The 2026 mandate for Chief Data Officers at JSE-listed companies

EXECUTIVE POINT OF VIEW

The global evidence is unusually consistent. Data and AI budgets are rising, Chief Data Officer remits are expanding and AI activity is accelerating. Yet the gap between ambition and enterprise value remains wide.

The constraint is generally not the absence of technology. It is the operating discipline required to connect trusted data, accountable decisions, redesigned workflows, adoption and measurable economics.

For Chief Data Officers at JSE-listed companies, this gap is particularly consequential. King V brings board oversight of data, information and technology directly into the governance of strategy. POPIA makes lawful, transparent and well-controlled use of personal information non-negotiable. At the same time, South African employees are using AI faster than many organisations can train, govern or assure it.

The CDO therefore sits at the intersection of value, risk, operating-model change and workforce behaviour—whether or not the formal job description says so.

Our central view is this:

The winning JSE CDO will be judged less by how much data is governed than by how many consequential decisions become faster, better and safer—and whether those gains appear in business performance.

AN IMPORTANT RESEARCH QUALIFICATION

For this review, the “big five” refers to Deloitte, EY, KPMG, PwC and Accenture.

The research is not symmetrical.

Deloitte, EY and KPMG have published recent standalone CDO-focused studies. The latest relevant PwC and Accenture studies are broader enterprise and AI surveys involving senior business, data and technology executives.

KPMG’s latest standalone CDO survey identified in this review is an India study from 2023. No recent standalone South African Big Five CDO survey was located.

The South African analysis therefore uses local AI adoption research, global studies that included South African executives, and the country’s governance and regulatory context.

PwC and Accenture findings are treated as adjacent execution evidence and are not presented as CDO-only survey results.

WHAT THE RESEARCH SAYS

1. DELOITTE — CHIEF DATA OFFICER SURVEY 2024

Deloitte’s research found that:

• 49% of CDOs prioritised AI and GenAI.
• 42% prioritised building the business case.
• 35% prioritised data governance.
• 35% prioritised insights and analytics.
• Although 72% reported into the C-suite, only 3% reported directly to the CEO.

The study also found that 66% of CDOs had enabled process efficiency and regulatory or legal compliance, while 63% had improved strategic decision-making.

The implication is that CDO expectations are increasingly enterprise-wide, but formal authority, sponsorship and access to the CEO may not match the mandate.

For JSE-listed companies, this raises an important question: is the CDO being positioned to influence strategy and business performance, or primarily to manage platforms, governance and regulatory obligations?

2. DELOITTE — FEDERAL CDO SURVEY 2025

Deloitte’s sixth annual United States Federal CDO Survey found that:

• AI use increased from 67% to 78%.
• 64% of CDOs were deeply involved in AI data-governance policy.
• 30% also held the role of Chief AI Officer.
• 96% collaborated with AI leaders monthly.
• 70% collaborated with AI leaders weekly.
• 57% operated with five or fewer full-time employees.

Although this is a public-sector study, it provides a useful governance comparison. It highlights the increasing convergence of data and AI leadership, while also exposing the capacity problem confronting many CDOs.

The mandate may be enterprise-wide, but the resources available to fulfil it are frequently small.

3. EY — CHIEF DATA OFFICER STUDY 2026

EY’s second global Chief Data Officer Study found that:

• More than 80% of respondents reported data-office budget growth.
• The average data-office team grew by 19%.
• 86% expected budgets to grow further.
• 61% reported full authority over data.
• Only 33% reported full authority over AI.
• More than 90% use AI in some part of data management.
• Only 25% have integrated AI deeply into data management.

EY also found that the strongest organisations were more than four times as likely as the weakest to have trusted measurement frameworks integrated with the business.

The study captures the central issue succinctly:

“Organizations winning with AI aren’t winning because of their models—they are winning because their data is better, more trusted…”

The CDO role is becoming more strategic, but strategy alone does not differentiate leaders. Deployment, measurement, operating discipline and influence do.

