Reskill, redeploy or hire?
When it comes up
When an AI rollout frees significant capacity in a role, when a business plan needs new AI-related skills, or when finance proposes headcount reductions based on AI efficiency.
Your options
Upskill in place
Fits when: Most affected tasks are augmented: the role stays, but people need to work with AI and take on higher-value work.
Watch out: Training without time to practise or redesigned work doesn't change behavior.
Reskill and redeploy
Fits when: A large share of hours is automated and there are growing roles with adjacent skills inside the company.
Watch out: Needs named target roles, funding and internal-hiring rules; slow programs lose people.
Hire externally
Fits when: You need skills you cannot build in time (for example AI engineering or AI governance).
Watch out: Expensive and competitive; can signal to employees that internal growth is closed.
Reduce
Fits when: Demand for the work has truly fallen and redeployment is not possible.
Watch out: Trust, legal consultation duties and employer-brand damage; loss of institutional knowledge.
How to decide
- Map the role's tasks and the exposure mix (automate / augment / keep human) — use the workforce impact mapper.
- Size the capacity freed in FTE and check it with a pilot before planning on it.
- Find adjacent growing roles and the skills gap between them.
- Choose per role: upskill in place when most tasks are augmented; redeploy when freed capacity is large and adjacent roles exist; hire for skills you can't build in time; reduce only when demand has fallen and redeployment isn't possible.
- Fund and track moves, not just programs: people trained, people moved, retention after the move.
Trade-offs
Reskilling and redeployment keep knowledge and trust but take time and money; hiring is faster for scarce skills but costly; reductions show savings quickly but carry trust, legal and knowledge costs.
Real examples
- Salesforce's CEO said in September 2025 that customer-support headcount fell from about 9,000 to about 5,000 as AI agents handled more conversations; a spokesperson said hundreds of employees were redeployed into professional services, sales and customer success. Source: Fortune — Salesforce CEO Marc Benioff says the company cut 4,000 customer support roles as AI steps in (Sep 2, 2025) ↗
- Amazon's Upskilling 2025 pledge committed over $1.2 billion from 2019 through 2025 to give 300,000 employees access to free training programs, including college tuition for front-line employees. Source: Amazon — Upskilling 2025: over $1.2 billion for 300,000 employees ↗
- The WEF Future of Jobs Report 2025 estimates that of every 100 workers, 29 will be upskilled in their current roles, 19 reskilled and redeployed within their organization, and 11 will need training they are unlikely to get by 2030. Source: UNLEASH — 5 things HR needs to know from WEF's 2025 Future of Jobs report (Jan 10, 2025) ↗ · World Economic Forum — Future of Jobs Report 2025 (Jan 2025) ↗
What to check first
- The exposure mix and hours freed for each affected role.
- Which growing roles exist internally and their skill requirements.
- Your internal fill rate today and what blocks internal moves.
- The budget available for reskilling, separate from the savings target.
Learn and use
Sources
- Fortune — Salesforce CEO Marc Benioff says the company cut 4,000 customer support roles as AI steps in (Sep 2, 2025) ↗
- Amazon — Upskilling 2025: over $1.2 billion for 300,000 employees ↗
- UNLEASH — 5 things HR needs to know from WEF's 2025 Future of Jobs report (Jan 10, 2025) ↗
- Coursera — WEF Future of Jobs Report 2025: key findings (Jan 2025) ↗
Last reviewed 2026-10-01.