Leadership vacuum widens as enterprises mandate AI without clear roadmaps
A veteran enterprise leader says agentic AI will push organizations toward ambiguous, high-impact problem solving
The rapid push by corporate leaders to embed artificial intelligence (AI) across organizations is exposing a growing gap between ambition and execution. While senior executives are increasingly mandating AI adoption, many companies still lack a clear roadmap for how employees should use the technology in day-to-day work.
That disconnect is creating uncertainty inside enterprises at a moment when AI tools are becoming easier to access but harder to govern. As generative systems spread beyond technical teams, organizations are grappling with how to align strategy, workforce confidence, and long-term value creation.
“We have a little bit of a leadership vacuum,” said Atif Rafiq, CEO and co-founder of Ritual, a software company that uses AI and workflow tools to support enterprise decision-making. “You’ve seen statements in the last two months by CEOs of very prominent companies setting their expectation that every employee must use AI in their work. But there is a little bit of a gap in terms of the roadmap.”
That absence of direction is feeding unease rather than empowerment.
“People have a lot of anxiety in the workforce around how to embrace it,” Rafiq said.
The challenge is not that executives are being overly prescriptive. In many cases, they deliberately avoid rigid instructions because they sit too far from the specifics of daily work. But without translation into concrete workflows, high-level mandates risk becoming hollow signals rather than catalysts for transformation.
Event context
Rafiq made the remarks during a keynote address at Momentum AI London 2025, a conference organized by Reuters and held in London. The event brought together enterprise leaders, technologists, and investors to discuss the transition from experimental AI deployments to production-scale use, with sessions spanning foundational technologies, real-world use cases, risk management, and return on investment.
Drawing on a career that includes senior roles at McDonald’s, Volvo Cars, and MGM Resorts, Rafiq framed the current AI moment as less a technology problem and more a leadership one. He argued that organizations are struggling to connect rapid advances in tooling with a coherent operating model for knowledge work.
As attention shifts toward agentic AI systems that can act autonomously within defined boundaries, Rafiq said the real opportunity lies in how companies redeploy human effort.
“Once we free humans from routine work, then we get to redeploy people for higher-value activities,” he said.
He described this shift as moving organizations toward what he calls the “problem-solving frontier,” the part of the business where ambiguity is highest and competitive advantage is created.
“It’s where there’s a ton of ambiguity, and it’s where growth actually comes from,” he said.
Rather than focusing solely on automation metrics, Rafiq urged leaders to think about how AI changes the nature of work itself. Routine tasks can be delegated to machines, but judgment-heavy problems, strategic trade-offs, and cross-functional decisions remain fundamentally human.
“As agentic takes place, it’s not too early to start thinking about problem-solving on the frontier,” he said.
From drafts to workflows
One reason the leadership gap is widening, Rafiq suggested, is that today’s AI tools create a misleading sense of progress. Generative systems excel at producing first drafts, prototypes, and rough outputs, but they do not automatically deliver high-quality, end-to-end outcomes.
“Today’s tools allow almost anyone, even a novice, to generate a great first draft of something,” he said.
That capability is reshaping how teams interact. Designers can produce code-like artifacts, while engineers can spin up early prototypes without traditional handoffs. “A designer can generate code, and an engineer can generate a prototype,” Rafiq said. “But when we think about high-quality, end-to-end work, people are going to step on each other’s toes more.”
He warned that without rethinking workflows, organizations risk amplifying friction rather than productivity. Product teams that once relied on clear role boundaries must now redefine how collaboration, accountability, and quality control work in an AI-assisted environment.
‘Self-driving business’
Rafiq traced these challenges back to an idea he first floated nearly a decade ago. In 2017, while speaking at a major AI conference, he explored the idea of a “self-driving business,” inspired by developments in autonomous vehicles.
“I began to talk about a world where we would have humans as sensors in organizations with thinking machines in the middle,” he said.
At the time, the concept was deliberately provocative and loosely formed.
“I didn’t really know what I was talking about at that point, but I didn’t let the idea die,” he said. “I kept developing it.”
Today, he sees elements of that vision resurfacing as enterprises experiment with agentic systems that can coordinate tasks, surface insights, and support decision-making. Yet he emphasized that fully autonomous organizations remain a distant prospect, and that the transition will be uneven and, at times, uncomfortable.
Turning opportunity into clarity
Against that backdrop, Rafiq argued that the central task for leaders is interpretation. Technology alone cannot resolve the uncertainty facing employees; clarity must come from how AI is framed, governed, and embedded into real work.
“There is a need for clarity in our organizations,” he said. “We need to translate what this big opportunity means to both the business and the people in the workforce.”
At Ritual, the company he founded in 2021, Rafiq said the focus is on embedding structured problem-solving and decision-making workflows across product development, innovation, and strategic planning. The aim is not to replace human judgment, but to support it with better preparation, alignment, and context.
Looking ahead, he suggested that enterprises that succeed with AI will be those that treat leadership, workflow design, and human confidence as first-order issues.
As AI continues to evolve, the competitive gap may widen less between companies with better models and more between those that can clearly articulate how people and machines are meant to work together.