CFO AI Optimism Hits 73% - But the Productivity Promise Still Has to Pay Out

The Confidence Curve

The numbers are striking enough to quote directly. <a href='https://www.deloitte.com/uk/en/about/press-room/deloitte-cfo-survey-quater-two-2026-growing-ai-optimism.html'>Deloitte's Q2 2026 CFO Survey</a>, conducted across 58 UK finance chiefs between 1 and 13 July, found that 73% now report growing optimism that AI will materially improve their own business's performance. That compares to 59% at the end of 2025, and just 39% in Q3 2024. In two years, the sentiment has very nearly doubled.

Nearly all respondents (96%) expect investment in digital technology and assets to rise over the next five years. Of those, 93% say that rise is coming within the next 12 months. Half of CFOs expect to see actual productivity gains from AI within a year.

For consulting and business services firms selling transformation programmes, those figures read as a green light. Clients want to spend. But there is a meaningful gap between intent and execution that the survey does not quite close.

The Utilisation Problem Nobody Wants to Name

The Management Consultancies Association has separately found that 77% of UK consulting firms have integrated AI into their systems or enabled employees to use AI models, with 68% increasing automation. The adoption rate is high. The ROI clarity is not.

Dayshape's analysis of professional services economics puts it plainly: AI saves time, but not always in neat, reusable blocks. In service lines where work is booked in fixed chunks, shaving hours off a task does not suddenly free someone up for another engagement. The firms that benefit most are those that can deliberately turn fragmented time savings into usable capacity across teams. That makes resource planning more critical, not less.

The CFO optimism is real. The mechanism for capturing it in billable hours or fee income is still being worked out.

Graduate Hiring and the Quiet Signal

Buried in the Deloitte data is a detail that professional services firms should read carefully. Cost control remains the single biggest dampener on graduate hiring: 64% of FTSE 100 and FTSE 250 CFOs say broader cost-control efforts will reduce graduate intake over the next 12 months. AI is reshaping workforce planning, but many organisations are making hiring decisions primarily through budget discipline rather than any formal headcount-replacement strategy.

That is a problem for consulting, recruitment, and training businesses simultaneously. Fewer graduate hires means fewer candidates flowing through assessment pipelines, less demand for the onboarding and early-career training programmes sitting squarely at SIC 85.59, and a tightening talent pool at the entry point that consultancies rely on for delivery.

The ICO's Recruitment AI Compliance Deadline Is Live

The CFO story lands at the same moment that the ICO's final automated decision-making guidance is due any week. The consultation closed on 29 May 2026, and the regulator has signalled summer 2026 as the publication window. Recruitment businesses and HR teams at consulting firms should treat that as imminent.

The ICO's March 2026 "Recruitment Rewired" report, based on evidence from more than 30 employers, found most are non-compliant with automated decision-making rules. The central problem: employers said their AI tools were used only for decision support, with a human making the final call. The evidence showed something different: tools making substantive decisions, and human review amounting to rubber-stamping. The ICO is explicit that approving an AI-generated shortlist without genuine capacity to override it does not qualify as meaningful involvement.

The Data (Use and Access) Act 2025, in force since 5 February 2026, replaced the old near-total prohibition on significant automated decisions with a permissive, safeguard-led regime. That is not a free pass. Any AI that filters, ranks or rejects candidates now requires a Data Protection Impact Assessment, bias monitoring, and a documented human review process. The ICO has already written to 16 named organisations.

For recruitment agencies and HR consultancies building their AI stack right now, the compliance clock has already run.

Formation Data: The Optimism Has a Floor

AI Business Dispatch analysis of Companies House data shows 3,125 new SIC 70.22 companies registered in 2026-Q3, a 75.6% fall versus the prior period. Trademark filings tell a similar story: Class 45 UK filings at the IPO reached just 530 in Q3, down 66.7% on the prior period. And 99.7% of active SIC 70.22 companies hold no Class 45 trademark at all (AIBD analysis of Companies House and IPO data, as of July 2026).

That last figure is not a curiosity. Class 45 covers legal and security services: the intellectual property protection layer that a management consultancy ought to hold if it is genuinely building a proprietary brand around an AI methodology or service name. The near-total absence of trademark protection in the cohort suggests most new entrants are moving fast on delivery and slow on brand infrastructure. When CFO clients start consolidating vendors as AI matures, the firms without defensible brand positions will be the first to compete on price alone.

What the Data Actually Says

The Deloitte survey is a sentiment measure, not a revenue measure. Fifty per cent of CFOs expect productivity gains this year; 50% do not, or have not committed to a view. On the buy-side, 37% of CFOs said they plan to expand to new products or markets, up 12 points from Q1. That is the number consulting firms should track: it is the proxy for discretionary transformation spend.

Geopolitical risk eased this quarter. Concern over the conflict in Iran drove the average geopolitical concern rating down from 79 to 68. Energy price concern fell from 70 to 60. Those are not resolved problems; they are pauses. CFOs who have been in cost-control mode for two years are not suddenly running open-ended transformation budgets.

The firms that convert this optimism into revenue will be specific about outcomes and honest about timelines. The ones that pitch AI transformation as self-evidently worthwhile will find CFOs are now experienced enough to ask the uncomfortable follow-up: compared to what, and by when?