In August 2022, four months before ChatGPT’s launch, we predicted AI would change content marketing — and still its impact caught us off guard when AI really hit.
Our AI Journey: Lessons from Failed Experiments and Where We Are Now
We knew rejecting AI completely wasn’t an option. It represented the biggest shift our industry had ever seen. If clients weren’t turning to AI themselves, they expected us to use it to their advantage.
It felt like the invention of the power drill in a world full of hand-wielded hammers. Either we pick up this new tool or get left behind building the old-fashioned way.
We started with small-scale experiments, but soon realized we had to rethink everything about our agency’s operations — an entirely unique challenge.
Here’s our journey so far: what we’ve learned, where we’re going, and why we’ve put someone with a reputation for editorial quality, not code, in charge of our AI content operations. [Editor’s note: spoiler alert — it’s Nathan himself.]
Our AI Reality Check (2023)
We started our AI journey where most companies do: chasing efficiency.
In early 2023, we assembled a specialized team armed with the latest AI tools — Jasper, Writer, and other GPT-powered platforms. We used those tools to offer clients AI-powered content at a fraction of our usual rate.
The idea was higher volume, lower costs. We expected to create more content in less time. And the math looked promising: lower production costs would offset the reduced pricing, and clients would get more content for their budget.
The reality was different.
Client expectations were high
We tried to set quality expectations upfront, but customers measured our AI-generated content against our usual high standards. And to make things worse, the gap between AI drafts and publishable content proved wider than we anticipated. Our team spent hours wrestling AI-generated content into shape, adding nuanced insights and matching client voices. We didn't just pay extra for AI tools — we paid for the extensive human cleanup required to meet our quality bar.
This experiment taught us firsthand what we later published about AI's substantial weaknesses: it creates fiction and presents it with absolute certainty, struggles with context, and can't understand how particular ideas fit into the bigger picture. Most crucially, at that time, it couldn’t conduct research, form opinions, or share real experiences — though it pretended to do all three.
The 2023 models weren’t what they are today
Tim Metz, our Director of Marketing, says about AI’s quality during that time: “From a distance, it looks good. But then you start to read it and you realize it doesn't mean anything.” We saw the future potential, but the technology just wasn’t good enough yet to fully integrate AI into the business.
Those weaknesses had real financial impact. We charged less for this content but invested more time fixing AI's limitations. Instead of finding efficiency, we subsidized mediocre first drafts with expensive editorial time.
Yet the AI models kept improving, and so we tried again.
Our Alternative Approach (2023 - 2024)
Our second experiment took the opposite approach. We encouraged an ad hoc adoption of AI tools across our editorial team with no real strategy or guardrails.
We gave the entire team access to OpenAI and Claude and hoped organic experimentation would reveal the best ways to use AI. This bottom-up strategy seemed less risky than our first attempt. Let the team discover what works, we thought, and best practices would emerge naturally.
But without a unified strategy, we created dozens of different approaches with inconsistent results. Some writers used AI for research, others for outlining, and some generated full first drafts.
And while everyone found it to be a personal force multiplier, those efficiencies didn’t scale without a systematic way to capture and validate them.
We learned AI has no inherent value — you won't see ROI just because you have AI. Like any tool, its worth comes from how skillfully you wield it and how well you integrate it into your existing processes.
Our Integrated AI Approach (2024 - 2025)
That insight shaped our current approach: we’re integrating AI into many of our processes, sometimes redesigning them from the ground up.
We’ve evaluated and selected a single AI platform for building our workflows. We've methodically examined each part of our content creation process, identifying specific points where AI reduces time without compromising quality.
We’ve started with two types of content:
- SEO content, because it follows clear patterns: identify search intent, analyze gaps, create briefs, answer user questions. This prescriptive nature makes it perfect for testing AI integration. Early results show specific gains: faster keyword research leads to more comprehensive gap analysis. AI-assisted brief creation quickly surfaces valuable angles from top-performing content. Each improvement amplifies our existing expertise rather than replacing it.
- Social thought leadership content, because the process is replicable and requires some amount of volume to do well, something AI is adept at. Writers transform voice memos and interviews into LinkedIn posts faster, helping thought leaders maintain consistent output while preserving their authentic voice. The original thinking stays human — AI just accelerates the journey from idea to published post.
Of course, this is an iterative process. We still have open questions about how far AI can go (or how far we can take AI), how we’ll measure a deeper ROI, and how to not just maintain but increase quality.
Our Ongoing Quality Obsession (2025)
More than anything, we’ve learned that AI integration demands an understanding of content quality, which is why I’ve stepped into a new role as Director of AI Content Operations [Editor’s note: And not an AI tech automation wizard.] This role represents our belief that AI implementation should be guided by content expertise and an obsession over quality, not just technical knowledge.
I oversaw our quality standards for several years. First as editor, then as our managing editor, and the goal was to always improve content for the human reader. We now bring that same rigor to our AI implementation.
We document, measure, and improve. Our new platform allows us to document workflows, measure outcomes, and build repeatable processes that scale across our team. As we learned, random efficiency gains mean nothing without a framework to replicate them.
We experiment (with our own marketing as "customer zero"). An example is our recent deep dive into ChatGPT Pro examined whether its $200 monthly price tag delivers tangible value for content teams. These aren't academic exercises — we need to understand exactly where and how AI improves our work.
We rethink what’s possible because of AI. Case in point: our team (with no coding background) just built a free SEO traffic forecasting tool that launches next week.
We share our learnings. We’ll continue to talk about what we learn on our AI journey here on the blog, but also on other channels.
We're launching a new season of the Animalz podcast in February focused on AI in content marketing titled “AI & Content: Hello, is there any value out there?”. We've gone on a search for the real pioneers — the ones who've ventured beyond the hype to succeed (or fail) spectacularly. Through their hard-won insights, we'll discover if there's actually something of value hiding in the noise, or if we're all just shouting into the void.
Our AI Strategy is Long-term
You can slap a “we use AI!” sticker on your product, but it’s a red herring unless you can prove meaningful change — improved speed, more consistent quality, and a bigger creative payoff.
Most AI initiatives fail because companies fixate on having the technology instead of asking, “What are we trying to solve?” or “How will AI help us achieve better outcomes?” Sure, efficiency gains are nice, but quality can’t stagnate or suffer, and so you also want to spend the time gained improving the output.
At Animalz, we believe AI has reached a tipping point where it’s now delivering real benefits in quality and efficiency: faster content delivery, better research capabilities, and more time for strategic thinking.
But none of that value resides in AI alone — it comes from how we humans apply AI, integrate it into our workflows, and back it with the hard-won expertise of our strategists, writers, and editors. Like a power drill in the hands of an experienced contractor, AI finally helps us build better, faster, stronger.
Share
