During a recent session on AI-powered marketing and selling, I showed participants how artificial intelligence could develop different campaign approaches for the same product. One targeted young professionals, another targeted corporate buyers, while a third explored overseas Filipino workers as a potential customer segment.
When I asked which campaign they thought would perform best, many chose the OFW version. It was compelling and easy to imagine succeeding. But the OFW market was an exploratory segment. We had yet to gather enough evidence to conclude that it would outperform the others.
I reminded the participants not to let their biases make them fall in love with an idea too soon.
That exercise captured something I have increasingly noticed in my own use of AI. Generating an initial analysis has become easier. Graphics can be produced and refined with greater consistency. Presentations can be developed while maintaining a recognizable brand identity. The distance between having an idea and turning it into something tangible has become much shorter.
These are genuine productivity gains. They also raise an interesting question: When polished output becomes easier to produce, what becomes valuable?
For years, production itself was a constraint in many kinds of knowledge work. Developing several campaign concepts, analyzing information, preparing a presentation, or creating visual material required considerable time and sometimes specialized skills.
Generative AI is raising the floor of what individuals and smaller organizations can produce. Someone without extensive design resources can create respectable visuals. A small business can explore several marketing concepts instead of developing just one. An employee can turn information into an initial analysis without spending hours starting from a blank page.
At the same time, AI is raising the ceiling of what people can attempt. In my own work, it has reduced the effort required to move from an idea to an initial result. I now have more ideas that I could realistically pursue. My constraint increasingly becomes deciding which ones deserve my attention.
The bottleneck has moved.
Organizations may encounter the same shift. A marketing team capable of producing three times as much content has increased its production capacity. The more important question is whether that additional work produces better results.
How are we measuring its effectiveness?
The value of additional advertisements, reports, presentations, images, or videos ultimately depends on what they help the organization accomplish. Faster production becomes meaningful when it leads to better customer engagement, stronger decisions, improved processes, or other outcomes that matter.
This is why the process extends beyond what AI produces.
Idea → AI → Output → Test → Evidence → Learn.
The campaign exercise illustrated this well. The OFW concept might eventually prove to be an excellent market opportunity. Its polished execution made it attractive, but the market still had to validate it.
Originality can make an idea worth testing. Evidence tells us whether it deserves to survive.
There is another consequence of making production easier. The speed of AI can tempt us to accept the first competent result. Yet the time saved allows us to provide more context, inject our own ideas, question assumptions, and improve the work further.
I have discarded AI-generated material that was technically acceptable because it lacked the perspective and distinctiveness I expected from the inputs I had provided. As more people gain the ability to produce polished AI-assisted work, polish by itself becomes less distinctive.
This is becoming particularly visible as AI capabilities expand from text and images into video. Creating impressive material is becoming accessible to more people. The differentiator increasingly lies in what we bring into the process: original thinking, experience, customer understanding, domain knowledge, creativity, and the questions we choose to ask.
AI can strengthen those contributions as well. We can ask it to challenge an assumption, propose alternatives, critique an idea, and analyze the evidence that comes back. AI may increasingly help us evaluate work that AI itself helped create.
The organization still determines what evidence is sufficient, what actions a system is authorized to take, and who remains accountable for consequential decisions.
For this reason, I am more concerned about companies that are underinvesting in AI access for their employees. These tools can help people analyze information faster, communicate ideas better, explore more alternatives, and accomplish work that previously required considerably more time or resources.
Such access should come with approved tools, appropriate policies, data protection, training, and clear boundaries. Organizations need both capability and governance as AI becomes part of everyday work.
The return on that investment should also extend beyond hours saved.
If AI helps an employee finish something in two hours instead of six, what becomes possible with the four hours that were freed? Perhaps that person can serve more customers, investigate a recurring problem, improve a process, develop an idea, help colleagues, or accomplish something the organization previously lacked the capacity to pursue.
That is where the productivity opportunity becomes much more interesting. AI gives people the ability to do more. The value comes from what they choose to do with that ability.
As polished production becomes increasingly accessible, organizations will need more than the ability to generate. They will need worthwhile ideas, distinctive inputs, meaningful measures, and the discipline to test assumptions against reality.
The question may soon be less about whether AI can produce something impressive. Increasingly, it can.
A better question is what we contributed that made the work worth producing, what we learned after putting it into the world, and whether it ultimately helped someone.
