For leaders, the real question is not whether agentic AI can automate work, but which work should be redesigned first. The articleโs zero-based process redesign approach implies a management shift: treat AI as a trigger to reissue process ownership, decision rights and performance measures, rather than as a bolt-on productivity tool. That matters because partial automation can preserve handoffs, delays and controls that were designed for a human-only model.
Executives should expect a portfolio-level trade-off. Agentic redesign is not a broad โscale everywhereโ programme; it is a sequence of bets on processes where cycle time, compliance burden, exception rates and labour intensity make the case. The practical next question is which workflows have enough volume and decision consistency to justify redesign, and which should remain human-led because judgement, liability or customer sensitivity outweigh automation gains.
Operating model is the hinge point. If agents begin orchestrating end-to-end work, governance must move from task supervision to outcome supervision. That means defining who owns the process, who signs off on exceptions, what audit evidence is retained, and how incidents are escalated when agent decisions drift. Without this, organisations may increase automation speed while weakening accountability.
Leaders also need a deliberate change plan, not just a technical pilot. Zero-based redesign will expose roles that shrink, shift or disappear, so reskilling and communications should be built into the business case from day one. A useful next step is to create a small number of process โdesign authorityโ workshops that pair operations, risk, finance and technology leaders to decide where human oversight is essential, where it is optional, and where it adds cost without value.
Radical, not incremental, process redesign
Rather than deploying agentic AI incrementally within existing processes, ZBPR harnesses the contextual awareness, reasoning, planning and acting capabilities of agentic AI to radically transform processes. The result is workflows that optimize costs and efficiency, support risk management and compliance, and are more scalable and flexible than legacy processes without linear cost increases. For example, an employee onboarding process that used to require signing in to multiple platforms to handle account creation, payroll and equipment could be redesigned using a team of orchestrated agents. Those agents could handle all onboarding tasks for each new hire through a single point of contact. The agents and process could be adapted across different geographies and acquisitions for flexible scalability. Other benefits of agent-native processes include the ability to run around the clock with minimal human oversight, shorter cycle times, more accurate data entry and real-time data visibility for compliance and insights. By eliminating rote, repetitive tasks, ZBPR can also increase employee availability for higher-value, more engaging work, such as dealing with edge cases and developing strategies based on agentic process insights. For example, with an expense-processing agent handling all employee expense reports below a certain threshold of value, the human manager can focus on higher-value reports and those with flagged anomalies.The business impact of ZBPR
Organizations that have already implemented ZBPR for their agentic process transformation report higher AI ROI than organizations using less holistic automation strategies. The ROI improvement is due to not just greater efficiency. In some cases, the ZBPR plus AI approach allows organizations to automate workflows that previously couldn’t be automated. Consider an insurance contact center, where policyholders call or message with many different types of claims, varying levels of coverage and a patchwork of state laws governing their policies. An incremental agentic strategy would use a chatbot to handle the most basic inquiries, saving some time but not fundamentally transforming the experience for policyholders or service agents. A ZBPR redesign of the contact center workflow can fully leverage agentic capabilities. For example, one general AI agent can triage and process most basic contacts from end to end. A team of more specialized agents can handle a large portion of the contacts that the main agent isn’t trained to deal with. That leaves a smaller set of more complex or high-value issues for human agents to handle. By using ZBPR, the contact center can reduce costs, resolve issues faster for better customer experience and allow human agents to focus on the areas where their judgement and empathy matter the most. This example highlights a key trend. The most innovative adopters of agentic AI are shifting away from task-level automation to build multi-agent, end-to-end workflows that deliver more value than automating individual tasks. This shift is critical for organizations that want to future-proof for efficiency, agility and resilience.A practical roadmap for ZBPR and agentic AI
Adopting a zero-based process redesign mindset requires a shift that starts at the top. Executives need to develop and share a clear vision of what’s possible with agents and identify high-impact processes to pilot this approach. Next, zero-based process design workshops create agent-native workflows that achieve process outcomes more efficiently and support business goals for value creation. For each redesigned workflow, the organization must orchestrate multi-agent teams to handle all relevant processes. Humans must be in the loop as safeguards for edge cases and for compliance monitoring using real-time process data. As agentic pilots scale, organizations will need to reskill growing numbers of employees to manage AI agents or end-to-end agentic workflows. Reskilling should be part of a larger, ongoing cultural shift that positions agentic automation as a way to elevate employees’ capabilities rather than replace them. Successful ZBPR transformations will depend heavily on compliance, governance and change management to ensure that employees and agents work together.From AI-assisted to AI-orchestrated value
As organizations build out agentic workflows, change their culture and reskill their employees, they may benefit from creating a center of excellence for automation that tracks value at each step of the agentic transformation. A center of excellence can also help develop the next iteration of agentic AI value creation, whatever form that may take. For now, however, the key fact is that the future of processes and workflows isn’t simply AI-assisted: It’s AI-orchestrated and largely self-managing if organizations are bold enough to reimagine the way they work.Enjoyed this article? Sign up for our newsletter to receive regular insights and stay connected.