Deloitte created the Chief Intelligence Officer role on 19 May 2025 — twelve months ago this week — when Joseph Ucuzoglu, Global CEO, announced the appointment of Ranjit DeSouza to the position from his prior role running the firm's US Strategy and Analytics practice. The CIO role was the first Big Four executive appointment to carry "Intelligence" rather than "AI" in its title, a deliberate framing that DeSouza pushed in the appointment announcement on the basis that the AI category was, by mid-2025, too narrow to describe the operational change Deloitte was building toward. The twelve-month review of the role — conducted internally by Deloitte's Office of the CEO in March and April 2026 and shared with the global executive committee at its May 2026 meeting in Lake Como — concludes that the CIO function has met its primary objectives, exceeded two of its three secondary objectives, and missed one of its operational targets by a margin large enough that the executive committee has authorised a structural reorganisation that the review proposes for the second half of 2026. The conclusions matter beyond Deloitte. The CIO role was the template against which EY, KPMG, and PwC built their equivalent functions in late 2025; the twelve-month review is the first multi-quarter operating data on the role across any Big Four firm. The data is granular enough to inform the trajectory of the broader professional-services category. The patterns are clear enough to draw structural conclusions. The review is the document the broader consulting industry will read most carefully this calendar year.
Practitioner Augmentation: the program that drove the consultant-utilisation shift
The Practitioner Augmentation program — the largest of the four named programs by budget — has produced the most measurable operational effect across Deloitte's twelve-month review period. The program's core deliverable is a unified AI workbench, internally called "Deloitte Quant," that integrates Claude Sonnet 4.6 (the firm's primary frontier-model vendor under the September 2025 framework agreement with Anthropic), Microsoft 365 Copilot (the firm's primary productivity-AI surface under the existing Microsoft enterprise agreement), and a series of internal retrieval-augmented capabilities that ground against the firm's proprietary methodology archive. The workbench was the first production deliverable shipped under DeSouza's tenure, with the initial release in December 2025 and an enterprise-wide rollout that completed in March 2026. By the time the twelve-month review was conducted, 87 per cent of Deloitte's global practitioner base had been provisioned access to the workbench, with active daily-usage rates running at approximately 64 per cent of provisioned users.
The consultant-utilisation curve is the metric the executive committee has examined most carefully. Deloitte's pre-deployment utilisation rate — the percentage of practitioner billable hours actually billed against client engagements, the foundational productivity metric in professional services — ran at approximately 73 per cent across the firm in fiscal 2025. The twelve-month review measured the post-deployment utilisation rate at approximately 71 per cent across the firm in the trailing six months of the review period, with the difference distributed unevenly across practice areas. The headline reading is that utilisation has dropped two percentage points. The reading the review proposes is structurally different. The 2-point drop is a composite of three offsetting effects: a 4-point drop in pre-shipment time-on-engagement (practitioners are spending less time on each engagement because the AI workbench accelerates the work), partially offset by a 2-point increase in total engagement count (practitioners are taking on more engagements because each engagement requires less time), and a smaller residual drop attributable to onboarding-and-training time that the workbench rollout itself consumed in the second half of fiscal 2025.
The revenue implication is the metric DeSouza presented to the executive committee as the operational justification for the program. The 2-point utilisation drop multiplied by Deloitte's $68.6B revenue base would, in a naive read, represent approximately $1.4B in annualised revenue impact — a number that would dominate any conversation about the program's ROI. The composite read produces a different number. The 2-point increase in engagement count, scaled against the average revenue per engagement, contributes approximately $1.8B of incremental annualised revenue. The 4-point reduction in pre-shipment time-on-engagement, when properly attributed to engagement-margin improvement, contributes approximately $1.1B of incremental annualised gross margin. The net effect — incremental revenue plus incremental margin minus the workbench's operating cost — is approximately $2.3B of annualised value, against a program operating cost of approximately $480M. The roughly 4.8x ROI is the operational result that the executive committee approved in the May 2026 Como meeting and that informs the budget allocation for the second twelve months of the program.
We measure the consultant-utilisation curve at the composite level, not at the per-practitioner level. The composite reveals what the per-practitioner number conceals.
