S&P Global AI as a Growth Catalyst Across Benchmarks and Workflows

S&P Global used its fireside chat at the Raymond James Institutional Investor Conference to make a simple, consequential claim: artificial intelligence is not just an efficiency play — it is a structural growth catalyst the company can monetise across benchmarks, ratings, market intelligence and workflows.

Background / Overview​

S&P Global enters 2026 on a footing that mixes legacy durability with rapid product evolution. The company’s core benchmarks, ratings and price-assessment businesses continue to generate the majority of operating income; management reiterated at the conference that roughly three-quarters of operating income still comes from these core assets. That revenue mix gives S&P a large, recurring base to deploy data, analytics and AI-powered workflows on top of.
The headline financial backdrop matters. In its Q4 2025 results the company reported mid-to-high single‑digit organic growth, margin expansion and sustained cash returns — facts that management used to frame AI as an incremental lever rather than a speculative pivot. Management’s guidance and Investor Day messaging set an explicit target: deliver 50–75 basis points of adjusted margin expansion annually (on an organic basis), a goal S&P ties in part to productivity and AI-driven deployment inside the business.
This context — stable, high-margin core franchises combined with a growing catalogue of workflow tools and proprietary data — is central to understanding why S&P’s leadership characterises AI as a growth catalyst instead of a threat.

---

What S&P Announced at Raymond James​

The core message: AI across products and processes​

During the fireside conversation, CEO Martina Cheung and CFO Eric Aboaf walked through a three‑pronged approach for AI:

• Embed generative and retrieval‑augmented capabilities into product workflows to increase customer value and expand usage.
• Deploy AI internally to reengineer large, repetitive operational processes to capture productivity savings.
• Protect proprietary data and control access, ensuring S&P content remains monetisable and not freely available to LLM training sets.

Management described ongoing pilots and rollouts across specific products — from iLEVEL (private markets/workflow) to Capital IQ Pro, Platts Connect, and core ratings/indices workflows — and stressed that many of these platforms are industry-critical, used in decision‑making and regulatory compliance, which raises the bar for commoditisation risk.

Concrete metrics management disclosed​

S&P offered several measurable datapoints at the event that clarify how management expects AI to scale:

• Targeted margin expansion: 50–75 basis points annually on an organic basis (excluding certain divestitures).
• Enterprise Data Office (EDO) program: the EDO covers about 9,000 people out of roughly 40,000, with management targeting a 20% run‑rate expense reduction by the end of 2027 through process reengineering and AI tooling.
• Product adoption signals: S&P reported that an automated data‑ingestion add‑on for iLEVEL attracted nearly 20% of clients in six months; around 80 customers have adopted the MCP connector; and roughly 60 energy customers are using integrations with Microsoft Copilot to surface S&P content.

Those are not casual, marketing‑only claims — they are operational metrics management repeatedly cited in the Q4 earnings cycle and the Raymond James discussion, and they form the basis of management’s productivity and revenue thesis.

How S&P plans to protect its IP and monetise data in an LLM world​

S&P made several technical and commercial commitments to prevent its proprietary content from being absorbed as free training data for large language models:

• Permissioning and connectors: S&P's MCP connector and “grounding agents” are used to permission external AI systems to access S&P content, ensuring the relationship remains contractual and traceable.
• Throttling & monitoring: technical throttles and usage monitoring are in place to limit volume, detect anomalous access patterns and enforce commercial terms.
• Contractual restrictions: S&P is explicit that customer licences do not grant the right to use S&P data to train third‑party models; training remains the domain of the LLM providers unless separately negotiated.

This combination of technical controls and commercial guardrails is the linchpin of the company’s argument: proprietary datasets can be simultaneously useful to customers and monetised by S&P — provided permissioning and enforcement are effective.

---

Why the market narrative matters: growth drivers and opportunity buckets​

1) Workflow monetisation and stickiness​

S&P’s workflow products are tightly coupled with proprietary datasets (loan references in WSO, benchmarks for indices, price assessments for commodities). That integration creates customer lock‑in: a workflow loses much of its value without the embedded data. By adding AI features that improve speed, searchability and decision support, S&P expects to increase usage, attach rates and average revenue per customer.

2) Private markets momentum​

S&P emphasised private markets as a high‑growth area: private credit, infrastructure and alternative funds saw fundraising increases in 2025, with S&P pointing to an 11% year‑over‑year increase in billed issuance in 2025 and a rotation toward Europe and Asia. The acquisition of With Intelligence and product integrations (for example, iLEVEL upgrades) are presented as building blocks for a comprehensive private‑markets stack.

