The ROI on Intelligence: Turning AI Investment into Enterprise Transformation | RAISE Summit 2026

ROI Metrics and Investment Shifts

• Jordan Topolesky (Cursor) reports a market shift from measuring AI adoption inputs to tracking outputs and organizational outcomes over the last 18–24 months.
• CFOs have increasingly taken ownership of AI strategy as costs exploded, necessitating a move from simple tool adoption to investment optimization.
• Cursor introduced the role of "forward-deployed ROI specialist" to help organizations implement controls for responsible scaling and cost management.
• A case study with a large hardware company reduced AI spending by 26% while maintaining a 60% rate of AI-generated code reaching production.
• Nokia initially imposed token quotas per engineer but abandoned strict limits after receiving negative feedback that hindered innovation.
• Pallavi (Nokia) shifted the ROI conversation from budget constraints to business outcomes, specifically focusing on product delivery speed and quality.
• Organizations are moving beyond "token spending" leaderboards (cited at Meta and Tesla) to prioritize value creation per project rather than per developer.
• The cost optimization strategy involves dynamically selecting the most cost-effective model for specific tasks rather than defaulting to frontier intelligence for all use cases.

Organizational Restructuring and Role Compression

• Topolesky identifies the bottleneck in AI transformation as culture and operating models, not technology itself.
• Middle management layers are being reinvented from coordination roles to support new AI-native workflows where automated systems reduce coordination needs.
• Nokia is experiencing "role compression," collapsing distinct titles (e.g., software engineer, test engineer) into broader "part builder" roles.
• To adapt to AI agents and systems, organizations are shrinking team sizes; a large insurance client split eight-person teams into two four-person teams, achieving a 3x uplift due to reduced coordination costs.
• The software development lifecycle is rebalancing: while "write and review" phases are automated, resources are shifting to "plan, design," and "test/deploy" phases.
• Topolesky notes that organizations often stall at 35–40% of code written by AI unless they restructure their system architecture to support multiple parallel agents.

Innovation and Economic Impact

• Philippe Agnon (INSEAD/Nobel Laureate) posits that AI efficiency boosts GDP by only 0.7% annually, whereas increasing the volume and quality of ideas could yield significantly higher growth.
• Both speakers agree that AI is enabling a shift from "how" (execution toil) to "what" (strategic ideation and business context).
• Nokia is leveraging freed-up engineering time to innovate on "AI-native networks," moving away from static connectivity toward intelligent systems for physical AI (AR/VR, drones).
• Cursor observes that the primary constraint for enterprises was not a lack of ideas but the capacity for execution, which AI has expanded.
• Organizations are now utilizing AI to incorporate previously unfeasible "below-the-line" roadmap items, accelerating innovation cycles and expanding the scope of organizational output.