Rethinking the Organization | Article 1 of 4

AI and the Question Before the Question

For most of history, technology helped people find answers after they recognized the need. AI is beginning to change the front of that chain: it can increasingly identify that something deserves attention before anyone requests the analysis.

Most AI still waits for a prompt

An HR leader asks an AI assistant to compare turnover in critical roles, recent movement in employee sentiment, internal mobility and changes in hiring velocity. The system can assemble information from several sources, interpret relevant context and return a useful view in seconds.

That is a meaningful advance over reconstructing the same picture manually. Yet the HR leader has already performed the essential first step. They recognized that workforce continuity might be at risk, selected the signals worth examining and framed the analysis. The technology improved the response; it did not identify the need for the response.

This is the dominant pattern in enterprise AI today. A system may be conversational, highly capable and connected to multiple applications, but it usually waits for a person, a process event or a prescribed trigger to tell it what to do.

The relationship is older than software. Lists, files, databases, reports and search engines all made information easier to store and retrieve. Enterprise platforms made work more controlled and connected. Across each generation, the person remained responsible for noticing the issue, deciding what they needed to understand and organizing the available information around that purpose.

Most AI can answer the question. The next shift is recognizing that the question exists.

From response to recognition

Now the same HR leader begins the day with a different message. A combination of changes across several critical roles may warrant review. Internal movement has slowed, hiring cycle times have lengthened and a recent organizational change has altered the context around the affected teams. No single signal is conclusive, but the pattern is material enough to bring forward.

The system does not claim that particular employees will leave or prescribe an intervention. It explains the pattern, assembles the supporting evidence, makes uncertainty visible and routes the issue to the people accountable for evaluating it.

The shift occurs at the beginning of the chain. Intelligence is no longer limited to responding after a person frames the request. It is interpreting activity in relation to an objective and recognizing that something may deserve attention. This is anticipatory intelligence.

Anticipatory intelligence changes who must recognize the issue; it does not automatically change who makes the decision.

Recognition does not require autonomy

AI discussions often move quickly from assistance to autonomous action. That can obscure a nearer and potentially more valuable opportunity. An organization does not need to delegate a consequential workforce decision to benefit from earlier recognition.

In the example above, the system has not changed compensation, initiated an employee intervention or determined that a retention problem exists. It has reduced the chance that a material pattern remains invisible until it becomes a larger problem.

Recognition, recommendation, preparation, initiation and execution are different forms of authority. Treating them separately allows organizations to increase awareness while preserving human judgment and accountability.

The same boundary applies across a Workday customer environment. Finance can surface an unusual pattern in operating expense without changing the forecast. IT can identify a combination of access and role changes without revoking access automatically. Employee service can recognize that two policies conflict without deciding which exception should apply. In each case, earlier attention creates value before autonomy enters the design.

An organization can notice sooner without delegating the decision.

Human attention becomes the scarce resource

Organizations have spent decades overcoming information scarcity. They collect more data, create more reports and make more knowledge available. Those problems remain, but as access improves, another constraint becomes clearer: leaders and practitioners cannot review every possible signal.

Every workforce contains weak signals. A recruiting process slows in one business unit. A critical skill becomes harder to find. Internal movement changes after a reorganization. Time-off patterns shift. A project begins to drift. The relevant information may already exist, but the pattern remains unseen because no one has time to look across all of it continuously.

The answer is not a larger stream of alerts and summaries. That simply transfers the information burden into an attention burden. Useful intelligence has to distinguish ordinary variation from material change and route only the situations that merit awareness, review or decision.

This makes the quality of the organizational definition more important than the volume of the data. A longer time-to-fill may be expected during a deliberate hiring pause or consequential when it affects a role required for a growth plan. The same metric carries different meaning because the objective and context differ.

What anticipatory intelligence depends on

Earlier recognition becomes useful when intelligence understands what matters, what the signals mean and how they should be interpreted responsibly.

A model can detect a pattern without these conditions. Determining whether it matters, who should see it and what happens next requires organizational clarity.

Workday provides organizational context, not just data

This is where Workday becomes especially relevant. Worker and position records, supervisory and financial structures, roles, security, skills, compensation, business processes and effective-dated changes describe more than transactions. Together, they represent relationships through which the organization understands itself.

That context allows intelligence to interpret the same signal differently. A recruiting delay carries greater significance when it affects a position linked to a committed growth plan. A compensation change may be routine or material depending on role, geography, timing and approval authority. An employee-service response depends on worker type, jurisdiction, policy version and the effective date of the underlying change.

Workday will not contain every fact an intelligent capability needs. Relevant context may also sit in policy libraries, planning systems, learning platforms, service data, project records and the knowledge held by HR, Finance and IT. The value lies in connecting those sources to an authoritative organizational model rather than treating each as an isolated pool of information.

The same foundation establishes boundaries. Identity and security determine who may receive a signal. Organizational relationships help route it to the right owner. Effective dating separates current truth from valid history. Business-process and control definitions show where a recommendation can inform the work and where a human decision must remain explicit.

The value of Workday in an AI-enabled enterprise is not only the transactions it records. It is the organizational context it can provide.

Greater capability makes clarity more valuable

AI does not resolve ambiguity merely by interpreting more information. An outdated job profile does not become accurate because a model can summarize it. Conflicting policies do not become authoritative because a system produces a fluent answer. A loosely defined critical role does not become meaningful because it can be monitored at scale.

Previous generations of technology exposed similar problems. Databases forced organizations to define data; enterprise platforms required clearer transactions, roles, security and process. AI extends that discipline upward. It depends on clearer objectives, thresholds, decision rights and accountability because it is being asked not only to store or move information, but to interpret what the organization should care about.

This is both the risk and the opportunity. AI can give unresolved assumptions more speed, reach and apparent confidence. It can also help an organization that understands its priorities notice important change earlier and direct scarce human judgment more deliberately.

Greater capability does not reduce the need for clarity. It increases the value of having it.

From better answers to earlier recognition

Anticipatory intelligence is not a prediction of a fully autonomous enterprise. It is a practical extension of what organizations already need: earlier awareness, better context and more deliberate use of human attention.

Its value appears when a system can monitor a defined objective, recognize that several signals matter together, explain the evidence and route the issue without pretending that uncertainty has disappeared. People remain responsible for interpreting consequences, balancing trade-offs and deciding what happens next.

For leaders, the central consideration is the operating model for attention: the outcomes that merit continuous interpretation, the changes that justify interruption and the points at which accountable judgment must remain visible. For practitioners, the work is equally consequential: maintaining meaningful definitions, connecting authoritative sources, preserving security and effective dating, testing for false confidence and ensuring the capability continues to reflect the organization as it changes.

The promise is not that people stop asking questions. It is that the organization no longer depends exclusively on busy people to notice every issue before technology can help. AI can monitor, connect and surface. People continue to determine meaning, make trade-offs and accept responsibility for the response.

That introduces the next problem. Once intelligence can recognize what deserves attention, it still needs a basis for deciding what is worth preserving. Every organization is different, but not every difference creates value. AI can learn a distinctive capability and accumulated complexity with equal effectiveness.

Before organizations teach intelligence to reproduce how work happens today, they need to distinguish the differences that strengthen the enterprise from those that simply record its history.