Sector Verticalisation: the program that has not yet delivered
The Sector Verticalisation program is the one operational target the twelve-month review concludes Deloitte has missed by a margin large enough to require structural reorganisation. The program's design intent was to ship sector-specific AI tooling for Financial Services, Life Sciences, Energy, Consumer, and Public Sector — the five verticals that Deloitte's global executive committee designated as priority sectors in the original CIO appointment briefing. The program has shipped sector-specific tooling for Financial Services (the most mature of the five verticals, with Almeida himself coming from that practice) and for Life Sciences (the second-most-mature, where the regulated nature of pharmaceutical client engagements created the strongest internal demand signal), but has not shipped general-availability tooling for Energy, Consumer, or Public Sector across the twelve-month period. The slippage against the original roadmap is the principal operational miss in the review document.
The slippage has three causes that the review document analyses in detail. The first is a vendor-integration cause: the Sector Verticalisation program's architecture depends on the same Claude Sonnet 4.6 inference layer that Practitioner Augmentation uses, but the sector-specific tooling requires per-vertical context retrieval against the firm's proprietary sector-research archives. The retrieval architecture, contracted with Cohere for the embedding layer and built in-house for the orchestration plane, took approximately five months longer to ship than the program plan anticipated. The delay was concentrated in the integration work required to make the retrieval architecture queryable from the practitioner-facing Deloitte Quant workbench in real time. The second cause is a content-quality cause: the firm's proprietary sector-research archives, which the retrieval architecture grounds against, were inconsistently digitised across the five verticals. The Financial Services and Life Sciences archives had been digitised and indexed in 2023 and 2024 as part of the firm's broader research-modernisation program; the Energy, Consumer, and Public Sector archives required substantial digitisation work that the program had budgeted but had not properly sequenced.
The third cause — and the one the review document treats with the most scrutiny — is an organisational-design cause. The Sector Verticalisation program was structured as a horizontal function reporting into DeSouza's organisation, with dotted-line accountability to the vertical practice leaders. The horizontal-vertical accountability structure produced what the review calls "tension-resolution friction": when the Sector Verticalisation team's priorities conflicted with the vertical practice leaders' priorities, the conflict resolution defaulted to the practice leaders' near-term P&L considerations rather than the firm-wide capability-building considerations the program was designed to prioritise. The friction has been visible in the program metrics — the Energy and Consumer verticals, in particular, have had repeated delays attributable to practice-leader-driven re-scoping — and has been visible in the program's own staffing churn, with three of the program's five vertical leads having rotated out of the program in the twelve-month period.
The structural reorganisation that the review proposes — and that the executive committee approved at the Como meeting — addresses the third cause directly. The Sector Verticalisation program will be restructured for the second half of 2026 to give DeSouza direct authority over the per-vertical sector leads, with the practice-leader involvement formalised as advisory rather than co-accountable. The restructure is the principal organisational change coming out of the twelve-month review and is the explicit acknowledgement that the original organisational design under-estimated the friction between horizontal capability-building and vertical P&L management. The implication for the rest of the Big Four is structural: EY, KPMG, and PwC built their equivalent CIO functions with co-accountable horizontal-vertical structures broadly modelled on Deloitte's original design. The Deloitte restructure will be read by the other Big Four firms as an early-mover signal about the operational architecture that the role's first twelve months has selected for.
Audit-Tech and Tax-Tech: the regulated practices
The Audit-Tech and Tax-Tech programs operate under regulatory constraints that the Practitioner Augmentation and Sector Verticalisation programs do not. The PCAOB has been actively engaged with the Big Four firms on AI deployment inside audit practice since 2024, and the equivalent regulators in Deloitte's international jurisdictions — the UK FRC, the German Apas, the Dutch AFM, the Japanese FSA — have been similarly engaged at varying levels of formality. The Audit-Tech program has been the most regulator-facing of DeSouza's four programs, and Whitfield's tenure has been disproportionately occupied with regulator engagement rather than tooling shipment. The twelve-month review's conclusions on Audit-Tech are accordingly more nuanced than its conclusions on Practitioner Augmentation: the program has shipped a meaningful set of capabilities, but the shipping cadence has been governed by regulator-engagement timelines rather than the program's own engineering capacity.
The capabilities that have shipped under Audit-Tech are concentrated in two categories. The first is workpaper-preparation automation — the AI-augmented assembly of audit workpapers from the underlying client data, the firm's audit methodology templates, and the relevant accounting standards. The capability has shipped to general availability across Deloitte's US and UK audit practices and is in active deployment for approximately 64 per cent of Deloitte's audit engagements in those geographies. The second is anomaly-detection assistance — the AI-augmented identification of transactions, journal entries, or financial-statement line items that warrant further audit attention. The capability operates in advisory mode only — the AI flags potential anomalies for human-auditor review but does not make any audit-conclusion recommendation — and has shipped to general availability across the same US and UK practices.
