Data

Make enterprise data, knowledge and context trustworthy, usable and ready for AI and agents.

Dimension version v1.0

Definition

Make enterprise data, knowledge and context trustworthy, usable and ready for AI and agents.

The complete document structure, including its introduction, chapters and closing sections.

—Introduction and source informationThe Data Dimension evaluates whether the enterprise can turn data, knowledge and operational context into trusted intelligence and responsible action. It focuses on the information foundations required for AI-supported decisions, automated workflows and agentic operations: the right data must be available, understandable, current, permitted and tra —From People-First Data to Agent-First Data FoundationsFigure 1 illustrates a fundamental shift in the role of enterprise data. Traditional environments are primarily designed to help people analyze the past through reports, dashboards and manually operated tools. Agentic environments must also support systems that interpret a situation, make or recommend a decision, invoke tools and participate direct —Guidance for AssessorsThe sections that follow examine the organisation. This section examines how the assessment itself should be conducted. Anti-patterns are errors of judgement and scope on the part of the assessor: they narrow the evidence base, distort findings, or lead to recommendations that are disproportionate to the decisions at stake. Red flags are conditions D.1D.1 Enterprise Data, Knowledge & Context LandscapeThis section determines whether the enterprise understands the information foundations of its most important decisions, workflows and AI-enabled services. Three things are examined together: data, which records what happened; knowledge, which explains how the enterprise interprets it; and context, which determines which of both applies at the momen D.2D.2 Data Quality, Accessibility & Operational ReadinessThis section determines whether critical information, in records, in documents and in motion, is correct enough, current enough and accessible enough to be relied upon by the workflows, decisions and AI services that consume it. Quality is contextual: an address may be complete enough for marketing but unsafe for emergency response. Readiness ther D.3D.3 Data Accountability, Privacy & ProtectionThis section determines whether the enterprise has the accountability, purpose limitations, privacy obligations, permissions, security controls and escalation paths required for responsible data use in AI-enabled operations. The section does not attempt to redesign the entire enterprise data governance model; it concentrates on the decisions and pr D.4D.4 AI and Agent-Ready Data EnablementD.4 assumes that the intended data use has been authorised under the accountability and safeguards established in D.3. It answers how permitted data and knowledge are technically prepared, versioned, retrieved, assembled and delivered reliably to AI systems and agents at the moment of decision. The section spans classical machine learning, generati D.5D.5 Data Trust, Risk & Continuous ImprovementThis section determines whether the enterprise can sustain confidence in AI data over time. Trust is not a one-time certification. It is the result of visible provenance, appropriate controls, ongoing monitoring and a closed learning loop that connects data conditions to AI behaviour, operational outcomes and business performance.—Dimension Synthesis and Transformation RoadmapThe Data Dimension should conclude with a transformation narrative, not a list of isolated findings. The synthesis has five steps: judge the evidence, place the enterprise on a maturity pattern, decide where it needs to be, sequence the work, and package the result. The takeaways at the end of each section are inputs to that narrative, not a substi —Source headers, footers and notes—Practices & playbooksThe practical companion documents linked to this section.—Version historyChanges to this document, separate from the overall framework release history.

Related practices

PlaybookGuidance

Introduction and source information

AIDIVE AI Maturıty Framework™ - Architecture & Rhythm - DATA PRACTICES & PLAYBOOK

Data

AR-D01

AR-D01 · Decision-to-Data Dependency Mapping

Expose the data, knowledge, rules and context on which a priority decision or AI-enabled workflow depends, establish what must be kept out of it, and connect each dependency to its business consequence.

Data

AR-D02

AR-D02 · Critical Asset and Ownership Mapping

Use when critical assets are orphaned, ownership is system-based, issue resolution stalls or multiple domains claim authority.

Data

AR-D03

AR-D03 · Semantic Context Mapping

Define the meanings, relationships, temporal conditions and context signals an AI-supported decision must understand.

Data

AR-D04

AR-D04 · Data Fitness-for-Decision

Set and test decision-specific quality, freshness and latency requirements linked to business impact.

Data

AR-D05

AR-D05 · Accessibility and Interoperability Assessment

Determine whether authorised human and machine consumers can obtain and interpret required information with acceptable effort.

Data

AR-D06

AR-D06 · Event and Operational Readiness

Validate that event-driven or real-time information can support safe operational action under normal and failure conditions.

Data

AR-D07

AR-D07 · Accountability and Decision Rights Design

Clarify who decides, who advises, who executes and who provides independent challenge for critical data-use decisions.

Data

AR-D08

AR-D08 · Privacy and Data-Purpose Boundary

Trace purpose, legal basis or consent, minimisation, retention and individual rights through the AI data lifecycle.

Data

AR-D09

AR-D09 · Contextual Permission Design

Design runtime permissions that account for identity, purpose, task, risk, customer relationship and transaction context.

Data

AR-D10

AR-D10 · AI Data Lifecycle Boundary

Separate and govern training, evaluation, inference and operational feedback data to prevent leakage, contamination and uncontrolled reuse.

Data

AR-D11

AR-D11 · Retrieval and Context Evaluation

Demonstrate that the right authorised and current context reaches the AI system, including cases where retrieval should return nothing.

Data

AR-D12

AR-D12 · Enterprise Memory Design

Define what an AI system may remember, why it may remember it, who can correct it and when it must expire.

Data

AR-D13

AR-D13 · Agent Data and Tool Contract

Expose enterprise data and actions to agents through bounded, interpretable and auditable business-capability interfaces.

Data

AR-D14

AR-D14 · End-to-End AI Decision Lineage

Create proportionate evidence linking user or system intent, context, policies, models, tools, actions and business outcomes.

Data

AR-D15

AR-D15 · Data Trust Monitoring and Remediation

Detect material changes in quality, freshness, access, retrieval, memory and population conditions, and in the behaviour of the services that consume them, and resolve them through an owned operational process.

Data

AR-D16

AR-D16 · Outcome-to-Data Improvement and Risk Integration

Convert AI outcomes, overrides and incidents into prioritised data improvements and visible enterprise risk treatment.

Data

AR-D17

AR-D17 · Content and Knowledge Quality Assessment

Establish whether the documents, procedures and recorded material an AI service draws on are authoritative, current and correctly scoped to where they are used.

Data

AR-R01Rhythm