If it could happen to one of the nation’s largest fund managers, it could happen to you.

T. Rowe Price Group Inc. accidentally voted in favor of a deal it was publicly opposed to — the 2013 buyout of Dell — because it relied on an automated system that was programmed to vote in favor of management-led initiatives.

The firm had to reimburse $194 million to customers who lost out because of the blunder. The Wall Street Journal reports that, according to analysts, T. Rowe Price is “among the publicly traded asset managers with the strongest balance sheet.”

The Risks of Automation

So how could this happen, and what does it mean to ad tech? The T. Rowe Price team could have manually overridden the computerized voting system, but it failed to check the machine’s work.

Automating marketing functions frequently causes errors in our industry. The New England Patriots let an automatic bot tweet a racial slur from its Twitter account.

Shutterfly sent congratulatory notes to “new parents” via automated emails to customers who hadn’t actually had children.

Uber probably didn’t benefit from running its banner ad below a photo of an Uber driver who tried to break into his passenger’s home.

Automated solutions still require human intervention, even in the programmatic advertising space.

Because creating multiple ad versions can be challenging and time-consuming, the industry is quick to embrace automated solutions for creative customization, and understandably so. But if we forgo human touch entirely, we risk marketing errors or, at the least, running campaigns that aren’t as effective as they could be.

How to Improve Your Automated Creative Process

An increasing number of marketers are automating their creative processes with dynamic creative optimization (DCO), a highly-automated approach, or creative management platforms (CMPs), which require more human input.

If human intervention isn’t naturally built into your workflow, you have to create a system for it. T. Rowe Price’s error resulted from someone failing to oversee a machine-based action.

Learning Opportunities

Even if your programmatic creative solution doesn’t require you to review the ads you are serving, and even if you are tempted to wash your hands of this responsibility, completely trusting in a machine is risky. Programmatic advertising benefits from human judgment.

Have someone — your art director, perhaps — review your ads to confirm they’re on brand and high-quality.

Then, keep an eye on your campaign once it is live. If you are using a DCO, watch the default rate, which is how often you serve a generic ad rather than a personalized one.

You have invested resources to understand the marketing message that will best resonate with a target group. If you don’t serve that identified message, you are missing out on programmatic’s unique value proposition and compromising ROI.

People Power

Recently, Mercedes made headlines for replacing robots on the assembly line with humans, because humans were better equipped to handle the automaker’s advanced customization needs.

Technology can make our jobs easier and improve performance, but we have to understand its limitations. You would never suggest a team member work completely independently, excused from all check-ins and team meetings.

Why trust in a machine so blindly? A machine can’t be held accountable for its actions. If something goes wrong, blame technology and see what happens!

Take responsibility for your creative process and manage your machines, or risk making headlines for all the wrong reasons.

fa-solid fa-hand-paper Learn how you can join our contributor community.