The PCAOB's posture toward Audit-Tech has been broadly supportive but bounded. The board's December 2025 staff guidance on AI-augmented audit work — published after extended engagement with all four Big Four firms — established that AI tools may be used in audit work provided that auditor judgement remains the controlling input on every audit conclusion, that the AI tool's outputs are appropriately documented in the workpapers, and that the firm maintains a documented evaluation of the tool's reliability against the specific audit assertion it supports. The guidance is operationally workable, but it imposes documentation requirements that have slowed the shipping cadence relative to the unregulated Practitioner Augmentation program. The twelve-month review concludes that the Audit-Tech program is on track against the regulator-adjusted roadmap but behind the original roadmap, with the variance attributable to the documentation burden that the PCAOB guidance imposes.
The Tax-Tech program has produced more measurable shipping progress than Audit-Tech because the tax practice operates under less continuous regulator engagement than the audit practice. Olausson's program has shipped capabilities across direct-tax compliance (the AI-augmented preparation of corporate tax filings), indirect-tax automation (the AI-augmented preparation of VAT and sales-tax returns), and transfer-pricing analysis (the AI-augmented preparation of transfer-pricing documentation for multinational client engagements). The shipping cadence has tracked the original roadmap closely, and the program has produced what DeSouza described to the executive committee as the cleanest demonstration of AI's operational value across the four programs: a 38 per cent reduction in preparation time for direct-tax filings, a 41 per cent reduction in preparation time for indirect-tax returns, and a 47 per cent reduction in preparation time for transfer-pricing documentation. The percentages are higher than Practitioner Augmentation's equivalent figures because the tax-preparation workstream is more structured and more amenable to AI augmentation than the broader practitioner-engagement workstream.
The partnership matrix: Anthropic, OpenAI, Microsoft, and the dual-frontier posture
Deloitte's vendor partnership matrix in May 2026 is the operational expression of DeSouza's procurement strategy and reflects the broader Fortune 500 dual-vendor pattern noted in the procurement aggregates. The firm signed a framework agreement with Anthropic in September 2025 that named Claude Sonnet 4.6 as the firm's primary frontier-model vendor, with reserved-capacity guarantees, custom rate limits, and a price floor against Anthropic's published rate changes for the contracted volume tier. The agreement covers all four of the named programs and is the principal commercial relationship under DeSouza's organisation. The contract terms were negotiated by DeSouza's procurement function over approximately four months in the summer of 2025 and reflect the firm's leverage as one of Anthropic's largest enterprise customers globally.
The OpenAI relationship is structured differently. Deloitte signed a non-exclusive framework agreement with OpenAI in November 2025 — two months after the Anthropic agreement — that names GPT-5 and o3-pro as approved frontier-model vendors for specific use-case categories that the firm has identified as best-fit for OpenAI's capability matrix. The use-case categories are narrower than the Anthropic agreement covers: open-ended Q&A surfaces inside Deloitte Quant, the long-context document-analysis surface used in the firm's M&A advisory practice, and a defined set of code-generation surfaces inside the firm's engineering-services practice. The OpenAI relationship is the firm's dual-vendor counterweight to Anthropic and the structural mechanism by which DeSouza preserves procurement leverage across the frontier-model layer. The two contracts together cover the bulk of the firm's inference spend, with the remainder distributed across smaller vendor relationships for specialised capabilities.
The Microsoft relationship is the third structural partnership and operates at a different layer of the AI stack. Deloitte's existing global Microsoft Enterprise Agreement, renewed in 2024 with a five-year term, includes the productivity-AI surface (Microsoft 365 Copilot) that integrates into the firm's existing Microsoft 365 deployment across all 415,000 practitioners. The Copilot relationship is operationally distinct from the Anthropic and OpenAI relationships: Copilot is the integrated-productivity-AI surface for email, document creation, and meeting transcription, while Claude and GPT-5 are the workbench-integrated AI surfaces for client-engagement work. The three relationships, taken together, give Deloitte coverage across the productivity, frontier-reasoning, and specialised-task surfaces. The partnership matrix is the procurement-architecture template that the other Big Four firms have studied closely, and the May 2026 twelve-month review documents the structure with the level of detail that suggests the firm intends the structure to be replicable across its peer group.
