S&P Global Turns to AI to Strengthen Data Business and Cut Costs

S&P Global Puts AI at the Core of Data Management - What Managers Should Do Next

S&P Global Inc. (NYSE: SPGI) is leaning into artificial intelligence to upgrade its products and internal processes. On September 9, CEO Martina Cheung reiterated that AI integration and data management are the company's top strategic priorities.

Two-thirds of employees are already using the S&P Spark Assist platform. Leadership expects automation to bring efficiency gains and, over time, a reduction in headcount.

S&P's clients-including banks and other financial institutions-are testing AI to boost productivity and lower costs. The company is also exploring partnerships with AI vendors to make its data easier to use inside customers' AI tools.

Why this matters for management

• Productivity and cost: AI assistants compress research time, reporting cycles, and routine analysis. Budget shifts from manual work to higher-value tasks.
• Customer pull: As clients adopt AI, they expect vendors to provide structured, API-accessible data that plugs into their models.
• Data as a product: Clean pipelines, clear permissions, and usage-ready packaging become core capabilities-internally and for customers.

Operating model shifts to plan for

• Workflow redesign: Map end-to-end tasks, automate the repetitive segments, and keep human review where accuracy and judgment matter most.
• Data governance: Standardize metadata, lineage, and access policies. Use an AI risk framework to guide approvals and monitoring (NIST AI RMF).
• Tech stack choices: Prioritize secure data access (APIs, governed lakes), retrieval augmentation for generative tools, and logging for auditability.
• Vendor contracts: Lock in service levels, data retention limits, training restrictions on proprietary data, and audit rights.
• Workforce planning: Pair automation with reskilling and redeployment plans. Be transparent about role changes and growth paths.

Metrics that signal real impact

• Efficiency: Time-to-insight, cycle time for client deliverables, cost per analysis.
• Quality: Error rates, factual accuracy of AI outputs, compliance exceptions.
• Adoption: Percentage of staff using assistants weekly, satisfaction scores, usage depth by function.
• Financials: Opex savings, margin improvement, and revenue from AI-enhanced products.

Partnership playbook

• Access models: Favor clean APIs and controlled embeddings over ad hoc file drops. Keep customer data and proprietary data strictly segmented.
• Security and privacy: Require encryption, strict identity controls, and logs. Define breach notification and remediation steps upfront.
• Interoperability: Push for open formats and clear schemas to avoid lock-in and speed integration.

What to execute this quarter

• Pick two high-volume use cases (e.g., research summaries, client Q&A) and run a four- to eight-week pilot with baselines and targets.
• Stand up a small AI review board (data, legal, security, business) to approve use cases, monitor results, and publish standards.
• Document a human-review step for external outputs and set thresholds for when to escalate.
• Upskill your team with focused, role-based training (Courses by Job).

The takeaway for leaders

S&P Global's move signals where information businesses are heading: AI-first workflows, cleaner data products, and deeper customer integration. Managers who systematize governance, measure outcomes, and reskill teams will capture the upside while controlling risk.

If you need a governance standard to anchor your program, consider an AI management system framework such as ISO/IEC 42001 alongside the NIST AI RMF. Start small, measure hard, and scale what works.