4. KPMG — CDO OUTLOOK 2023

KPMG’s study of more than 75 CDOs, CDAOs and data-office leaders in India found that:

• More than 75% focused on data strategy and scaling through business collaboration.
• Fewer than 25% had begun a data-value or monetisation journey.
• Three in five experienced budget challenges.
• 53% cited change-management barriers to governance adoption.
• 67% wanted automated data catalogues or marketplaces.
• 62% intended investing in GenAI and large language models, accompanied by appropriate ethical guardrails.

The study provides a useful emerging-market comparison.

It also illustrates an important distinction: organisations may have sophisticated data strategies while remaining immature in the disciplines needed to convert those strategies into measurable value.

5. PwC — AI METRIC SURVEY 2026

PwC’s survey of 70 senior business and technology leaders found that AI expenditure was distributed as follows:

• 62% to technology.
• 34% to process redesign and change.
• Only 4% to workforce training.

PwC recommends that organisations measure AI across financial, operational, workforce and trust dimensions.

The expenditure pattern is revealing. Many organisations continue to invest heavily in technology while underinvesting in the people, process and operating-model changes required to produce value.

For CDOs, this means the AI portfolio needs more than technical performance measures. It needs an economic and trust scorecard.

PwC’s sample was small and predominantly United States-based. Its reported relationships between AI investment and financial performance should be interpreted as associations rather than proof that spending alone caused superior results.

6. ACCENTURE — PULSE OF CHANGE 2025

Accenture surveyed 3,000 C-suite executives across 18 countries, including South Africa.

The study found that:

• 85% planned to increase AI investment.
• 67% expected AI to drive revenue growth.
• 86% said they were preparing their workforces for AI.
• 75% nevertheless said the speed of change was outpacing training.
• 63% were investing in AI agents.
• Only 27% were integrating agents across functions.
• Just 20% were rebuilding processes around AI.

This suggests that many organisations are inserting AI into existing workflows rather than redesigning work around the new capability.

For CDOs, the implication is clear: AI value depends on process reinvention and workforce readiness, not simply technology deployment.

7. ACCENTURE — MAKING REINVENTION REAL WITH GENERATIVE AI

Accenture’s analysis of more than 2,000 projects and 3,000 C-level respondents found that:

• Only 36% of organisations had scaled GenAI.
• Just 13% reported significant enterprise value.
• Value leaders were 2.9 times more likely to have comprehensive data strategies.
• They were 2.7 times more likely to have responsible-AI principles and governance.
• Organisations combining the full set of enabling actions were 2.5 times more likely to report enterprise-level results.

Trusted data, responsible AI, executive sponsorship and process redesign are not supporting activities. They are enterprise-value enablers.

SEVEN NUMBERS THAT SHOULD CONCERN CDOs

• 49%: Deloitte CDO respondents prioritising AI and GenAI.
• More than 80%: EY respondents whose data-office budgets increased.
• 25%: EY respondents with AI deeply integrated into data management.
• Fewer than 25%: KPMG India CDO respondents that had begun a value or monetisation journey.
• 20%: Leaders in Deloitte’s 2026 enterprise AI research reporting AI-driven revenue gains.
• 4%: Workforce training’s share of AI expenditure in PwC’s benchmark.
• 63%: South African respondents who had relied on AI output without evaluating its accuracy.

THE COMMON SIGNAL: STRATEGY IS NOT THE DIFFERENTIATOR

Most large organisations already have data strategies, AI strategies, cloud platforms, governance policies and lengthy lists of use cases.

The surveys increasingly separate leaders from laggards on execution:

• Deployment into real workflows.
• Clear ownership and decision rights.
• Trusted measurement.
• Process redesign.
• Adoption and change capacity.
• The connection between use cases and business economics.

The real divide is not between companies that have an AI strategy and those that do not.

It is between companies that can convert strategy into changed decisions and those that remain trapped in pilots, platforms and presentations.

SIX SHIFTS DEFINING THE JSE CDO MANDATE

1. FROM DATA STRATEGY TO AN ECONOMIC CONTRACT WITH THE BUSINESS

The CDO needs an explicit contract with each business domain covering:

• The decision or outcome to be improved.
• The economics at stake.
• The accountable business executive.
• The required data, process and control changes.
• The adoption path.
• The method for validating benefits.

