And it’s a wrap! We just closed two intense weeks shaped by our AI Forums in Singapore and Sydney. The events revealed both the extraordinary momentum behind enterprise AI and the scale of the reinvention still ahead. Across the two cities, we brought together close to 400 business and technology leaders. In the tech track, our latest AI research was reinforced by perspectives from Grab, CLP Hong Kong, and DBS in Singapore, and from Canva and Lendi Group in Sydney. Questions from the stage, workshops, and attendee conversations converged on one issue: How do enterprises turn rapidly advancing AI capabilities into sustained outcomes? Back home, I’ve had time to reflect with my colleagues on what these discussions revealed.

AI Adoption Is Outpacing Organizational Adaptation

AI capability is advancing faster than most organizations can absorb. Enterprises are experiencing the fastest technology adoption in recent history, but adoption alone does not produce business outcomes. They must integrate AI into existing architectures, redesign workflows (and work itself), develop new skills, manage emerging risks, and change the way people make decisions every day. Each of these activities moves at a different speed, creating a gap between the capabilities entering the market and the4 ability of organizations to use them safely, consistently, and economically.

  • Enterprise value emerges at the process level. Many organizations can identify productivity improvements for individual employees. For instance, Ken Wong, Grab’s head of engineering, described the difficulty of translating tool-level time savings into enterprise value. Grab is now moving from estimates of individual productivity toward baselined, process-level measures owned by functional teams. Its legal contract-review pilot reduced first-pass review time from 60 minutes to two minutes across 104 NDAs, while its finance pilot reduced monthly close time by 10%. AI value becomes visible when organizations transform workflows rather than automate isolated tasks.
  • Control planes translate policy into practice. Governance was a prominent topic at last year’s forums and remained so this year. Organizations increasingly understand the policies and principles they need. They are now asking how to translate them into enforceable controls across models, agents, data, applications, and workflows. Nimish Panchmatia, chief data and transformation officer at DBS, made this concrete through GenAI Flock, its enterprise AI gateway for governed model access, usage monitoring, operational controls, and financial oversight. Managing more than 300 billion tokens each month highlights the growing importance of centralized control as AI adoption expands.
  • Operating models determine AI absorption. Bailing Zhang, principal enterprise architect, and Manoj Parpiani, principal data architect, described how CLP’s AI task force brings together value realization and prioritization, data and architecture, responsible management, and workforce capability. Its enterprise architects connect business outcomes to platform choices, build-versus-buy decisions, data protection, residency, supplier diversification, and guardrails. As a fully integrated energy and critical infrastructure provider in Hong Kong, CLP treats these decisions as part of a cohesive operating model. The ability to absorb AI at scale increasingly depends on how effectively organizations align governance, architecture, workforce readiness, and business outcomes within that model.

AI Reinvention Requires A New Unit Of Transformation

As organizations gain experience with AI, attention is shifting from individual activities toward the broader systems of work that connect customers, employees, decisions, and outcomes. These work systems are becoming the new unit of transformation. The most useful examples shared during the forums showed how organizations are redesigning journeys, redefining roles, adding new skills and rethinking accountability across people and AI. AI reinvention is becoming an organizational challenge as much as a technological one, where:

  • Customer journeys provide a better design point. Lendi Group rebuilt its mortgage journey around Guardian, a swarm of agents supported by roughly 350 reusable tools. Parts of an experience that previously took days can now be completed in as little as 10 minutes, contributing to a reported 25% improvement in lead conversion. The more important lesson lies in the design approach. Travis Tyler, Lendi’s CPO, shared how they organized the journey around customer intent, separated deterministic decisions from probabilistic interactions, and introduced human gates where risk parameters required additional oversight. Its experience illustrates how customer journeys can become a more effective starting point for AI-enabled transformation than existing process boundaries.
  • AI shifts work toward judgement and supervision. The adoption of AI is changing the composition of work across a growing number of roles. Lendi described engineers spending more time reviewing AI-generated code, while brokers focus more heavily on customer relationships and judgment. Jackie Hill, AI Marketing Lead at Canva, characterized a similar shift as an expansion of employee capability, coupled with responsibility for validating results and understanding accuracy. These examples point toward a progression from execution to orchestration and supervision, creating new expectations for how organizations develop skills and define expertise.
  • Context engineering makes AI work executable. As AI moves deeper into journeys and roles, organizations need to define the context that allows models and agents to operate reliably inside real work. That context includes business intent, data boundaries, task instructions, workflow state, memory, escalation paths, and guardrails. Treating context as an engineering discipline helps organizations move beyond prompt experimentation toward repeatable work systems. It also creates a clearer connection between how work is designed, how AI participates, and where human judgment remains necessary.
  • Hybrid workforces depend on shared accountability. Grab offers a practical example of the accountability this requires. Its central technology team owns identity, security, deployment standards, and platform guardrails, while functional centers of own business rules, test sets, and workflow outcomes. Internal forward-deployed engineers (FDEs) help functions establish the scaffolding before ownership transitions to the function. This handover of ownership prevents the central team from becoming a delivery bottleneck and keeps accountability with the team that understands the work.
  • The target operating model remains a work in progress. Organizations are still learning which practices scale, how much autonomy agents should receive, how human roles will change, and which controls create confidence without constraining innovation. Forrester’s work on the cognitive operating model and the workflows, roles, and skills framework provides an initial structure for these decisions. Many of the management practices required to coordinate people and AI remain in formation and will continue to develop through experimentation and operational experience.

The Race Will Be Won Through Organizational Reinvention

Forrester’s AI Voyage emphasizes four dimensions of readiness: becoming customer-led, driving transformation from the CEO level, building the required data and technology platforms, and raising the organization’s artificial intelligence quotient, or AIQ. The discussions during the forums reinforced two important lessons. First, the foundations of successful transformation remain essential for AI reinvention. Customer obsession, executive commitment, and modern technology platforms continue to shape enterprise outcomes. Second, organizations are developing new management practices for an environment shaped by agents, probabilistic systems, and hybrid workforces. The gap between what is necessary for AI readiness and what is sufficient defines the next phase of AI reinvention. So remember:

  • Outcomes will matter more than AI adoption. Enterprise leaders should measure progress by their ability to produce different outcomes, rather than by the number of AI tools and use cases they deploy. Wider access to models, agents and harnesses will make technology increasingly abundant. The differentiator will become an organization’s capacity to direct that intelligence toward customer and business outcomes, redesign work around it, and govern the resulting system with confidence.
  • Organizations must learn to orchestrate intelligence. Frontier labs will continue to move quickly, model capabilities will continue to improve, and tech vendors will continue to close gaps in the enterprise stack. Enterprises have a different race to run. They must build the structures, skills, platforms, controls, and management practices that allow people and AI to perform together as a coherent system.
  • The future will be shaped through experience. This is what makes the current moment demanding and exciting. Forrester has built a substantial body of knowledge about what creates successful transformation, and enterprises still have work to do to put those foundations in place. At the same time, organizations are entering territory whose operating models, roles, and practices will be shaped through experimentation and experience. The organizations that begin redesigning how work, decisions, and accountability come together will help define what AI reinvention ultimately becomes.
Share