Findings are manually reviewed Tracker-level summaries linked to their supporting Signals.
Reviewed finding
AI is changing work before it clearly changes workforce size
Across the current evidence base, AI is already compressing, substituting and reorganizing specific tasks, while corresponding changes in workforce size are often not directly evidenced. Workflow change and headcount effects should be treated as separate questions.
This does not establish that AI generally has little employment impact, or that workforce reductions will not occur. It means only that within the current published evidence base, task and workflow changes are often more directly evidenced than net workforce effects.
Reviewed finding
Measured AI gains are clearest at the workflow level
The strongest quantified outcomes in the Tracker are tied to concrete work units—coding, review, planning, reporting and logistics. Institutions evaluating AI should therefore measure changes in specific workflows, not rely only on enterprise-wide adoption metrics.
These metrics come from different organizations, functions and measurement methods. They must not be aggregated into a common productivity rate or treated as an expected return from AI generally.
Reviewed finding
Human authority remains explicit in some high-stakes AI workflows
Several healthcare and public-sector cases combine AI delegation with explicit human control points. Where the evidence is strongest, humans retain confirmation, review or final-judgment authority.
Do not generalize this into a claim that human oversight is the norm across AI deployments. Human decision rights are not documented for many Tracker Signals.
Evidence explorer
Browse evidence cases
7 of 41 visible
Filter the evidence
Active
Anthropic
United States · Artificial intelligence
Anthropic employees report more AI-agent coordination and supervisory meta-work
Explicit outcome claims
AI mechanism → work change
AI agents execute multi-step work → Role Redesign
How the organization responded
Not yet evidenced
What changed
Employee work increasingly includes writing specifications, monitoring agent pipelines and reviewing agent output alongside other tasks.
Key outcome
Estimated meta-work share of the average employee’s interactive Claude sessions — Company-reported estimate from automatically coded one-line session summaries; session share is not working time, FTE displacement or a causal employment effect. — Before 10 percent of interactive sessions · After 22 percent of interactive sessions
Yadea
China · Two-wheeler manufacturing / e-commerce
A junior designer coordinates an AI-assisted creative workflow
Explicit outcome claims
AI mechanism → work change
AI-assisted work → Role Redesign
How the organization responded
Not yet evidenced
What changed
宋雨蒙 built an agent spanning creative stages.
Key outcome
Reported product-detail production time — One-worker self-report, not a cohort outcome — Before 23 hours · After 8 hours
JD.com
China · Retail & e-commerce
Customer-service worker transition into AI training roles
Explicit outcome claims
AI mechanism → work change
Automation → Role Redesign
How the organization responded
Redeploy · Create Roles
What changed
Customer-service experience is being repurposed into training, configuring and improving AI service systems, with reported job conversion at scale.
Key outcome
Employees completing job conversion or upgrading — Thousands of employees achieved job conversion/upgrading over roughly two years
LONGi Green Energy
China · Solar / advanced manufacturing
AI inspection shifts frontline quality work toward model training and verification
Explicit outcome claims
AI mechanism → work change
Automation → Role Redesign
How the organization responded
Redeploy · Reskill
What changed
AI performs substantial visual quality inspection, human workers verify flagged defects, and a documented former frontline quality inspector transitioned into AI-model training using frontline inspection expertise.
Key outcome
Inspection labor cost — Inspection labor cost was reported as reduced to approximately one-eighth of its previous level
Salesforce
United States, Global · Enterprise software
AI agents take over routine support as support engineers move into new roles
Explicit outcome claims
AI mechanism → work change
AI agents execute multi-step work → Role Redesign
How the organization responded
Redeploy · Reskill · Reduce
What changed
Routine customer-support conversations increasingly move to Agentforce, while human support engineers concentrate on more complex or proactive customer work and hundreds have moved into other functions.
Key outcome
Customer conversations handled by Agentforce — Agentforce had handled 2.6 million customer conversations
The Tracker reviews public sources and distinguishes operational evidence, observed work change and reported outcome claims. Claims are not necessarily independently verified or caused by AI. Records are reviewed periodically as evidence becomes available.