AI Finds the Talent, But Who Designs the Work?

Enterprises now identify employee skills more accurately using AI that mines work records from systems like Salesforce and Jira. Workday's planned adoption of TechWolf's technology and Microsoft People Skills' use of Microsoft 365 signals illustrate this trend. However, the article argues that identifying who has a skill does not equal getting work done.

A case study makes the gap concrete: a regional margin decline requires a two-week diagnosis. The talent system recommends three people — an operations specialist who contributed to a prior analysis, a senior financial analyst with diagnostic experience, and a regional backbone who knows stores and promotions. Yet work stalls: the specialist lacks clarity on deliverables, the analyst is tied up in budget review, and the backbone cannot reduce store visits.

The core problem is that skill recognition solves information asymmetry, but the remaining barriers — time, backup coverage, managerial support — are arrangement problems. The article then walks through three organizational mechanisms that close the gap:

Task structuring: The analysis lead first agrees with the general manager on a concrete deliverable — a diagnosis listing key factors, evidence, and open issues — rather than a vague "find the cause." This defines scope, comparison standards, and acceptance criteria, turning "need someone who knows margin analysis" into specific roles: data comparison, field verification, professional judgment.

Task contracts (任务契约): Each participant and their manager makes explicit commitments: the specialist confirms data scope and submission time; the analyst defines judgment scope; the finance head confirms review slots; the regional head adjusts visit schedules and revises sales targets. The general manager accepts the diagnosis per pre-agreed standards. These contracts make division of labor enforceable.

Task-driven responsibility network (任务驱动的责任网络): When the budget review suddenly demands more analyst time, the network surfaces the conflict. The analysis lead quantifies the gap, the finance head offers alternative slots, and the general manager decides to delay the diagnosis by two days (accepting the trade-off). All affected parties receive the updated plan, and obsolete commitments are retired.

The article notes that existing project charters or resource plans can serve as contracts if they already capture these elements; software like Mido's MOOS (菌丝体组织操作系统) is designed to connect skill recommendations, contracts, dependency tracking, and evidence across HR, finance, and business systems. However, software cannot replace managerial prioritization, turn notifications into commitments, or claim outcome improvements without empirical validation.

Closing recommendation: leaders should pick a stalled task where people are identified but work hasn't started, diagnose whether the blocker is skill evidence, task definition, role constraints, or resource conflicts, and fix that specific arrangement. Until the discovered capability is organized into a deliverable cooperation, even perfect recommendations remain unused potential.