“Single customer view”, “AI enablement” and “modern data platform” are not business outcomes.

Reduced churn, better credit decisions, lower claims leakage, faster asset turnaround, improved forecast accuracy and shorter working-capital cycles are outcomes.

This does not mean every dataset must be separately monetised. It means the value logic must be explicit enough for Exco to allocate capital and hold executives accountable.

2. FROM AI PILOTS TO A PRODUCTION-GRADE DATA FOUNDATION

AI has created a sharper version of an existing problem: models can move faster than the organisation’s ability to assure the data, definitions, ownership and permitted uses behind them.

Although more than 90% of EY’s respondents use AI in some part of data management, only 25% have integrated it deeply.

Governance, quality and metadata are receiving more investment because they are prerequisites for safe scale.

The answer for diversified JSE groups is not necessarily a perfect enterprise data model. It is a deliberately sequenced foundation consisting of:

• Authoritative master data in the domains causing the greatest cross-business friction.
• Fit-for-purpose data-quality thresholds.
• Critical data lineage.
• Common business definitions.
• Identity and access controls.
• Reusable data products supporting high-value decisions.
• Clear ownership and service levels.

This is an important area for Emergent Africa.

Our Master Data Management as a Service proposition can help organisations create accountable and continuously improving customer, product, supplier or asset data without waiting for a multi-year enterprise programme.

3. FROM GOVERNANCE POLICY TO A LIVE CONTROL SYSTEM

Governance must travel at the speed of the use case.

A static policy library cannot manage the risks arising from:

• Rapid AI adoption.
• Third-party models.
• Sensitive personal information.
• Automated decisions.
• Embedded AI agents.
• Unapproved employee use.
• Models that change after deployment.

Effective governance is risk-tiered and embedded in delivery.

It includes permitted-use rules, model and data inventories, approval gates, privacy and bias checks, human oversight, incident escalation, audit evidence and continuous performance monitoring.

The KPMG South Africa AI snapshot found that:

• 63% had relied on AI output without evaluating its accuracy.
• 56% had made mistakes because of AI.
• 48% had used AI contrary to organisational policy.
• 72% said their organisations had an AI strategy.
• Only 53% said responsible-AI training was provided.

A strategy can therefore exist on paper while the real operating model is being shaped by ungoverned employee behaviour.

4. FROM FORMAL AUTHORITY TO ENTERPRISE INFLUENCE

The CDO rarely controls every variable needed to produce value.

Data platforms may sit with the CIO. Models may sit with a CAIO or analytics function. Privacy resides with the Information Officer and Legal. Process ownership remains with business executives. Risk appetite belongs to the board. Adoption depends on line leaders.

EY found that 61% of CDOs had full authority over data, compared with only 33% for AI.

Even CDOs with full data authority frequently shared responsibility for AI.

Influence is therefore not a soft skill. It is part of the CDO’s operating capability.

The practical response is explicit role design:

• Who sets enterprise data and AI standards?
• Who owns source data?
• Who funds shared foundations?
• Who accepts model risk?
• Who owns the affected business process?
• Who signs off benefit realisation?
• Who is accountable when an automated decision causes harm?

A small number of cross-functional forums should have real decisions to make. Another layer of reporting will not close the accountability gap.

5. FROM A SPECIALIST TEAM TO A FEDERATED ENTERPRISE CAPABILITY

Scarce skills and constrained budgets make a purely centralised model unrealistic. Uncontrolled federation, however, makes standards and trust brittle.

A stronger model is a strong centre with accountable domains.

The enterprise data office owns policy, architecture guardrails, platforms, assurance, shared services and portfolio visibility.

Business domains own priorities, stewardship, process adoption and business outcomes.

Technology teams provide reliable, scalable delivery. Risk and control functions provide proportionate oversight. The board and Exco establish appetite and accountability.

Data and AI literacy must also be role-specific.

Board members need to interrogate value, risk and accountability.

Executives need to sponsor decisions and redesign processes.

Product and operational leaders must understand when human judgement overrides an algorithm.

Data stewards need authority, time and service levels.

