For many marketers, 2025 seemed to only get busier as the year went on. With 76% reporting their jobs had grown more challenging and 58% feeling overwhelmed.
This was in large part, due to the fact that this was the first year where marketers actually started to see somewhat useful output from their AI counterparts, with 85% achieving faster campaign execution. However, rather than giving us all more time to pause and adjust, as these capabilities materialised, pressure only increased to keep pace.
By looking back at the year that was, hopefully we can take stock of the moment in which we find ourselves and set up some guardrails going into 2026 to protect our teams, our clients, and our sanity.
Slop was the word of the year
As Merriam-Webster put it, “our human editors have chosen slop as the 2025 word of the year” defining it as “digital content of low quality produced usually in quantity by means of artificial intelligence.” This year, generative AI made below-average content almost too easy to produce at scale, resulting in brands getting carried away in a sea of sameness that ate away at brand perception.
It’s been said many times but if frontier AI labs like OpenAI aren’t using these tools for their own marketing, instead focusing on human-led content, that should give us all pause. Just because we can, doesn’t mean we should. AI content will become great, some already is, but for now, we should aim to use it to make work better, not just cheaper.
As Co-Intelligence author Ethan Mollick noted: “The traditional metrics of time saved or costs reduced may miss the more transformative impacts of these systems.”
Instead, we’ve been inundated with brain rot and uncanny valley deepfakes. One recent study found 20% of videos shown to new YouTube users were AI slop. It’s cheap, easy, and racking up billions of views. It’s a bizarre meeting of two macro AI trends, namely the broader quasi-geopolitical race to AGI and the algorithmic race to the bottom of the brain stem.
What made all this slop possible?
1- We reached human parity with visual assets. Google’s Nano Banana Pro delivered the biggest visual leap, with video models like Kling 2.6 and Veo 3.1 bringing these visuals to life. But 95% of the work is getting the last 5% right, and most AI content still only delivers only the bare minimum. The next unlock will be multi-modal world modals with enough human and cinematographic understanding to deliver emotion.
2- Marketing budgets flatlined. 59% of CMOs reported insufficient budget at 7.7% of revenue for the second year running. Many resorted to AI to cut costs, a paradox where demand for niche skills skyrocketed while perceived value plummeted.
3- We started marketing to AIs. Search traffic dropped 15% YoY, while ChatGPT hit 800M WAU. Marketing’s battlefield moved inside the model, spawning the GEO industry and automated content pipelines aimed at being served as the top option within these platforms.
4- Automation flows crawled out of their POC era. 2025 was when agentic content automation actually worked. The internet flooded with replicable n8n flows, letting anyone set up large-scale automated content generation. These style of automations allowing one state-funded network to produce 3.6 million articles in a single year.
5- Reasoning models made AI strategy viable. O3 models were huge for strategic thinking, research time plummeted and logic became commoditized. But great ad strategy is lateral, not logical. We’re still missing the friction that leads to great work, resulting in jagged intelligence that squeezes creatives into delivering premature work.
Workslop ate away at our time
Workslop, as you may have guessed, is “AI generated work that masquerades as good work, but lacks substance to meaningfully advance a task.” While some employees used AI to improve output, others used it to shirk responsibility. Pushing the burden onto receivers to decipher, despite missing context, rationale, or clarity.
A Harvard study found that 40% of workers who received workslop spent almost two hours deciphering each instance. As anyone in an agency in 2025 will tell you, this made for incredibly verbose client emails, convoluted briefings, and juniors who took “done is better than perfect” to new extremes.
Please consider this a PSA: there is no word count to meet. Send the dot points, not the GPT essay. In 2026, brevity will be a winning differentiator.
Is all this completely unsloppable?
Right now audiences are pushing back. But this won’t last. Even as ads like the infamous McDonald’s Netherlands commercial get revoked, these receding tides are temporary. We could all soon be washed away by content automation, with infinite slop machines like Zuck’s Meta AI App and his $76bn push to automate away agencies.
But it doesn’t have to be that way. Marketing has always been about zigging when others zag. Use AI in all your processes but again aim for better, not just cheaper.
Looking forward at 2026
1- We’ll all get busier. Despite what Bill Gates might have you think, AI won’t give us three-day work weeks, at least the data doesn’t show that. Once each of us is able to do more, we will in turn drive each other to also do more. We can’t afford not to. When even Andrej Karpathy (co-founder OpenAI) tweets about never having felt more behind, this is likely the permanent state of things going forwards.
2- Everything gets smarter. 2026 will see platforms finally deliver useful, contextual intelligence. Shopify’s recent Sidekick assistant for example lets users query data and forecast at unprecedented speed. Hopefully these kinds of jumps make for better work and less slop.
3- Cursor for marketing is born. Tools like Cursor changed coding forever. This year, vibe-marketing platforms will come into their own. Data-driven, end-to-end marketing creation at scale. Think Brandwatch meets Adaily, meets Weavy, meets Creatomate, etc.
4- Working will seem like gaming. With multiple research agents running, browser agents updating decks, and models rendering mood films, work will start to look more like playing Starcraft, high actions per minute, intense multitasking, and constant rapid context switching.
5- Community, network and the validation of reality. AI content has slid from cool to cringe and will pop back into normalcy soon. But most audiences will accept it with natural mistrust. Content featuring real, verifiable people and the natural chaos of life will matter more than great visuals. For a time, the behind the scenes will matter more than the scene.
Staying sane
Look, maybe I’m just a grumpy old man fed up with short-term, short-sighted decisions. And maybe slop is just the necessary crawling before amazing content can run.
Either way, throughout 2026 we’ll likely all need to regularly touch grass and smile at a stranger. The road to level 5 AI (models that can do the work of organizations), will no doubt become increasingly demanding on all of us.
We can however protect ourselves with some up-front boundaries: for instance the brand makes 70% human-led content, 20% augmented, 10% automated. Or my team works max 10% overtime. Or I’ll spend one hour weekly learning new AI tools. Whatever you commit, treat it as non-negotiable. As these boundaries will count for everything as the pressure to deliver better, faster, cheaper grows. My hope is that we all put cheaper and faster to the side for a moment and just try to make better work. The efficiencies will follow.
We clearly already have (what we previously deemed to be) AGI, albeit jagged. The next step (adding infinite learning and memory) won’t be far from ASI. And if we truly create superintelligence, maybe just maybe we’ll also get super-creative. The best films you’ve ever seen, your favorite songs created for your exact biology, art that brings you to tears at the mere thought, may all lay ahead.
The challenge will be protecting our brands, our clients, our teams, and our sanity along the way to that creativity… or whatever comes next.
