Amazon Ads Director Says 2 Percent Fully Automate. The Other 98 Percent Are Correct. Frankly.

Amazon Ads director Kapil Sharma reports that only 2 percent of Indian SMBs have used AI to fully automate advertising decisions without human review. AI is nonetheless reshaping how SMBs experiment with advertising by allowing them to test more ad variants and optimize campaigns faster at reduced cost. Human decision-making remains central to the process, with AI serving as an experimentation accelerant rather than a replacement.

The principle here is human-in-the-loop augmentation. The mechanism is variance expansion. AI lets you generate and test far more ad variants than any human team could manually produce, which increases the probability of finding a high-performing combination. But the 2 percent figure reveals something important. Full automation without oversight is still considered reckless by the people actually doing this well. You let the machine generate options. You let the human choose. That is the correct division of labor.

Amazon Ads director Kapil Sharma shared data on Indian SMB advertising practices, noting the 2 percent full-automation figure and the broader trend of AI-assisted ad experimentation among smaller advertisers.