Goldman Sachs Bets AI Will Grow, Not Shrink, Its Workforce

Goldman's AI Bet: More People, Not Fewer

David Solomon sees AI expanding-not shrinking-Goldman Sachs' workforce over the next decade. He expects automation to clear busywork while demand rises for new, higher-value roles. The core idea: AI scales capacity and ambition, which pulls in more human talent.

This view counters the prevailing cost-cutting mindset. History supports it: major tech shifts boosted productivity and headcount in firms that leaned in early.

AI As a Growth Engine, Not a Headcount Axe

Goldman is investing heavily to improve risk, deal execution, and client service. Solomon's point is simple: greater throughput invites tougher mandates, bigger pipelines, and new lines of service. That means more people in specialized roles.

He also flagged a likely market reset tied to AI hype-think a 5% to 10% drawdown in the next 12 to 24 months-but framed it as a healthy filter. Long-term value still stands.

Where Headcount Grows

• AI product ownership for revenue teams (deal origination, coverage, advisory).
• Model risk, audit, and compliance for regulatory-grade oversight.
• Data engineering and governance to improve data quality and lineage.
• AI platform integration across research, sales, trading, and operations.
• Client-facing specialists to translate AI outputs into advice and action.

Operating Model: Human-In-The-Loop

Machines handle repeatable tasks; people handle judgment, relationships, and escalation. This raises service levels and deal velocity without sacrificing control. The result is more mandates supported by fewer bottlenecks.

Practical Playbook for Executives

• Define "capacity goals" before tools. Decide what volume or service uplift you want, then back into use cases.
• Fund a core AI platform team that serves multiple businesses. Avoid one-off pilots that never scale.
• Start with high-frequency workflows: research synthesis, pitch prep, KYC/AML checks, and model documentation.
• Treat governance as a product. Build model registries, monitoring, human approval gates, and immutable audit trails.
• Adopt a portfolio approach: 70% proven use cases, 20% adjacent bets, 10% experiments. Kill or scale based on clear gates.
• Revise job architectures. Write role profiles that blend domain expertise with AI fluency across every front-office and support function.

Metrics That Matter

• Deal cycle time: days from pitch to mandate; mandate to close.
• Analyst throughput: hours per task before/after AI assistance.
• Win rate and fee capture on AI-assisted pitches.
• Risk event rate and time-to-detect under human-in-the-loop controls.
• Model change failure rate and time-to-remediate.

Budget Guidance

• Anchor on platform ROI, not single-use ROI. Shared services outperform tool sprawl.
• Expect uneven payoffs. Some bets will miss; the portfolio should net positive within 12-24 months.
• Co-invest with infra partners to lower compute and data costs while improving latency and privacy.

Risk, Compliance, and Trust

• Mandate human sign-off for material outputs and client-facing content.
• Log every prompt, output, and approval for audit readiness.
• Segment data by sensitivity; use retrieval layers for confidentiality.
• Stand up bias testing and red-teaming as recurring routines, not one-time checks.

Workforce Planning: 12-24 Month Snapshot

• Hire: platform engineers, data governance leads, model risk officers, prompt engineers with domain fluency, and AI product managers.
• Upskill: coverage, research, and operations teams on AI-assisted workflows, controls, and client communication.
• Partner: legal and compliance early to codify acceptable use and approval flows.

Why This Matters Now

AI infra buildouts are pushing economic activity, even as markets may cool. Firms that treat AI as capacity expansion will capture share while others cut into bone. Solomon's stance signals where competitive advantage is moving: smarter throughput with human judgment at the core.

Next Steps

• Pick three workflows to automate end-to-end with human approval gates.
• Stand up an AI steering committee with P&L, risk, legal, and tech representation.
• Publish an AI operating standard for your firm and train managers on it.

For ongoing market context, track broad AI coverage from major outlets like CNBC. If you're building internal capability fast, explore finance-focused toolsets and certifications that align teams on common methods.

Resources
AI Tools for Finance - a curated starting point for deal, risk, and ops use cases.
Popular AI Certifications - options to formalize skills across roles.