The promise of artificial intelligence often collides with a stark reality: when an organization declares that AI is everybodyโ€™s job, the initiative tends to stall. This paradox surfaces because diffusion of responsibility removes clear accountability, leaving no one empowered to drive experimentation or allocate dedicated effort. At the recent Zapier Outpost Windsor gathering, leaders shared a cautionary tale from the era of robotic process automation, where a wellโ€‘intentioned pledge to shield 20โ€ฏ% of an AI championโ€™s calendar was rescinded after just a month, as competing priorities reclaimed that time. The episode underscored a fundamental lesson: without explicit protection of time and visible recognition of the work, even the most enthusiastic advocates see their momentum evaporate. Participants left the retreat with a renewed focus on how to carve out protected space for those who can translate AI curiosity into tangible outcomes, and why that protection must be backed by concrete incentives, clear metrics, and unwavering managerial support.

The retreat, held at Oakley Court on the Thames from September 8 to 10, brought together a select group of operations and technology professionals whose travel, lodging, and meals were sponsored by Zapier. Facilitated by Philip Lakin, the companyโ€™s director of AI transformation, and Alex Maxwell, who leads the AI Leaders Lab, the sessions moved beyond theoretical frameworks into concrete stories of what works and what fails when organizations try to embed intelligent automation into daily workflows. Lakin opened with a blunt observation: unless leaders actively create the time they claim to value, back it with compensation structures, and signal its importance through incentives, the initiative will never gain traction. His candor set the tone for a day of frank discussion about the structural and cultural levers that determine whether AI experiments become isolated pilots or scalable capabilities that reshape how work gets done.

A core insight that emerged was the direct link between time protection and perceived value. Lakin argued that merely announcing an AI goal is insufficient; organizations must demonstrate commitment by removing a measurable slice of an employeeโ€™s existing duties and shielding that slice from encroachment. When leaders fail to honor that promiseโ€”as illustrated by the anecdote of the reclaimed 20โ€ฏ%โ€”they send a signal that the initiative is expendable, which in turn discourages investment of cognitive energy and experimentation. The remedy, according to Lakin, lies in aligning three elements: what people are paid for, what performance metrics reward, and how calendars are structured. When these levers move in concert, champions receive the legitimacy they need to persist, iterate, and ultimately showcase results that justify continued support.

Zapierโ€™s own internal architecture offers a living example of this principle in action. Members of the AI transformation guild remain embedded in their respective lines of business, yet the company carves out 30โ€ฏ% to 40โ€ฏ% of their workload for dedicated transformation activities. This reallocation is not a vague aspiration; it is accompanied by a formal reduction of their other responsibilities, ensuring the protected time is real and defensible. The guild role is deliberately triangulated: champions spend time listening to endโ€‘users to uncover pain points, they build prototypes or workflows to test solutions, and they translate the broader AI mandate into concrete, measurable outcomes that their teams can track. By integrating listening, building, and measurement into a single role, Zapier creates a feedback loop that turns user insight into demonstrable value while keeping the champion anchored in the business context they seek to improve.

Conversely, attempts to democratize AI by assigning a small fractionโ€”say 10โ€ฏ%โ€”of every employeeโ€™s time to the technology often backfire. When the expectation is spread thinly across the workforce, no individual feels a strong sense of ownership, and the lack of dedicated focus makes it difficult to develop the depth of expertise needed to spot subtle opportunities or navigate technical hurdles. Employees quickly perceive the request as an additional burden rather than a strategic investment, leading to superficial compliance or outright resistance. Lakinโ€™s succinct summaryโ€”’When everyoneโ€™s responsible for something, no oneโ€™s responsible for it’โ€”captures the essence of this pitfall. The takeaway for leaders is clear: meaningful AI adoption requires concentrated effort in the hands of a few wellโ€‘supported individuals, not a diffuse sprinkling of expectation that yields little measurable change.

The traditional sales pitch for AIโ€”emphasizing pure efficiency gains such as time saved or costs cutโ€”frequently misses the mark with the people who would actually use the tools. Frontline workers often interpret such messaging as a threat: if AI can shave hours off their tasks, does it also imply that their roles are ripe for reduction? This perception triggers anxiety, reluctance to experiment, and, in some cases, guilt over adopting technology that feels like it might replace them. Lakin proposed a reframing that centers on relief rather than mere efficiency. By positioning AI as a means to eliminate the most tedious, frustrating, or errorโ€‘prone components of a job, or as a way to make an individualโ€™s contributions more visible and impressive, the narrative shifts from threat to benefit. When workers see AI as a tool that removes the parts they hate, adoption becomes a personal win rather than an organizational abstraction.

