That is the “AI time dividend”—in theory. In practice, it has proved difficult to cash it in. One factor is that AI adoption is highly uneven. Some people are not using any AI tools; others are using them to save double‑digit hours a week. A second factor is that AI’s potential is not being fully realized. A recent survey of CEOs and senior executives found that while AI saved an average of 5.7 hours per employee per week, only 1.7 of those hours were redirected to work that improved business outcomes. 

It is easy enough to use AI to improve how specific tasks are done. Substantial value creation, however, will only come from reconfiguring an organization around AI, rethinking design practices and operating models. That will not be easy: traditional org charts don’t have boxes for “Unassigned Saved Time” or “Ad-hoc Strategic Initiatives.” 

Moreover, employees have little incentive to risk stepping out of their boxes. In a global survey in 2024 of more than 17,000 workers, nearly half said they would feel uncomfortable telling their manager they used AI to speed up a task. 

For executives, though, how to scale up AI is top of mind. Here is how to get started on creating value from the AI time dividend.

Develop a blueprint for the future. Start by defining where technology can drive as much automation as possible.  Map these new organizational possibilities onto the operating domains that generate the most economic value.  Based on this blueprint, prioritize the where, when and how to reconfigure the organization based on an assessment of value, technical feasibility, and degree of change management needed.

To advance this transformation, change can be categorized into two levels.  In “Level 1,” AI augments work activities, such as responding to emails or speeding up decision-making; the time saved can boost productivity by up to 20%. Because Level 1 applies to existing ways of working and structures, however, it can only go so far. In “Level 2,” the organization takes full advantage. Existing linear and sequential workflows can be replaced by teams of parallel processing AI agents managed; these are managed by people, who have broader spans of control than under legacy ways of working.  The potential to boost growth and productivity is massive.

Leaders need to decide where they are comfortable with the modest benefits of Level 1 and where they want to pursue Level 2: this is one of the greatest resource reallocation challenges of our time. 

Clarify AI’s role. Trust and clarity are critical to getting AI deployed on a scale that matters. Without explicit permission to use AI, support to develop related skills, and reassurances that efficiency won’t be punished with layoffs or unrealistic demands, companies will not reap the benefits AI can bring. To start, be explicit about acceptable use, risk, and how quality is measured. This removes ambiguity that could otherwise push employees to play it safe; it also helps leaders see where to double down on tooling, data access, and enablement. Training is essential—and, at the moment, hit or miss. In a global survey, 61% said they had received less than five hours of AI-related training. 

Set clear expectations. It is the responsibility of CEOs and managers to direct time toward the company’s goals, and to identify high-potential change agents who can discern how different parts of the organization can create value. No two organizations will make the same choices, but in every case, it is important to be intentional, so that employees believe that saving time is rewarded, not penalized. Transparency is critical: Spell out how time savings translate into benefits such as higher win rates, more skills, new career opportunities, and, where appropriate, compensation. 

Make time valuable. Executives are accustomed to reallocating capital and headcount; they should add the time freed up by AI to that list. Do not leave those hours to chance; instead, create mechanisms that productively redeploy them. Some forward-thinking companies have implemented formal time reallocation programs alongside their AI rollouts, so that employees can go where the work is. Examples include time‑savings dashboards at the team level to show where hours are being freed and how they are being reinvested. Internal gig marketplaces allow people to spend freed-up hours on projects outside their usual role. Monthly innovation days bring attention to new ideas.  

Think outcomes, not hours. Saving and reallocating hours is a means; the end is better decisions and faster growth. Incentives should be aligned accordingly, for example by rewarding teams for improved AI-driven business outcomes such as customer satisfaction, revenue per seller, or cycle‑time reduction.  Convert some of the time dividend into benefits—bonus pools, career‑advancing stretch roles, or fewer meetings—to cultivate a sense of self-interest in promoting these outcomes. 

                                             🞐  🞐  🞐

In addition to strategy, financial management, and business development, today’s leaders need to become experts on time allocation. When clear expectations, trustworthy incentives, and innovative structures are in place, the next time a smart algorithm saves an hour of work, both the employee and the company know exactly how that hour can be put to good use. That cannot be left to chance: time is a terrible thing to waste.

The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.