Table 1 GenAI quasi-agency across DDDM... | Emerald Publishing

Table 1

GenAI quasi-agency across DDDM phases

DDDM phase	GenAI role	Dominant knowledge dynamic	Main governance concern	Governance boundary
Problem identification	Helps articulate, compare and refine problem framings	Knowledge hiding through omission-by-framing	Assumptions, alternatives or concerns may disappear once a plausible frame stabilizes	Managers retain ownership of objectives, trade-offs, success criteria and final problem definition
Information collection	Retrieves, aggregates, summarizes and organizes evidence	Knowledge coupling through evidence aggregation	Retrieval and synthesis may hide transformations, caveats, minority signals or uncertainty	Evidence must remain traceable through traceability records, auditable prompts and versioned outputs
Alternative identification	Generates scenarios, counterfactuals and candidate courses of action	Knowledge coupling through generative recombination	Generated options may create premature convergence or narrow the visible option space	GenAI proposes; managers screen, justify, include, exclude and commit
Alternative prioritization	Compares, ranks, summarizes and structures options	Knowledge decoupling between ranking support and human justification	Rankings may become implicit organizational rationales; disagreement may disappear	Humans must author the “because” section, document dissent and retain override authority
Implementation	Supports operational coordination and bounded execution	Knowledge decoupling through explicit delegation boundaries	Responsibility displacement, workflow drift or hidden discretion during execution	Execution must be logged, interruptible, reviewable and constrained by nondelegable decision classes
Results measurement	Synthesizes performance information into dashboards, summaries and evaluative categories	Knowledge coupling through shared performance representations	Generated summaries may define salience, compress uncertainty or obscure source records	Generated narratives must remain separable from source-of-truth records and traceable metrics
Results review	Organizes, retrieves and reconnects prior decisions and lessons	Knowledge coupling through learning retention and memory integration	Organizational memory may become sanitized, compressed or stripped of rationale and context	Lessons must be validated, contextualized and stored with rationales, not only metrics

DDDM phase	GenAI role	Dominant knowledge dynamic	Main governance concern	Governance boundary
Problem identification	Helps articulate, compare and refine problem framings	Knowledge hiding through omission-by-framing	Assumptions, alternatives or concerns may disappear once a plausible frame stabilizes	Managers retain ownership of objectives, trade-offs, success criteria and final problem definition
Information collection	Retrieves, aggregates, summarizes and organizes evidence	Knowledge coupling through evidence aggregation	Retrieval and synthesis may hide transformations, caveats, minority signals or uncertainty	Evidence must remain traceable through traceability records, auditable prompts and versioned outputs
Alternative identification	Generates scenarios, counterfactuals and candidate courses of action	Knowledge coupling through generative recombination	Generated options may create premature convergence or narrow the visible option space	GenAI proposes; managers screen, justify, include, exclude and commit
Alternative prioritization	Compares, ranks, summarizes and structures options	Knowledge decoupling between ranking support and human justification	Rankings may become implicit organizational rationales; disagreement may disappear	Humans must author the “because” section, document dissent and retain override authority
Implementation	Supports operational coordination and bounded execution	Knowledge decoupling through explicit delegation boundaries	Responsibility displacement, workflow drift or hidden discretion during execution	Execution must be logged, interruptible, reviewable and constrained by nondelegable decision classes
Results measurement	Synthesizes performance information into dashboards, summaries and evaluative categories	Knowledge coupling through shared performance representations	Generated summaries may define salience, compress uncertainty or obscure source records	Generated narratives must remain separable from source-of-truth records and traceable metrics
Results review	Organizes, retrieves and reconnects prior decisions and lessons	Knowledge coupling through learning retention and memory integration	Organizational memory may become sanitized, compressed or stripped of rationale and context	Lessons must be validated, contextualized and stored with rationales, not only metrics

Source(s): Own elaboration