Indeed, broader market trends reflect this duality of promise and peril. Analyses have shown that while AI can boost productivity, the gains are often redirected into additional work rather than leisure time, a dynamic noted by Metaโ€™s chief technology officer. In manufacturing sectors, Hyundaiโ€™s workforce recently voted to strike over concerns that automation would erode job security and intensify workloads. Even among younger professionals, surveys reveal a sense of guilt when using AI at work, as if leveraging the technology constitutes cheating or a betrayal of effort. These reactions highlight that any AI strategy must address the human dimension headโ€‘on: acknowledging fears, providing clear pathways for skill development, and ensuring that productivity improvements translate into genuine improvements in workโ€‘life balance or job enrichment, not merely higher output expectations.

Identifying the individuals worthy of champion investment becomes a practical challenge when most teams already operate at or beyond capacity. Grady Sokaras, Zapierโ€™s head of sales, outlined two complementary pathways for talent spotting. The first route involves hiring candidates who already possess a proven track record of building automation or AIโ€‘driven solutions, thereby bringing immediate expertise into the organization. The second, more developmental approach consists of gathering a broad crossโ€‘section of employees for a short, handsโ€‘on workshop, observing who gravitates toward the material, who asks probing questions, and who begins to prototype solutions organically. Those who demonstrate genuine curiosity and aptitude can then be earmarked for deeper training, with their roles gradually reshaped to allocate a substantial portion of their time to transformation work.

Zapier puts this methodology into practice through a tightly curated 90โ€‘minute session that walks a small group through the creation of their first workflow using the companyโ€™s platform. The brevity of the exercise ensures low barrier to entry, while the tangible outputโ€”a functioning automationโ€”provides an instant sense of achievement. Facilitators watch closely for the spark of engagement: participants who linger after the session, who tweak their workflows out of pure interest, or who start envisioning how the tool could solve other problems in their domain. Once identified, these individuals receive targeted instruction, progressively taking on more complex projects, and ultimately seeing their transformation responsibilities grow to 30โ€ฏ%โ€“40โ€ฏ% of their roleโ€”or, in some cases, becoming fullโ€‘time specialists. The underlying theory is simple: people who have built something, even a prototype, develop an intuition for where friction exists, enabling them to spot higherโ€‘value problems even if they never move the prototype into production.

Yet building a solution is only half the battle; the innovation must also be visible and valued within the organization. Lakin stressed that, particularly in operations, excellent work does not advertise itself. Champions therefore need to adopt a marketerโ€™s mindset, translating technical achievements into stories that resonate with stakeholders, highlighting metrics such as reduced processing time, error reduction, or freedโ€‘up capacity for higherโ€‘order tasks. By presenting concrete beforeโ€‘andโ€‘after numbersโ€”turning โ€˜this numberโ€™ into โ€˜that numberโ€™โ€”they give managers a rational basis to protect and even expand the championโ€™s allocated time. Evidence of impact transforms abstract goodwill into a compelling business case, making it far harder for leadership to retract support when competing demands arise.

To systematically diagnose where an AI initiative might be stuck, Lakin proposed a threeโ€‘layer framework that moves from the grassroots to the executive suite. At the base level, the focus is on enablement and skill: if frontline employees are not surfacing ideas to the AI guild, the problem likely lies in insufficient training, lack of access to tools, or a culture that does not reward experimentation. The middle layer examines crossโ€‘functional flow: even when numerous suggestions emerge, they may stall because data silos prevent integration, or because each department develops its own variant of a processโ€”say, ten different ticketโ€‘triage workflowsโ€”rather than a unified pattern. Here, the need is for someone to synthesize disparate inputs into a standardized approach that can scale. Finally, the top layer evaluates strategic alignment: when the lower two layers function well yet no change is observed, the root cause may be an overly vague AI mandate or a failure of viceโ€‘presidents, directors, and senior managers to translate that mandate into measurable objectives that influence their own performance incentives and reward structures.

Armed with these insights, leaders can move from theory to concrete action. First, secure explicit time protection for prospective AI champions by negotiating a defined percentageโ€”ideally 30โ€ฏ% to 40โ€ฏ%โ€”of their role and formalizing that commitment in job descriptions and performance agreements. Second, implement a lightweight observation workshop, such as Zapierโ€™s 90โ€‘minute flowโ€‘building session, to identify innate talent and enthusiasm; then invest in targeted training that scales their responsibilities as competence grows. Third, institute a regular cadence of outcome reporting where champions translate their work into clear, comparable metrics that connect directly to business goals, thereby giving managers the data they need to justify continued investment. Fourth, address the human side of AI by framing initiatives around reliefโ€”eliminating hated tasks, enhancing visibility, and fostering skill growthโ€”while openly discussing concerns about job displacement and offering pathways for upskilling. Finally, apply the threeโ€‘layer diagnostic periodically to pinpoint whether bottlenecks reside in skill gaps, crossโ€‘functional coordination, or leadership alignment, and adjust tactics accordingly. By combining protected time, deliberate talent spotting, measurable impact, and a relatable narrative, organizations can transform AI from a fleeting experiment into a sustainable source of competitive advantage.