A polished presentation can arrive before the team responsible for delivering it has even discussed the problem. The argument is coherent, the recommendations are neatly organized, and the plan appears ready to move. Then the questions begin. Which conclusions reflect decisions the business has made? What evidence supports the investment? Can the teams involved execute it?
I see more leaders across industries using generative AI to develop their thinking before bringing colleagues into the conversation. I understand the appeal. A leader can shape an idea and prepare for a meeting in a fraction of the time it once took. In media planning, that speed helps us examine more scenarios and spend less time assembling a first pass. But faster preparation does not always produce faster decisions.
The problem starts when a well-written output is treated as settled. AI can give an initial assumption the language and appearance of a strategy while leaving the underlying business questions unanswered. Because these tools tend to agree with the person using them, the leader may be the last to notice. The people asked to act then have to work backward, separating what the business has agreed from what the model inferred.
Polish Can Hide an Unfinished Decision
Consider a business that has relied heavily on performance media and now wants to reach new audiences. An AI tool can quickly produce a persuasive case, suggest an allocation, and outline a measurement approach. Those are useful starting points. They do not establish how much the business can invest before seeing a return, whether those audiences offer room for growth, or what the finance team needs to understand about the timing of that return.
I have seen how difficult those conversations become when a company is under pressure to meet its next earnings target. Leaders may recognize that they should have invested in future demand years earlier, yet their immediate circumstances leave little room to change course. No amount of polish resolves that tension. The decision depends on a candid account of the business’s constraints and agreement on which outcomes matter over which period.
If those questions remain open, an AI-generated plan can make progress look further along than it is. Teams may spend several meetings challenging assumptions that would have surfaced in the first conversation. The faster draft has slowed the decision by hiding an incomplete brief.
The quality of a decision depends in part on how seriously the underlying idea is challenged. A model responds to the context and direction it receives, so it can strengthen a proposed approach even when the premise deserves scrutiny. Research from Stanford University, published in Science in March, tested 11 leading AI systems. The systems affirmed users seeking advice on interpersonal conflicts far more often than people did, and participants left those conversations more convinced they were right. The study focused on personal advice, not marketing strategy, so it does not establish the same pattern in business decisions. Still, it gives leaders reason to ask whether their use of AI invites disagreement or simply sharpens a view they already hold.
Better Decisions Start with Open Questions
In integrated media planning, objectives regularly pull investment in different directions. A business may want immediate sales while also needing to build consideration among people it does not reach today. A forecast can estimate outcomes under different allocations. It cannot decide which trade-off the business is prepared to make, especially when the answer depends on inventory, creative readiness, or expectations beyond marketing.
At PMG, integrated media planning brings those business considerations into the recommendation as it develops. Alli’s shared data foundation and forecasting capabilities support the analysis, while our planners examine what the proposed investment requires of the business and pressure-test the assumptions behind it.
If a plan depends on reaching new audiences, we need to understand whether the creative and audience strategy are right, whether the business can support the time required to build demand, and what near-term tradeoffs may be necessary to achieve the longer-term objective. We also consider the conditions and dependencies that could influence success, from timing and investment thresholds to measurement, inventory, and the broader customer experience. Resolving those questions early gives our customers a clearer understanding of not only what we recommend, but what needs to be true for it to work and how we will adapt if conditions change.
"AI can give an initial assumption the language and appearance of a strategy while leaving the underlying business questions unanswered."
Teams get the most from AI when they use it to expand the range and speed of thinking throughout the planning process. It can synthesize large amounts of information, connect signals across audiences, channels, and business inputs, develop and compare scenarios, and surface patterns or implications that may not be immediately obvious. It can also help expose gaps in the logic behind a recommendation by identifying assumptions, testing alternative explanations, or showing what different stakeholders may question. The value is not in asking AI to make the decision, but in using it to strengthen the analysis and give planners more informed choices to evaluate.
AI is most effective when it builds on what is already known. Breakthrough ideas often require teams to question familiar patterns, imagine what has not been tried, and accept the uncertainty that comes with experimentation. For brands seeking to stand out, AI can support that process, but it cannot replace the human creativity and judgment needed to pursue an original direction.
Before a recommendation reaches a broader team, people who know the customer, media environment, and operating constraints need to examine it. A confident account of audience behavior may rest on an unsupported assumption. A shift in spending may be sensible in principle but impossible to execute within the quarter.
Specialists can offer the strongest challenge while the thinking is still open. A finished-looking deck creates social and organizational pressure to focus on execution. Involving specialists while the problem is being defined gives them room to ask whether the proposed solution addresses the business need
Shared Reasoning Turns Speed Into Progress
The answer is not to abandon AI-assisted strategy work, but to carry the reasoning forward with the output. A brief should make clear which business outcome is being pursued, what evidence informs the recommendation, which assumptions remain untested, and who can resolve the outstanding trade-offs. The level of detail will vary, but the people receiving the work should be able to tell where analysis ends and judgment begins.
"For brands seeking to stand out, AI can support [the] process, but it cannot replace the human creativity and judgment needed to pursue an original direction."
Leaders can apply this discipline to the next AI-assisted recommendation before it reaches an approval meeting.
Make the unresolved questions visible. Ask the person preparing the recommendation to distinguish confirmed business decisions from assumptions that still need testing. If the plan depends on a particular return horizon or budget flexibility, state that dependency explicitly and identify who can confirm it.
Invite challenge while the recommendation can still change. Share an early version with the specialists whose knowledge could materially alter the plan. Ask what would make the recommendation impractical and which evidence would change their view. AI can help generate competing explanations, but the people responsible for delivery need to assess them against the business’s circumstances.
Record what would trigger a reassessment. Once a decision is made, document the assumptions it depends on and agree when to review them. A recommendation to invest in new audiences, for example, should specify the expected time to results and the evidence that would justify continuing or adjusting the investment.
The time saved in preparing a draft gives teams more room to examine the choices it contains. Using that time to resolve uncertainty before approval makes the recommendation easier to implement and gives the business a clearer basis for learning from its results.