3) Ratings & debt issuance tailwinds (data centre / hyperscaler financing)​

Management flagged hyperscaler data‑centre financing and project finance as an incremental issuance opportunity that helped drive ratings revenue last year and could continue to do so — with a potential multi‑point revenue contribution depending on how that market evolves. This is an area where S&P’s ratings and indices franchises intersect with AI infrastructure themes.

4) Internal productivity and margin expansion​

The EDO programme and software‑engineering productivity initiatives are where S&P expects to capture direct cost savings. Management quantified early software‑developer productivity improvements (double‑digit gains in some teams) and sees opportunities to compress other labour‑intensive tasks via agents and process automation across research, data ingestion, sales operations and support. The 20% EDO target is the clearest numeric articulation of that ambition.

---

Critical analysis — why S&P’s thesis is credible​

• Proprietary data moat: S&P’s core IP — indices, ratings methodologies, price assessments — are not mere datasets but regulatory and market infrastructure. That creates higher switching costs for customers than for commodity data suppliers. Embedding AI on top of this IP raises the effective value of S&P’s subscriptions.
• Recurring revenue base: The majority of operating income derives from subscriptions and recurring services, which provide cash flow to fund AI investments without destabilising the P&L. This makes S&P’s multi‑year productivity investments less risky than a capital‑intensive pivot.
• Measured, multi-layered approach to IP protection: Management didn’t rely solely on legal promises — it outlined technical permissioning layers (connectors, grounding agents, throttles) that, if implemented robustly, materially reduce the risk of unintended reuse. That technical-first posture is the right risk‑control architecture for data vendors in 2026.
• Tangible adoption signals: Early product adoption metrics (iLEVEL add‑on uptake, MCP connector customers, Copilot pilots) provide more credibility than generic AI rhetoric. Those datapoints let investors and customers validate whether the enhancements are moving from proof‑of‑concept to paid adoption.

---

Critical analysis — where the thesis faces significant risk​

While S&P’s plan is defensible, several structural and execution risks deserve scrutiny.

1) Execution risk on the EDO programme and realised savings​

S&P’s EDO covers thousands of employees and multiple legacy processes. Achieving a 20% run‑rate expense reduction by 2027 depends on flawless process decomposition, robust tooling, change management and sustained productivity gains. Historical examples across the industry show that large‑scale reengineering often underdelivers on timing or magnitude. Investors should view the target as achievable but execution‑sensitive.

2) IP leakage and the limits of permissioning​

Technical throttles and contractual guardrails materially reduce training‑risk, but they do not eliminate all paths to leakage. Third‑party LLM providers may create derivatives or paraphrased outputs that approximate proprietary insights without direct training, and enforcement across global jurisdictions is non‑trivial. S&P’s layered approach is prudent, but not proof against subtle exfiltration or ambiguous regulatory interpretations of data reuse.

3) Competition from AI‑native vendors​

Niche, AI‑first vendors with aggressive pricing and modern UX can disintermediate parts of the workflow stack unless S&P keeps pace on product experience. S&P’s advantage is the combination of workflow and proprietary data; however, if AI‑native players replicate the UX and integrate third‑party data cheaply, S&P may face pressure on smaller, non‑mission‑critical workflows. The company’s counter is co‑creation with large clients and deeper embedding in regulated tasks, but this is a continuous defensive requirement.

4) Regulatory & compliance exposure​

S&P operates in heavily regulated domains (ratings, indices, benchmarks, energy pricing). New rules governing model explainability, data sovereignty, or algorithmic accountability (for example, recent European regulatory movements) can add compliance friction to AI features and raise implementation costs. Management acknowledged regulation in passing; the company will need an elevated governance posture as it surfaces AI inside decision‑point workflows.

5) Macroeconomic sensitivity in issuance‑driven revenue​

Ratings and indices revenue remains sensitive to issuance cycles. S&P’s billed issuance was strong in 2025, but management warned of challenging comparables and potential quarter‑to‑quarter variability. If capital markets slow unexpectedly, AI‑driven topline lifts could be insufficient to offset a downturn in issuance‑dependent segments.