Front-line employees need simple rules for safe use and escalation.

Generic awareness training will not produce these behaviours.

6. FROM DASHBOARDS TO DECISION INTELLIGENCE

The end-state is not more reporting.

It is an organisation in which critical recurring decisions are deliberately designed:

• What decision needs to be made?
• Who makes it?
• At what cadence?
• Using which evidence?
• Subject to which guardrails?
• Embedded in which workflow?
• Measured against which business outcome?

Decision intelligence is the bridge between analytics or AI output and realised enterprise value.

It also changes the CDO’s success measures.

Alongside model accuracy and data quality, the CDO needs to measure:

• Whether the insight was used.
• Whether decision cycle time improved.
• Whether exceptions were handled safely.
• Whether the target business metric changed.
• Whether the benefit persisted.
• Whether users trusted and adopted the changed process.

This is how the Data Office becomes a Value Office without weakening its stewardship responsibilities.

WHY THE MANDATE IS SHARPER FOR JSE-LISTED COMPANIES

BOARD ACCOUNTABILITY IS BECOMING MORE EXPLICIT

King V’s Principle 10 states:

“The board governs data, information and technology in a way that enables the organisation to sustain and optimise its strategy and objectives.”

King V applies to financial years beginning on or after 1 January 2026, while revised JSE Listings Requirements came into effect during February 2026.

The governance context is therefore immediate for listed issuers.

The CDO should equip the board with a concise view of:

• Business value being generated.
• Principal data and AI risks.
• Critical dependencies.
• Material exceptions and incidents.
• Management actions.
• The quality of evidence supporting reported benefits.

The board does not need a catalogue of data activity.

PRIVACY AND AUTOMATED DECISIONS REQUIRE JOINED-UP ACCOUNTABILITY

POPIA gives the Information Officer formal responsibilities for the compliance framework, personal-information impact assessments, policies, controls and awareness.

The CDO should not absorb statutory accountability by assumption.

Instead, the CDO, Information Officer, CIO, CAIO, Legal, Risk and business owner need a designed control chain from data sourcing through model use, decision, customer impact and evidence retention.

This is especially important where personal information is used for profiling or automated decisions.

SOUTH AFRICA’S AI ADOPTION GAP IS A MANAGEMENT-SYSTEM PROBLEM

The KPMG South African evidence shows high expectations of AI alongside frequent unverified and non-compliant use.

This combination can produce:

• Privacy risk.
• Conduct risk.
• Cybersecurity exposure.
• Incorrect operational decisions.
• Reputational damage.
• Weak auditability.
• Inaccurate internal or external reporting.

Heavy central restriction is unlikely to succeed.

Safe behaviour must be easier than unsafe behaviour through approved tools, practical rules, contextual training, monitoring and accessible escalation.

CAPITAL DISCIPLINE FAVOURS SHORT, PROVABLE VALUE CYCLES

South African CDOs should avoid positioning the data foundation as an open-ended modernisation programme.

A 90-day cycle can:

1. 1. Establish a performance baseline.
   2. Select two or three important decisions.
   3. Resolve the most material data constraints.
   4. Embed an intervention in the workflow.
   5. Measure adoption.
   6. Validate the initial value signal.
   7. Use evidence to inform the next funding decision.

This gradually improves the data estate while building organisational confidence and financial credibility.

A 180-DAY MANDATE RESET

DAYS 0–30: MANDATE AND BASELINE

Management actions:

• Agree the CDO mandate with the CEO and Exco.
• Map responsibilities across the CDO, CIO, CAIO, Information Officer and business executives.
• Baseline current data, AI, governance, adoption and value measures.
• Identify the ten decisions with the greatest economic or risk exposure.

Evidence of progress:

• A signed CDO mandate.
• A clear role and decision-right map.
• One enterprise data and AI scorecard.
• A prioritised decision portfolio.
• A stop, start and continue list.

DAYS 31–60: VALUE PORTFOLIO

Management actions:

• Select two or three lighthouse decisions.
• Build outcome cases with accountable business executives.
• Identify critical data, MDM, process, model, privacy and change dependencies.
• Define adoption and benefit measurement before development begins.