The supporting partnership matrix — the vendors below the principal three — has been operationalised through a tiered-vendor framework that DeSouza's procurement function established in late 2025. The framework defines three tiers of AI vendor relationship: principal-tier (Anthropic, OpenAI, Microsoft), secondary-tier (Cohere for embedding, Voyage AI for specialised retrieval, NVIDIA's enterprise inference platform for the firm's on-premises deployments), and tertiary-tier (the longer tail of specialised vendors that the firm engages on a use-case-specific basis). The tiered framework allows the procurement function to apply differentiated negotiation discipline at each tier: principal-tier relationships are governed by framework agreements with reserved capacity and price floors; secondary-tier relationships are governed by standard enterprise agreements with annual renewal cycles; tertiary-tier relationships are governed by per-use-case purchase orders that the procurement function can authorise without executive-committee review. The tiered discipline has been operationally important as the AI vendor landscape has fragmented, and is the structural mechanism by which the firm has managed the vendor-count compression that the broader Fortune 500 has experienced over the trailing twelve months.
The vendor-relationship management discipline that supports the partnership matrix has been the operational backbone of the program's commercial performance. DeSouza's procurement function conducts a structured quarterly business review with each principal-tier vendor, with the review covering the trailing-quarter consumption against the contracted commitments, the operational reliability metrics across the production deployments, the roadmap alignment between the vendor's product roadmap and Deloitte's deployment requirements, and the commercial-terms calibration against any market-clearing changes. The quarterly business review discipline is unusual in its rigour for an enterprise-services firm and reflects the broader strategic positioning DeSouza has assumed: the AI vendor relationships are operationally critical and require the same discipline that the firm has historically applied to its largest cloud-vendor and technology-platform relationships. The discipline has produced what the May 2026 review documents as "predictable commercial performance" across the principal-tier vendor base, with no material surprises across the twelve-month operating period.
The training rollout: 415,000 practitioners, eight weeks per cohort
The training program that supports the Deloitte Quant rollout has been the largest concurrent learning-and-development effort the firm has run in its recent operating history, and the twelve-month review treats the training rollout as one of the program's operational achievements rather than as a supporting activity. The program — designed in Q3 2025 by the Practitioner Augmentation team under Samira Lakhani's leadership and delivered through Deloitte University and the firm's regional learning centres — has reached approximately 312,000 of the firm's 415,000 practitioners as of May 2026. The training is structured as an eight-week cohort program, with the first four weeks delivered as asynchronous self-paced modules covering the firm's AI policy framework, the Deloitte Quant workbench's capabilities and limitations, and the partner-level governance discipline that the AI-augmented workflows operate under. The second four weeks are delivered as synchronous practice-based modules where the practitioner applies the AI-augmented workflows to live engagement work under the supervision of a partner-equivalent reviewer.
The training program's cost is the operational metric that the executive committee has examined most carefully. Deloitte's investment in the training program across the twelve-month period has been approximately $94M, against an originally-budgeted $61M — a 54 per cent overrun that the review document analyses across three causes. The first cause is scope expansion: the original program design assumed an eight-week cohort program with approximately 250,000 practitioners reached in twelve months; the actual program has reached 312,000 in twelve months, with the additional 62,000 driving the bulk of the overrun. The second cause is delivery-format adjustment: the program's early cohorts produced feedback that the asynchronous self-paced modules were under-developed and required additional content investment to meet the operational standard the partner-level governance discipline required. The third cause is the cross-vertical extension: the training program has had to develop vertical-specific extensions for the Financial Services, Life Sciences, and Audit-Tech use cases, which the original program design had not budgeted at the level the practice-leader feedback indicated was operationally necessary.
The training-completion-to-active-usage conversion is the metric that the program's effectiveness ultimately depends on. The twelve-month review found that approximately 87 per cent of practitioners who completed the training program have authenticated against Deloitte Quant within thirty days of completion, and that approximately 73 per cent have continued to use the workbench actively (defined as more than three engagement-related queries per week) ninety days after completion. The conversion-and-retention metrics are operationally strong by professional-services-industry baselines for technology-tool rollouts and reflect what Lakhani's team has described in internal materials as the "embedded utility" of the workbench — the workbench is operationally integrated into the practitioner's daily workflow in a way that creates ongoing usage rather than one-time adoption. The retention pattern is the metric that DeSouza's organisation tracks weekly and that informs the broader strategic conclusion that the program has cleared the threshold from initial deployment to operational entrenchment.
