Generative AI is making content faster to produce, but the harder question is whether that content changes what people believe, choose and do
Recent research into AI in business writing shows that it works well as a productivity tool. A
study in Science found that access to ChatGPT reduced time spent on professional writing tasks by 40% and improved output quality by 18%. In customer support, an
NBER study found that generative AI assistance increased productivity and improved customer sentiment.
The real test, however, is what those words help people do.
A
2026 field experiment on video advertising found that AI-generated ads performed better on upper-funnel measures: higher click-through rates, stronger viewer retention and fewer negative reactions. But the same ads were less likely to convert than human-created ones. The researchers called this “attention without conversion”.
There’s a lesson here for business leaders. AI writing is useful when it helps teams clarify, structure, adapt and test ideas. It’s weaker when it replaces judgement altogether. We encourage our clients to ask whether the engagement AI delivers is actually worth anything. If it helps people understand, decide and trust, it has value. If it isn’t delivering results, it’s just noise.
