“[Generative AI] is an aggregated view of all we know,” said Tess Kornfield, vice president of product and data science at ThredUP, at our Outlook and Strategies for 2024’s Second Half EMARKETER Summit. Issues with generative AI aren’t always a result of the tech itself. Often, AI exposes existing issues which are exacerbated by the tech’s scale and efficiency. Here are three areas marketers should watch when working with generative AI.

Weakness No. 1—Data Hygiene: AI’s potential for personalization at scale is great, allowing for one message to translate across audiences. But AI is only as good as its input.

“Garbage in, garbage out. You have to invest in your data,” said Kornfield. While AI is valuable for personalization, without high-quality customer data, it’s impossible to know how to personalize messages.

Marketers looking to use AI to scale up personalization need a first-party data strategy in place, through data collection or partnerships.

Weakness No. 2—Analytical Thinking: Even if AI is working with high-quality data, users are still at risk of misinterpreting data, said Todd Hassenfelt, global digital commerce senior director at Colgate-Palmolive. Hassenfelt emphasized the tech’s ability to summarize things like earnings calls across quarters, years, categories, and countries. But marketers need to interpret that analysis correctly.

“The genAI only knows what we tell it,” said Jennifer Faraci, chief data officer at Digitas. “It doesn’t understand context. It doesn't understand things like recency.” While marketers should be experimenting with generative AI to analyze performance data, they need to make sure they’re paying close attention to output, as they would for output from a standard spreadsheet.

“A human always has to be in the loop,” said Hassenfelt.

Weakness No. 3—Flawed Processes: Generative AI can also highlight broken processes, said Hassenfelt. “It may just be putting a spotlight on something that currently needs to be optimized within your organization.” AI’s speed and efficiency will make broken processes such as poor communication between teams or lack of specifications for a desired ad design even clearer. Use that as an opportunity to identify and fix those break downs.

This was originally featured in the EMARKETER Daily newsletter. For more marketing insights, statistics, and trends, subscribe here.