Gabe’s current thoughts on AI x Marketing and Creative (+rants)
—

I have a hard time sitting down to write about AI, and I know a lot of us have struggled with this in the last few months. It seems clear that the initial hype and freakout (in the creative services space at least) is dying down quite a bit, and our jobs still exist. But AI’s ability to distract those of us in charge of brands, agencies, startups, marketing departments, C-suites… that ability remains impressive.

I usually end up ranting about the wasted time and energy I see people putting into the development of agents to automate 38 rote tasks that they never struggled to complete in the past. They spend 6 weeks heads-down learning and grinding out new, but (so far) barely useful tricks, and convince themselves that they’re working on modernizing their business for the AI age. Adapt or die! They’re all in, and I think they’re simply and entirely distracted by the hype.

It’s obsessive AI-transition group-thinking that has everyone feeling like they need to re-engineer everything about their careers and businesses to be AI-first. It’s an intellectual/techy rabbit hole full of dopamine hits that may eventually lead to 4% time savings on recurring tasks. We all know people who’ve been lost to this obsession.

Well. I went from talking about how I slip into ranting to actually slipping immediately into ranting. Let’s change gears.

As business leaders, we do all need to articulate a clear stance on the aspects of AI that impact our businesses, and the businesses of our clients. For people like me, in the agency world, clients and prospects need to know how our creative team thinks about, employs, or/and avoids different types of AI input on the strategic and creative processes that serve their brands.

So here’s the snapshot, and of course each of these topics has a hundred nuances behind them. ChatGPT could tell you all about them, extensively, but I can summarize everything in one foundational 3-part statement:

The use of AI depends on the context / Strategy dictates all technology, always / The lazy use of AI is a race to the bottom for your brand

To plant a flag on a few specifics, here’s a quick Q&A on how Studio Rover thinks about specific aspects of AI in our work for clients.

Q: Do you use AI in your creative process?

A: Sometimes for concept explorations, never for final artwork. If we’re creating something that is meant to serve as a brand asset for a business, we only deliver original work created by our team.

Q: Do you use AI in your research process?

A: Hell yeah, we use it for research. It does good basic research quickly, and the conversational interfaces are way better than old school Googling and wading through results to then go wade through web pages. Fast research with AI is truly amazing. Just check anything that looks like data. LLMs are famously bad at being objectively right about data. I guess humans are too though.

Q: Do you use AI in your strategy process?

A: Nope. AI-supported research gets woven in to our strategy work, but with heavy vetting. We insist that the new thinking we’re providing to clients on strategy engagements is driven by data and insights, but not built directly on top of them without creative alignment to your brand. If we only relied on AI for strategy recommendations, we’d pass over the same bland output that LLMs recommend to everyone, including your competitors. AI helps, but can’t be the star of the strategy.

Q: Do you use AI in your production/delivery process?

A: As a tool, like any other tool (which includes every piece of software we use on our fancy Macs), if AI can bring efficiency to production work, we’ll test and then incorporate into our process. A few years into the AI tsunami though, and we still have yet to find any truly reliable AI automations in the production delivery of creative work that don’t need heavy human oversight and fixing. As long as we can do a better job, faster, we’ll leave AI out of it.

Q: Do you use AI to write copy for yourselves?

A: Nope. We’ve spent a lot of time exploring LLMs for copy development, editing and consolidation of copy - and we’ve found that it does a really good job of giving you really bland copy. It can check boxes, but it is usually heavily vanilla’d out, and can be easily spotted as inauthentic. We’d not risk our human reputations on AI copy or messaging that’s posing as real copy written by us. Writing copy ourselves takes a lot longer, but it’s so so so much better when AI isn’t driving that bus.

Q: Do you use AI to write copy for your clients?

A: No we sure don’t. If our clients wanted AI copy, they’d prompt it themselves. In fact, most of our clients use AI to work through their copy and content ideas at a high level. Sometimes they send to us without much/any editing (which we notice), and other times they either ask our writers to humanize the core intention of the messaging/copy, or they do it themselves. AI copy isn’t inherently bad, but it tends to be long-winded, and it’s coming from the same pool of average that all of your competitors are swimming in as well. If words are important to your brand, you and we can’t just farm it out to AI. We all deserve better.

Q: How about code?

A: If use case calls for it, and strategy supports it - and if the QA team knows it’s coming - of course. This falls squarely under the ‘new tools’ perspective. If AI can improve on existing code, let’s do it. And test. If AI can build a piece of code from the ground up to support the bigger picture and strategy of a custom dev project, we’ll explore and implement it if it’s good code. There’s a very good chance the human developer will need to fix and finesse the details - just like writers fix AI copy and designers fix AI visuals.

Q: How do you plan for AI’s role in a project?

A: It’s all in the brief. A good brief will set a flexible but committed framework to a project, and it’s an easy add to include a section on AI tools or processes that the agency and client team agree may be beneficial to the project. Strategy should inform all use of technology, including and especially AI, and whether it’s written down in the brief or discussed on the kickoff call, it’s always best to reach clarity and agreement on the way into a project. Clients surprising agencies with AI, or agencies surprising their clients with AI, is never a good thing. Talk it out, humans, and write stuff down.

Q: Do you help your clients build agents and AI-ify their processes?

A: Mmm, no. A lot of agencies decided that they needed to fast-track in-house expertise on agentic work and efficiency-efficiency-EFFICIENCY! for their clients as a service offering. And literally thousands of AI-transformation agencies have oozed out of the internet woodwork doing JUST this work for clients. We’ve been successful as an agency by reducing our service offerings over the years - staying very focused on what we do best, and not adding anything that’s not already in our DNA. We’re not interested in dabbling in these types of AI services, although by all means, seek out the experts for anything that might truly benefit your business. Remember though - strategy first! There’s a lot of money going down the drain on AI playtime and distracted agenticification.

Q: Are you AI Optimists or AI Pessimists (or AI Doomsdayers)?

A: Yes, yes and yes. We’re lifelong fans and proponents of good technology, put to good use. Generally speaking, we’re AI Optimists, and our inner nerds are blown away almost daily by the potential of AI technologies. AI Pessimism steps in when we see how bubbly the market is, which results in a lot of scrambly AI technology desperately trying to show that it has value. So much of it doesn’t, or at least so much of our collective approach to using AI has led to slop at astronomical levels. Both in visual/written output, but also in less visible under-the-hood automations that seem to bring zero significant value to the vibe code jockeys who haven’t slept since 2025. Or worse, the tools are used to hack (like bro-‘hack’) systems and spam humans on levels we’ve never seen before.

As for AI Doomsday scenarios - humans seem committed as a species to destroy our planet and drain our limited resources. I’m pretty sure we’ll do ourselves in before evil sentient AI overlords take over. But there’s doom coming either way. I find it amazing that we’re unleashing an AI-fueled water and power consumption crisis before we’ve even come close to solving for our reliance on fossil fuels, although we’ve known that problem was ruining the planet for decades. Might as well pile on a fresh layer of negative ecological (and therefore human) impact with our heads in the sand… because there’s money to be made! And 4% increases in efficiency on… what exactly? I’m sure someone’s work is a little easier these days. Water be damned.

I told you there’d be rants!

Good luck out there, and happy prompting.