The potential of AI is considerable, but its value depends on solid foundations: reliable data, governance and redesigned processes. Without it, automation amplifies inefficiency.
The opportunities linked to AI are immense: according to McKinsey, it could generate up to $4.4 billion in productivity and growth over the long term. However, while the potential of AI is significant, the short-term returns we have are currently less clear. The Boston Consulting Group reveals that only 5% of companies are truly ready to embrace the future with AI, while 60% see little or no revenue gains or cost savings despite their investments.
The gap is not just about technology, but more about the foundations needed to support it: alignment, data, processes and governance. As we begin 2026, leaders must ask themselves a new set of questions. These go beyond the adoption of AI to question the very structure allowing sustainable and explainable organizational value creation.
Are we ready to realize the value of AI?
It’s tempting to take existing processes and simply automate them. But if the original data and workflows are failing, automation will only accelerate inefficiency.
AI is now integrated into major business management systems. Using agentic workflows, conversational intelligence, and natural language search, organizations can automate document reconciliation, identify exceptions, accelerate validations, and guide next steps, without added complexity. These benefits multiply when teams share the same data, processes and safeguards. Without a connected system – that is, shared data, clear governance and consistent adoption – the gains do not change in scale: it produces activity, but not necessarily operational impact.
This highlights a critical point: many “productivity metrics” reward activity, not necessarily impact. Before applying AI to an existing system inherited from another era and therefore obsolete, it is appropriate to question its purpose: why is our process structured like this? What would we do differently if we were starting from scratch?
Without questioning old patterns or reshaping the desired results, around a single data source, automation risks accelerating inefficiency instead of generating value.
How to have confidence in decisions supported by AI?
The shift to AI highlights a chronic pain: data fragmentation. Many companies hold a mass of information, but only a limited proportion is sufficiently coherent, governed and accessible to power an intelligent system. And if it is not possible to explain how the system works, then it should not be deployed: any responsible AI must be explainable. Good news: the “garbage in, garbage out” phenomenon is now widely understood by managers.
The mission is therefore clear: we must centralize real-time data from the entire organization – finance, operations, HR, supply chain – to make it accessible and connected between the different departments.
Visualizing data in a dashboard tailored to each employee’s role and responsibilities, and the ability to drill down and interrogate pain points, allows leaders to make decisions based on a single source of truth and not on instinct. And when more employees can access the same consistent data, organizations optimize the possibilities of detecting possible risks and increase the opportunities to surface relevant information.
What strategy regarding shadow AI?
Employees are already using AI tools, whether managers allow it or not. On the one hand, this demonstrates a certain interest in AI and its potential in terms of productivity gains. But on the other hand, the covert use of AI, called “shadow AI”, and outside of any framework planned by the organization raises serious security, governance and compliance issues when company data is used and shared outside of official systems.
Leaders must prioritize safe and governed ways of using AI, in line with company policy. Solutions such as NetSuite AI Connector Service, for example, offer employees a flexible and scalable way to connect their own AI to NetSuite, ensuring that results come from validated data, with full compliance. This approach restores control over the uses of AI without slowing down the innovation that teams can demonstrate.
How will teams interact with AI in 2026?
The relationship between employees and systems is changing rapidly. Conversational intelligence, autonomous workflows via AI agents and natural language search are becoming the norm in everyday life.
Rather than navigating dashboards or clicking through various menus, employees will increasingly be asked to question systems orally or in writing (“show me the revenues for the last two quarters”) or to interact with autonomous agents capable of carrying out complex workflows under human instructions. AI will be embedded in applications used on a daily basis by employees, without any need for a specialized interface or advanced technical skills to extract value.
This shift requires leaders to rethink skills, governance, and the employee experience. Next-generation ERPs thus exploit AI as a natural extension of the company’s existing operations.
What skills will matter when everyone is increased?
Initiatives focused solely on technical skills will not be sufficient. The competitive advantage represented by an AI-enriched workspace does not only belong to those who master the technological aspect, but to those who know how to interpret, challenge, and contextualize the information generated.
As smart tools amplify human capabilities, leaders must first and foremost value critical thinking and creative synthesis, and ensure that training and development plans reinforce these irreplaceable skills.
As 2026 dawns, the organizations that will fully benefit from AI will not be those that have accumulated the most tools, but those that have rethought the foundations on which these tools rest. The lasting impact comes from workflows redesigned around results, a single reliable data source, secure and controlled use of emerging AI capabilities; it also comes from the anticipation of new forms of interaction between employees and systems, in a more conversational environment and more driven by intelligent agents.