---

The technology and procurement angle — what CIOs and IT teams should note​

• Permissioned connectors change integration patterns. Clients using S&P content alongside LLMs will increasingly depend on connectors (MCP connector, grounding agents) to manage access. IT teams must plan for identity management, logging, and contract alignment when enabling these flows.
• Data governance moves from “nice‑to‑have” to strategic. If S&P’s approach becomes industry standard, enterprises will need stronger lineage, auditable consent mechanisms and service‑level observability on how proprietary data is used inside vendor‑supplied agents. Expect procurement and legal teams to add clauses around training‑use and third‑party model outputs.
• Vendor consolidation tradeoff intensifies. S&P argues customers want fewer vendors — not more. For enterprise architects, that creates a tradeoff: deeper integration with S&P brings fewer point vendors but concentrates operational dependency. Cloud, identity and security teams must design resiliency around vendor availability and data access controls.
• Copilot and platform partnerships matter. S&P’s customer pilots with Microsoft Copilot show how hyperscalers and LLMs plug into the incumbent data vendors. CIOs should evaluate Copilot/LLM integration not just from a productivity angle but from a contractual, compliance and audit perspective.

---

Valuation and investor implications — a measured view​

S&P’s AI argument is not a classic revaluation story based purely on hype. It’s an operational improvement thesis layered on stable, cash‑generating franchises. For investors that matters because:

• The margin expansion target (50–75 bps) is moderate and tied to internal productivity levers rather than aggressive top‑line assumptions. If delivered, it meaningfully improves free cash flow without requiring higher growth assumptions.
• Early monetisation signals (add‑on uptake, MCP connector adoption) reduce the binary risk that AI features will remain proof‑of‑concepts. These are incremental revenue levers that compound atop a large subscription base.
• Execution is the differentiator. Market multiples will increasingly price not only the quality of the data moat but management’s ability to deliver EDO savings and product attach rates at scale. Misses on either dimension are likely to pressure sentiment.

In other words, the stock’s prospective upside is correlated with measurable operational metrics — adoption rates, EDO run‑rate savings, and issuance trends — rather than speculative AI revenue attach figures. That alignment is attractive to investors who prefer operational proof points.

---

What to watch next — an action checklist​

• Quarterly telemetry on AI adoption: Watch for management to report attach rates, ARR from AI add‑ons and MCP connector growth. These numbers will be clearest validators of the growth thesis.
• EDO run‑rate progress: Interim updates on headcount redeployment, process automation milestones, and the incremental cost base impact will indicate whether the 20% target is realistic.
• Contractual language on training rights: Monitor customer contracts and S&P licensing language to see how the company operationalises its promise to prevent data use for model training. The market will react to evidence of enforceability.
• Regulatory signals in regions where S&P operates: New rules on data reuse or model governance (especially in Europe) could materially affect product rollouts and compliance costs.
• Private markets fundraising geography: Continued rotation of fundraising to Europe and Asia would validate management’s regional growth claims and expand TAM for S&P’s private‑markets products.

---

Final assessment: a reasonable, evidence‑based pivot with execution caveats​

S&P Global’s presentation at the Raymond James conference is notable not because it promised a speculative AI future, but because it linked AI to concrete adoption metrics, internal productivity programmes and a clear plan for protecting and monetising proprietary data. That combination — proprietary data + workflow stickiness + permissioned AI access + explicit productivity targets — is a compelling playbook for a data‑centric incumbent.
Yet the plan is not risk‑free. Execution on the Enterprise Data Office, the durability of permissioning against creative reuse, and macro sensitivity in issuance‑driven revenue are all material uncertainties. For enterprises and CIOs, S&P’s roadmap signals a near‑term imperative to rework data governance, procurement contracts and identity tooling to safely consume vendor AI features. For investors, the story is one of operationally anchored upside rather than speculative multiple expansion.
In short: S&P Global’s argument that AI is a growth catalyst is credible where it is tethered to product adoption and internal productivity; it becomes speculative when it relies on broad, unmeasured claims about AI‑driven top‑line acceleration. The next several quarters of adoption metrics and EDO milestones will determine whether the promise becomes predictable earnings power or remains a positive but imprecise tailwind.

Conference sources and corroboration​

This article synthesises S&P Global’s remarks at Raymond James with management commentary delivered during recent earnings and investor communications to validate claims of adoption, margin targets and programmatic savings. Key corroborating venue transcripts and official company notices were published in the immediate aftermath of the event.

---

Conclusion
S&P Global’s Raymond James appearance framed AI not as an experiment but as a layered commercial strategy: product enhancement, enterprise productivity and disciplined IP protection. That strategy fits the company’s asset base and risk appetite, and it offers a realistic path to incremental margins and monetisation. But the plan’s success depends on the company’s ability to translate pilots into billed add‑ons, to execute a complex global reorganisation without service disruption, and to maintain tight contractual and technical controls over proprietary data. For customers, investors and IT leaders alike, the next four quarters of measurable adoption and EDO progress will separate confident rhetoric from durable results.

---

Source: Investing.com S&P Global at Raymond James Conference: AI as a Growth Catalyst By Investing.com