A Good AI Plan Changes the Plan

As annual planning ramps up, leaders are increasingly being asked to define an “AI strategy.” In many organizations, this becomes a standalone planning exercise, one that runs parallel to defining the broader 2027 strategy.  

An AI lead or taskforce might set goals for AI, then identify use cases and tools to support them. The risk in this approach is that AI becomes something layered onto the plan rather than something that shapes it. Creating an AI plan for the sake of having an AI plan will likely only increase costs, decrease satisfaction, and not drive any meaningful value for your business.

The meaningful opportunity is to use AI in direct service of core business objectives. AI should empower teams to rethink how their goals are achieved, and even what goals are worth setting.

A good AI plan starts with the business strategy and improves it. This principle-driven approach is the basis of our AI planning work with clients. By working together through these five questions, we help teams build AI plans that are grounded in both ambition and execution.

1. What are we trying to accomplish?

You likely already know what your primary business objectives are: increase revenue, reach a new audience, improve service, launch a product, or reduce operating costs.

“AI transformation” is not a business objective.

It is not an achievable finish line, or an indicator that the needle has moved on any of your actual business objectives.

The best AI plan focuses on where AI could make a meaningful difference to goals the organization has already decided matters.

It also helps leadership distinguish a worthwhile initiative from an impressive demo. An AI application is only valuable in context: which goal it supports and whether its potential effect justifies the investment and organizational change required.

2. How could AI help us achieve our goals more effectively?

AI can help organizations pursue existing goals faster, at lower cost, at greater scale, and with better results.

There is no shortage of case studies to support these efficiency gains. For example, in a randomized study involving 758 Boston Consulting Group consultants, those using GPT-4 completed tasks within AI’s capabilities 25% faster, achieved 40% higher performance ratings, and ultimately finished 12% more tasks. The qualifier matters: on a task that fell outside the technology’s capabilities, access to AI made performance worse.

AI for the sake of AI will be a value detractor. The opportunity is to identify how AI can meaningfully impact the path to achieve your goals.

3. What new goals does AI make possible?

AI may even unlock new opportunities.

This can be because the economics change: a market becomes viable to serve, a service can be personalized at scale, or a new product can be tested with less time and investment. It can also be because the organization gains a capability it did not have before.

KPMG’s 2026 launch of KPMG Private illustrates the shift. For decades, the firm served large enterprises with highly sophisticated audit and advisory services. But delivering that same level of analysis to smaller companies was often too labor-intensive to do economically. AI changed those economics. By deploying AI agents that automatically analyze expense line items and reduce workpaper assembly time by 35%, KPMG was able to launch KPMG Private, expanding its product offering to reach a $18-billion market in new ways. This previously uneconomical business model was made viable by AI.

Good planning creates room to discover these possibilities. The result may be a more ambitious version of an existing goal, an adjacent opportunity, or a goal the organization would not have thought to set before.

4. What will it take to put these opportunities into practice?

Once you’ve identified how AI could advance existing goals or unlock new ones, the plan has to account for the organizational realities required to deliver:

• Technology & data
• Workflows
• Adoption
• Governance

These are the foundational pillars that determine how teams will actually move plans into action. 

An attractive idea may need to wait because it depends on inaccessible data, major integrations, or change the team cannot absorb. A less dramatic opportunity may be better prioritized because it can be actioned quickly in the organization today.

5. How will we know when to scale, adapt, or stop?

A good plan makes progress visible and creates decision points before the work begins.

For each priority, teams should define the business outcome they expect to change, establish a baseline, and identify the early signals that would show whether the initiative is moving in the right direction. Time saved may be an early signal, but it is not the end result. The next question is what business objectives are we impacting with that time. Those signals likely take longer to realize.

The plan should also define what would cause the organization to change direction. What result would make the team stop an initiative? What new capability would make a build decision worth revisiting? What assumption is the recommendation based on, and how often should leadership check whether it still holds?

This creates a practical way to know if the plan is working.

Turn AI possibility into a practical plan.

These five questions are the backbone of how we approach AI planning with clients. They turn a broad mandate to “do something with AI” into a coherent plan tied to real organizational outcomes.

Our work helping clients plan and execute AI strategies happens year-round. But as teams set priorities and budgets for next year, we’re offering a focused AI Planning Sprint to help leaders determine where AI can advance existing goals, what new goals it may unlock, and what needs to happen across technology, data, workflows, people, and governance to make those opportunities real.

In a matter of weeks, teams leave with a prioritized roadmap and execution plan grounded in the broader business strategy.

If your team is planning for 2027, we’d love to talk.