Evidence of progress:

• Named business owners.
• Quantified performance baselines.
• Delivery and control plans.
• An agreed benefit-measurement method.
• A data-foundation backlog linked directly to business use cases.

DAYS 61–90: DEPLOY AND LEARN

Management actions:

• Deliver the smallest end-to-end workflow change capable of improving a decision.
• Establish risk-tiered controls and human oversight.
• Train the roles affected by the change.
• Track use, overrides, incidents, cycle time and business outcomes.

Evidence of progress:

• Live use in an operational workflow.
• Adoption and usage data.
• Control and assurance evidence.
• An initial business-value signal.
• Documented lessons and remediation actions.

DAYS 91–180: SCALE THE OPERATING SYSTEM

Management actions:

• Expand reusable data products and shared services.
• Formalise domain stewardship and service levels.
• Reallocate funding using delivery and value evidence.
• Report value and risk to Exco and the board.
• Retire unsuccessful pilots and duplicative assets.

Evidence of progress:

• A repeatable delivery model.
• More trusted master data.
• Active federated stewardship.
• Visible portfolio economics.
• A board-ready value-and-risk narrative.

THE BOARD AND EXCO SCORECARD

A useful scorecard should fit on one page and cover five dimensions:

• Value: Validated revenue, margin, cash, cost or loss-avoidance outcomes, including the benefit owner and confidence level.
• Decision adoption: Workflow usage, time to decision, override patterns and user behaviour.
• Trust: Critical-data quality, lineage coverage, model performance, exceptions and incidents.
• Delivery: Time from prioritisation to live use, reuse of data products and retirement of duplicative assets.
• Capability: Domain ownership, stewardship service levels, critical-role proficiency and change saturation.

SIX QUESTIONS EVERY JSE CDO SHOULD PUT TO EXCO

1. 1. Which five recurring decisions matter most to our strategy and enterprise value—and where does poor data materially weaken them?
   2. Who owns the business outcome, not merely the platform or model?
   3. What must be true of data quality, permitted use, human oversight and process design before an AI use case can scale?
   4. Which shared data foundations are justified by the value portfolio, and which are expensive aspirations without a demand signal?
   5. Where are employees already using AI outside approved workflows, and how will we make safe behaviour easier?
   6. What evidence will we show the board next quarter that value is increasing while risk remains controlled?

WHERE EMERGENT AFRICA CAN ASSIST

Emergent Africa can help CDOs connect specialist data disciplines with strategy, operating-model change and value realisation.

Our proposition brings together local executive context and capabilities in strategy execution, decision intelligence, transformation, governance, performance improvement and master data management.

1. UNCLEAR MANDATE OR FRAGMENTED SPONSORSHIP

Emergent Africa intervention:

A CDO mandate reset incorporating executive interviews, a maturity and value diagnostic, decision-right design and Exco alignment.

Practical output:

A signed mandate, federated operating model, fit-for-purpose governance forums and a practical 180-day action plan.

2. WEAK BUSINESS CASE OR CROWDED USE-CASE BACKLOG

Emergent Africa intervention:

Value-portfolio design that maps important decisions, quantifies the economics, sequences dependencies and establishes benefits governance.

Practical output:

A fundable portfolio linked to growth, margin, cash, risk, resilience or customer outcomes.

3. INCONSISTENT CUSTOMER, PRODUCT, SUPPLIER OR ASSET DATA

Emergent Africa intervention:

Master Data Management as a Service covering multi-domain design, stewardship, quality, workflows, service levels and continuous improvement.

Practical output:

More trusted master data, visible accountability and a 90-day improvement rhythm.

4. ANALYTICS AND AI NOT EMBEDDED IN DECISIONS

Emergent Africa intervention:

Decision-intelligence transformation incorporating decision mapping, domain analytics, data-product design, workflow integration and adoption measurement.

Practical output:

Live decision interventions with measurable usage, control and business impact.

5. GOVERNANCE THAT IS POLICY-HEAVY BUT OPERATIONALLY WEAK

Emergent Africa intervention:

Risk-tiered data and AI governance covering decision forums, approval paths, data and model inventories, assurance evidence and board reporting.

