, customer effort score, redeployment count, training hours/person.

People-first practices that actually help

• Transparent criteria: Define how roles are selected for automation or redesign. Share the rubric.
• Redeploy before you replace: Internal mobility and short reskilling sprints beat external hiring for many roles.
• Manager scripts: Provide concise talking points for change, timelines, and support options.
• Financial runway: Offer meaningful severance and job placement support so people can re-skill without panic.

Skill stacks that move the needle

• Support pros: Case taxonomy, knowledge graph basics, prompt writing for support flows, QA frameworks, basic SQL, dashboarding.
• HR pros: Talent intelligence, skills ontology, HR data literacy, AI policy and vendor due diligence, change communications.
• Managers: Outcome-based goals, process mapping, risk controls, and performance coaching in mixed human-AI teams.

KPIs for an AI-augmented support org

• Containment rate (target by issue type), CSAT gap bot vs. human, AHT on complex cases, recontact within 7 days, cost/contact.
• Automation error rate and time-to-detect, percent of articles updated monthly, model feedback coverage, training hours per FTE.
• HR: internal fill rate for new roles, redeployment rate, time-to-redeploy, completion rate on learning pathways, attrition of critical skills.

Risk checklist

• Automated actions without audit trails or human oversight for high-stakes cases.
• Knowledge base drift causing wrong bot answers and higher recontacts.
• Silent skill loss: experienced agents exit before you capture workflows and edge cases.
• Shadow tools: teams adopting unvetted AI apps that create data and compliance risk.

Where to learn, fast

If you need focused training for Support and HR roles moving into AI-enabled work, explore practical course paths and certifications.

• AI courses by job for Customer Support and HR tracks
• AI Automation certification for ops and workflow roles

Further context

For a broader view on how roles are changing and which skills are gaining value, see the World Economic Forum's analysis of job trends.

• WEF: The Future of Jobs Report

This shift is real and ongoing. If you lead Support or HR, your edge is simple: redesign work, build skill pathways, protect people through clear transitions, and measure outcomes weekly. Do that, and you won't just absorb the change - you'll direct it.