Where AI actually moves the needle in marketing

AI does not make marketing better everywhere. It makes specific work dramatically better and other work quietly worse. The teams winning with it are not using more AI. They are pointing it at the right tasks.

Start with the size of the prize. McKinsey put marketing and sales among the four functions that hold about 75% of generative AI’s total value, and estimated the marketing lift alone at 5 to 15% of total marketing spend (McKinsey, 2023). That is real money sitting in the production line, not the strategy deck.

The needle moves on the work, not the slideware.

But more AI is not the answer, and there is a careful study that proves it. In a Harvard field experiment with 758 consultants, those using AI on tasks inside its range finished about 25% faster and produced work rated over 40% higher in quality. On tasks outside its range, they were 19 points more likely to get it wrong (Harvard Business School, 2023). Same people, same tool, opposite result. The researchers called it the jagged frontier.

So where does AI earn its keep in marketing? On the high-volume, well-defined work.

Production at scale. One approved idea turned into ten on-message assets. Drafting, resizing, adapting. Fast and consistent, and squarely inside the frontier.
First drafts and synthesis. Blank page to solid starting point, which is exactly where the speed and quality gains show up.
Personalization the team could never do by hand. Tailoring by segment and journey stage at a volume manual work cannot reach.

And where it quietly costs you? The defining calls. Positioning, the core message, the strategic bet. That work sits outside the frontier, and handing it to AI is how good teams ship confident, wrong work.

Aim AI at production and it compounds. Aim it at judgment and it backfires.

This is why a system beats a loose tool here. A system points AI at the work it is good at and keeps a human on the work it is not. The brief, the message, and the final check stay human. The production runs on rails.

If your AI results feel uneven, that is the jagged frontier, not your team. Put AI on the production and your people on the judgment, and the needle moves. Start a Conversation.

Sources

McKinsey, 2023. McKinsey estimated that marketing and sales sit among the four business functions holding roughly three quarters of generative AI’s total value potential, with the marketing gain alone worth the equivalent of 5 to 15% of total marketing spend. https://www.mckinsey.com/capabilities/tech-and-ai/our-insights/the-economic-potential-of-generative-ai-the-next-productivity-frontier
Harvard Business School, 2023. A field experiment with 758 consultants found that on tasks within AI’s current range, users worked about 25% faster and produced markedly higher quality work, but on tasks beyond its range they were significantly more likely to reach the wrong answer, a pattern the researchers named the jagged frontier. https://www.hbs.edu/faculty/Pages/item.aspx?num=64700