Practical output:

Governance that enables delivery and innovation while protecting privacy, security, conduct and accountability.

6. SKILLS AND CHANGE CAPACITY CONSTRAINING SCALE

Emergent Africa intervention:

Role-based capability building, executive alignment, stewardship activation, change networks and a performance-management cadence.

Practical output:

A federated capability that business leaders can use and sustain, with less dependence on a central specialist team.

A PRAGMATIC STARTING OFFER

CDO MANDATE-TO-VALUE SPRINT

In four to six weeks, Emergent Africa can help a CDO build:

• An evidence-based mandate.
• A decision-right map.
• A prioritised value portfolio.
• A data and AI governance agenda.
• A board and Exco scorecard.
• A 180-day delivery plan.
• Two or three lighthouse decisions around which to organise the first value cycle.

The intent is not a generic maturity assessment.

It is a working management agenda that tells the CEO, Exco and board what will change, who owns it, what it will cost, how risk will be controlled and how value will be evidenced.

CONCLUSION

The CDO role is gaining budget, scope and strategic relevance. That does not automatically produce influence or value.

The next phase belongs to CDOs who can improve the enterprise’s most consequential decisions while maintaining trust, accountability and control.

For JSE-listed companies, this is not simply a technology programme.

It is an enterprise management system spanning the board, Exco, business domains, data and technology teams, risk functions and the workforce.

Emergent Africa is positioned to help build that system: aligning strategy and execution, establishing the operating model, strengthening trusted master data, embedding decision intelligence and managing the transformation needed for benefits to persist.

SOURCES

Deloitte — Chief Data Officer Survey 2024
https://www.deloitte.com/nl/en/services/consulting-risk/research/chief-data-officer-survey.html

Deloitte — 2025 Federal Chief Data Officer Survey
https://www.deloitte.com/us/en/Industries/government-public/perspectives/federal-cdo-survey-2026.html

EY — Chief Data Officer Study, Second Edition, 2026
https://www.ey.com/content/dam/ey-unified-site/ey-com/es-es/insights/consulting/documents/ey-study-cdo-2026.pdf

KPMG India — CDO Outlook 2023
https://assets.kpmg.com/content/dam/kpmgsites/in/pdf/2023/10/kpmg-cdo-outlook.pdf

KPMG and the University of Melbourne — Trust, Attitudes and Use of AI: South Africa Snapshot 2025
https://assets.kpmg.com/content/dam/kpmgsites/xx/pdf/2025/05/trust-attitudes-and-use-of-ai-south-africa-snapshot.pdf

PwC — AI Metric Survey 2026
https://www.pwc.com/us/en/services/ai/ai-benchmarking-enterprise-decision-advantage.html

Accenture — Pulse of Change, September 2025
https://www.accenture.com/za-en/insights/pulse-of-change-september-2025

Accenture — Making Reinvention Real with GenAI
https://www.accenture.com/za-en/insights/consulting/making-reinvention-real-with-gen-ai

Deloitte — State of AI in the Enterprise 2026
https://www.deloitte.com/za/en/issues/generative-ai/state-of-ai-in-enterprise.html

Institute of Directors South Africa — King V Application Guidance
https://cdn.ymaws.com/www.iodsa.co.za/resource/collection/CFA6E544-15F6-4FE2-900F-872F39A21B52/Medical_Schemes_Application_of_King_V.pdf

Information Regulator South Africa — POPIA
https://inforegulator.org.za/popia/

Department of Communications and Digital Technologies — South Africa National AI Policy Framework
https://www.dcdt.gov.za/sa-national-ai-policy-framework/file/338-sa-national-ai-policy-framework

Werksmans — Governance Reforms for Listed Companies
https://werksmans.com/governance-reforms-for-listed-companies/

Emergent Africa — Decision Intelligence
https://emergent.africa/emergent-decision-intelligence/

Emergent Africa — How the Chief Data Officer’s Master Data Management Mandate Is Changing in 2026
https://emergent.africa/decision-intelligence/how-the-chief-data-officers-master-data-management-mandate-is-changing-in-2026/