‘Close to zero’: 4 AI chiefs say no companies have truly changed in response to AI — blame ‘politics projects’

AXA is targeting €500 million to €700 million a year in recurring pre-tax value from AI by 2029, after starting with about 400 use cases and narrowing its focus; it is also beginning to use roughly 4 billion unstructured documents. Executives said few large organizations have reinvented core operations around AI, emphasizing workflow redesign and organizational change over simply adding AI tools. Mastercard is rebuilding tools developed about a year ago, while Accenture’s Marc Warner said companies generally should not focus on blame for failed AI projects except in critical incidents such as major security breaches.

Analysis

The investable signal is that AI value may accrue to firms able to redesign workflows and overcome internal resistance—not simply to model providers or companies reporting high usage. That favors a longer implementation cycle for Accenture (ACN): complex change could sustain advisory demand, but the same workflow redesign could eventually reduce billable implementation work as clients build repeatable capabilities in-house. The event, sponsored by ACN, is not evidence of realized client ROI.

For Mastercard (MA), the described product-development changes are a plausible route to operating leverage, but they should not be conflated with payment-volume growth or broad enterprise-wide productivity. AXA (CS) has a potential cost/productivity catalyst, but its stated future value target needs validation against delivery costs, realized run-rate savings, and whether benefits are net of investment. Neither company’s comments establish those outcomes.

Over the next 1–3 months, watch earnings commentary for quantified AI-related productivity, headcount or cost changes, and guidance—not use-case counts. Over 6–18 months, the key risk is that pilots and faster individual steps fail to translate into end-to-end throughput because downstream processes, governance, and workforce adoption remain bottlenecks. A second-order risk for ACN is a split outcome: more transformation demand near term, but price pressure if AI tools make routine consulting work easier to reproduce. Energy constraints are a longer-dated risk, not a near-term catalyst. The contrarian point: broad AI enthusiasm may be underpricing organizational execution risk, while dismissing AI spend entirely may miss durable demand for difficult workflow redesign.

Key Decisions for Investors

• No event-driven position: the discussion supplies no independently verified financial impact or new guidance. Treat it as a framework for diligence, not an earnings signal.
• Watch MA for evidence that workflow redesign produces sustained expense leverage or faster product launches without weakening controls; absent quantified results, do not infer a change in payment economics.
• For CS, require reported progress toward the stated AI value ambition, including realized annualized benefits net of implementation costs. If subsequent disclosures show delays, weaker savings, or higher costs, reassess the productivity thesis.
• Keep ACN on a two-sided watchlist: stronger client demand for core-process redesign could support services demand, while evidence of reduced project scope, pricing pressure, or weaker consulting bookings would challenge that case.