The training program's vertical extensions have produced an unexpected operational dynamic that the review document flags as a structural change in the firm's training-organisation architecture. The vertical extensions — built initially as supplementary content for the core eight-week program — have been operationally adopted by the firm's practice-leader organisations as the principal vehicle for delivering ongoing methodology updates to the practice. The Practitioner Augmentation team has been asked by three practice-leader organisations (Financial Services, Life Sciences, and Audit-Tech) to extend the vertical-training architecture beyond the AI-rollout window and into a permanent practice-training mechanism, which would represent a structural change in how the firm has historically organised its training function. The decision is on the executive committee's agenda for the September 2025 quarterly review and would, if approved, formalise the AI-rollout training program's architecture as the firm's broader practice-training architecture for the next decade.
What to watch
The twelve-month review is the structural milestone for the CIO role at Deloitte. The next twelve months will resolve five open questions that the review surfaces.
- Whether the Sector Verticalisation restructure — giving DeSouza direct authority over the per-vertical sector leads — clears the horizontal-vertical accountability friction; the restructure is the principal organisational change coming out of the review, and its operational outcome will inform whether EY, KPMG, and PwC follow Deloitte's restructure or stay with the original co-accountable design.
- Whether the consultant-utilisation curve continues its composite-positive trajectory through the second twelve months of the program; the 4.8x ROI on Practitioner Augmentation is the operational justification that authorises the program's budget growth, and a reversion in the composite metrics would force a strategic review at the executive committee level.
- Whether the Audit-Tech program's regulator engagement produces a formal PCAOB framework — beyond the December 2025 staff guidance — that the broader Big Four can rely on for audit-AI shipment; the PCAOB has been signalling a formal rulemaking timeline that would conclude in 2027, and the framework's content will shape Audit-Tech's shipping cadence across the next two years.
- Whether DeSouza's discretionary pool — approximately $100M held against emerging priorities — gets deployed against a fifth named program in the second half of 2026; the review hints at three candidate fifth programs (M&A advisory automation, regulatory-affairs automation, and risk-advisory automation), and the choice will signal the firm's view on the highest-leverage next-deployment opportunity.
- Whether the partnership matrix holds against the commercial pressure from a single frontier-model vendor seeking exclusive commitment; the Anthropic and OpenAI relationships are both up for renewal in 2027, and the dual-vendor posture will be tested against vendor pricing strategies that may favour single-vendor commitment with the cost savings that exclusivity unlocks.
Frequently asked
- Why did Deloitte choose "Chief Intelligence Officer" rather than "Chief AI Officer" for the role?
- The framing was deliberate and was pushed by Ranjit DeSouza in the May 2025 appointment briefing on the basis that the AI category was, by mid-2025, too narrow to describe the operational change Deloitte was building toward. The "Intelligence" framing was intended to capture the broader scope of the role: not just AI tooling shipment, but the firm's strategic posture toward analytics, decisioning automation, and the integration of AI capabilities into the firm's broader knowledge infrastructure. The framing has subsequently been adopted by EY, KPMG, and PwC for their equivalent functions, with PwC the most explicit in modelling its appointment language on Deloitte's framing.
- What is Deloitte Quant, and how does it integrate Claude, Copilot, and the firm's proprietary capabilities?
- Deloitte Quant is the firm's unified AI workbench, shipped to general availability in March 2026 and now provisioned to 87 per cent of the firm's 415,000 practitioners. The workbench integrates Claude Sonnet 4.6 (the firm's primary frontier-model surface for client-engagement work), Microsoft 365 Copilot (the firm's productivity-AI surface for email, document, and meeting work), and a series of internal retrieval-augmented capabilities that ground against the firm's proprietary methodology archive. The workbench's architecture allows individual capabilities to be routed to different vendor surfaces depending on the use-case fit, and the abstraction layer that supports the routing is the firm's principal engineering investment in preserving procurement leverage across the frontier-model vendor layer.
- Why does Deloitte's consultant-utilisation rate look like it has dropped, when the program's ROI is positive?
- The 2-point drop in utilisation is a composite of three offsetting effects: a 4-point drop in pre-shipment time-on-engagement (the workbench accelerates the work), a 2-point increase in total engagement count (practitioners take on more engagements because each requires less time), and a smaller residual drop attributable to onboarding-and-training time. The net effect — incremental revenue plus incremental margin minus the workbench's operating cost — is approximately $2.3B of annualised value against a $480M program operating cost, or a roughly 4.8x ROI. The per-practitioner utilisation metric is a misleading frame for evaluating the program; the composite frame, which the executive committee uses, captures the operational value the program is producing.
- What went wrong with the Sector Verticalisation program?
