Imagine you’re the head chef at a restaurant known for its unique fusion dishes. No other place in town balances flavor and inspiration the way yours does, but you wish you could whip up meals more efficiently and get them out to your customers faster.
One day, a kitchen appliance company drops off a robot that can prepare 200 meals in the time it takes your team to make 20. You fire it up, and soon enough plates are headed out to customers at record speed.
There’s just one problem: The food the robot makes doesn’t taste like your food. The robot can follow instructions, sure, but it can’t test the sauce. It can’t step back and ask, “Does this taste too much like something you can get for cheaper down the street?” And it moves so quickly that by the time you realize you want to make an adjustment, there are 100 steaming dishes laid out in front of you, ready to be served.
When optimization tools are put to work without proper brand stewardship, the risks are real: off-brand copy, generic visuals, unclear differentiation, and, over time, erosion of brand equity.
Any marketer worth their salt will recognize this dilemma: How do you balance quantity versus quality, speed versus deliberation, automation versus a human touch? The tension is nothing new, but with the rise of tools like ChatGPT and Claude, which can output spreadsheets of copy faster than you can say “John Henry,” it’s never been more salient to the world of advertising than today.
Recently, Anthropic, the maker of Claude and Claude Code, posted a blog on their website detailing how their internal teams leverage AI tools. In a section we found particularly interesting, they described how their growth marketing team built a workflow that “processes CSV files with hundreds of ads, identifies underperformers, and … generates hundreds of new ads in minutes instead of hours.” The team’s use of AI, they went on to write, meant “saving critical time where every efficiency gain translates to better ROI.”
In growth marketing, where a well-timed pivot in response to emerging data trends can make or break your campaign, that speed is definitely mouth-watering, and the opportunity that AI offers in this space is undeniable. But the case study also raises a few follow-up questions, namely: Of the hundreds of new AI-generated ads, how many were good? How many nailed Anthropic’s brand voice, as opposed to sounding serviceable but generic? And how long did it take for human marketers to comb through the results and filter out what worked for the campaign from what didn’t?
These questions might seem incidental. After all, sheer volume plus rapid iteration will surface high performers anyway – the old “throw it all at the wall and see what sticks” method. But research shows that if creative quality isn’t prioritized, you could be handicapping yourself before the campaign even starts. A 2017 study published by Nielsen and NCSolutions found that creative is “the undisputed champ in terms of sales drivers,” accounting for almost 50% of incremental sales. Moreover, “weak creative results in weak sales lift, and … the reverse is true when the creative is strong – sales lift is higher and is affected less by media.” A follow-up study from 2023 confirmed those findings, noting that media mechanisms such as “reach, context, recency, and targeting are all ineffective if the creative is uninspired.” From this perspective, prioritizing starting with a high level of creative quality is its own form of efficiency.
Moreover, research has shown that bad ads do more than just fail to convert – they can actively undermine the perception of your brand. In a 2021 study, University of Washington researchers asked participants to evaluate a series of digital programmatic ads and share what made them “good” or “bad.”
When responding positively to ads, participants often praised the thoughtfulness or care they perceived to have gone into them. Good ads were seen as:
- Well put together and high quality
- Thoughtful, meaningful, or personalized
- Easy to understand
- Authentic and genuine
In contrast, some bad ads felt lazy or underwhelming:
- Low quality or poorly designed
- Unappealing style, colors, font, or layout
- Bland, dry, or unmemorable
Other bad ads were seen as pushing too hard:
- Designed purely to grab clicks
- Overly demanding
- “Too much” in tone or presentation
Interestingly, participants used the word “eye-catching” to describe both ads they liked and ads they disliked, showing that careful attention needs to be paid to execution, and going too far can backfire. In the case of “bad” ads, researchers wrote that “some participants felt that the poor quality of the ad reflected poorly on the product.”
Like our fusion kitchen example from earlier, balance matters a lot. Automating copy generation and other parts of your workflow might help you achieve a short-term increase in ROI, but if it comes at the cost of your unique brand voice and identity – if it sounds generic, impersonal, or lazy – in the long-term it could erode your brand and sales. And, at the end of the day, no AI ad will ever be more efficient than a brand loyalist who comes back with or without advertising dollars.
So what’s the best path forward? At TVGla, we’ve seen firsthand how useful automation can be, but we also know that getting brand positioning right can be tricky, both in a vacuum (“Will this copy line resonate with our audience?”) and in context (“Is it meaningfully different from what our competitors are doing?”).
To that end, we’ve used AI to gain efficiencies in ways that free up our creative teams to do deeper, more strategic work – not to bypass them entirely. Some recent examples include:
- Organizing client meeting notes and distilling hours of discussion into actionable creative direction.
- Identifying recurring themes in feedback about strategic goals and desired tone.
- Generating large volumes of initial ideas like concepts, taglines, and wordplay during the early “download” phase of brainstorming, for human refinement later.
- Producing quick visual inspiration or moodboards for an art style or color palette.
- Creating custom-made AI “player cards” featuring our team members to add a personalized touch to a pitch deck.
The throughline is clear: we use AI to streamline organization, briefing, and ideation, but the human touch is what ensures the final creative is on-brand, strategically sound, and emotionally resonant.
When optimization tools are woven into workflows and used by people who know the brand inside and out, the results can make life easier for everyone involved. But when they replace entire steps wholesale, without proper brand stewardship, the risks are real: off-brand copy, generic visuals, unclear differentiation, and, inevitably , erosion of brand equity.
After all, an impression is more than just a data point a user provides you when they see your ad. It’s something you leave with them after they scroll away. Whether they clicked through or not, you want that impression to be positive – and in that moment, production speed is irrelevant, and creative is everything.
AI may one day be advanced enough to create ads that are nuanced, considered, and unique without the need for human oversight – to combine ingredients for a fusion dish without anyone there to test the sauce – but we’re not there yet. In the meantime, it’s worth taking claims from AI companies with just a dash of salt.