- Three causes: a vendor-integration cause (the retrieval architecture took approximately five months longer to ship than the program plan anticipated), a content-quality cause (the firm's proprietary sector-research archives were inconsistently digitised across the five verticals, with Energy, Consumer, and Public Sector requiring substantial unbudgeted digitisation work), and an organisational-design cause (the horizontal-vertical accountability structure produced friction between firm-wide capability-building and per-vertical P&L management). The review proposes a restructure for the second half of 2026 that gives DeSouza direct authority over the per-vertical sector leads, formalising the practice-leader involvement as advisory rather than co-accountable. The restructure is the principal organisational change coming out of the twelve-month review.
- How does Audit-Tech interact with the PCAOB and the equivalent international regulators?
- The PCAOB's December 2025 staff guidance on AI-augmented audit work — published after extended engagement with all four Big Four firms — established that AI tools may be used in audit work provided auditor judgement remains the controlling input on every audit conclusion, that the AI tool's outputs are appropriately documented in the workpapers, and that the firm maintains a documented evaluation of the tool's reliability against the specific audit assertion it supports. The guidance is operationally workable but imposes documentation requirements that have slowed Audit-Tech's shipping cadence relative to Practitioner Augmentation. The equivalent international regulators — the UK FRC, the German Apas, the Dutch AFM, the Japanese FSA — have engaged at varying levels of formality and have generally tracked the PCAOB's direction. The PCAOB has signalled a formal rulemaking timeline that would conclude in 2027.
- What is the dual-vendor posture, and how does it operate across Anthropic and OpenAI?
- Deloitte signed a framework agreement with Anthropic in September 2025 that named Claude Sonnet 4.6 as the firm's primary frontier-model vendor across all four named programs, with reserved-capacity guarantees, custom rate limits, and a price floor. The firm signed a non-exclusive framework agreement with OpenAI in November 2025 that names GPT-5 and o3-pro as approved frontier-model vendors for specific use-case categories: open-ended Q&A surfaces, long-context document analysis in M&A advisory, and code-generation in engineering services. The dual-vendor posture preserves procurement leverage across the frontier-model layer and the technical abstraction layer in Deloitte Quant allows individual capabilities to be routed to either vendor without engineering rework. The two contracts are both up for renewal in 2027, and the dual-vendor posture will be tested against vendor pricing strategies that may favour single-vendor commitment.
Deloitte's twelve-month review of the Chief Intelligence Officer role produces the first multi-quarter operating data on a Big Four firm's AI executive function. The data describes a role that has met its primary objectives — the formalisation of an AI strategy across the firm's 415,000-practitioner global footprint, the shipment of the Deloitte Quant workbench to enterprise-wide general availability, the establishment of the partnership matrix that gives the firm dual-vendor coverage across the frontier-model layer — and that has missed one of its secondary objectives in a manner large enough to require structural reorganisation. The Sector Verticalisation slippage is the operational miss; the restructure for the second half of 2026 is the response. The review document is detailed enough that it will be read by the other Big Four firms as a template for their own twelve-month reviews of their equivalent functions, which will conclude across the rest of 2026 in roughly the order their CIO roles were created.
The peer-firm sequencing of CIO appointments — EY in September 2025, KPMG in November 2025, PwC in January 2026 — means that EY's twelve-month review will arrive in September 2026, with KPMG's in November and PwC's in January 2027. The sequencing gives Deloitte's review approximately four months of standing as the only published reference until EY's review arrives. The structural advantage is operationally meaningful: the other Big Four firms will calibrate their own twelve-month review processes against Deloitte's template, and the procurement-officer and client-facing communications about the role will reflect the framing that Deloitte's review has established. The first-mover advantage in the CIO-role narrative is the secondary benefit of having appointed the role first, and Ucuzoglu's decision to authorise the appointment in May 2025 — when the broader Big Four industry was still in the planning phase — has produced an operational return that the May 2026 review document treats with appropriate institutional satisfaction.
The structural conclusion the review supports is that the CIO role is not a transitional appointment. The compensation, budget authority, organisational positioning, and operational scope of DeSouza's function describe a permanent C-level role that has now been operationally validated by a full twelve-month review cycle. The role is becoming as structurally permanent inside the Big Four as the COO role became across the broader Fortune 500 in the 1990s. The reorganisation that DeSouza is leading for the second half of 2026 will not retire the role; it will reshape the role's internal architecture for the next phase of operational maturity. The twelve-month review is the document the next phase begins from.
More from Business →