The marketing playbook that worked in 2025 is actively hurting performance in 2026, says Donatas Smailys, CEO and co-founder of Billo.
“There is always a lag between when something stops working and when brands stop doing it,” said Smailys. “Right now, that lag is costing real money.”
Here are the three marketing tactics he says are already failing:
AI chatbots that function as brand personas
Last month, Snapchat launched sponsored AI chatbots inside users’ direct messages, letting brands deploy conversational AI agents into the same inbox where people talk to their friends. They claim these to be more immediate, more personal, and more useful.
Meta tried the same with Messenger and WhatsApp, faced backlash both times, and pulled it. When Snapchat first forced its own AI chatbot to the top of all user inboxes in 2023, the app’s US App Store rating dropped to 1.67, with 75 percent of reviews rated one star, data from app intelligence firm Sensor Tower then revealed.
Snap’s own CEO, Evan Spiegel, said recently that tech leaders underestimate the societal pushback coming against AI.
“Brands are being sold a chatbot strategy as if audiences were asking for it,” said Smailys. “They were not. People do not want to have a conversation with a financial services company in the same place they talk to their friends.”
Creating AI avatars or self-deepfakes
YouTube is rolling out avatar tools that let creators clone themselves on camera.
As much as 46 percent of social media users already said they are not comfortable with brands using AI influencers, according to the Sprout Social Pulse Survey.
“AI avatars are mostly getting backlash, and there haven’t been many positive use cases. In reality, we see that real content creators, including small UGC creators, are becoming more popular than ever, so the AI avatar trend is already past its time,” Smailys says.
Research studies already showcased that AI content is “dehumanizing advertising”, and customers perceive AI-generated content, including avatars, as lacking empathy and human touch.
“The reason for AI avatars was cost saving, but the performance and trust hit is also real,” said Smailys. “Creators are trading long-term audience trust for a short-term production budget win if they’re deciding to use their own deepfakes. That is a bad trade, and the data is starting to show it.”
Regulation is catching up too: The EU AI Act and FTC are already pushing mandatory AI content disclosure, with political deepfakes acting as the forcing function.
But the legal infrastructure being built around fake political figures won’t stop at politics, Smailys says. Mandatory labeling is coming regardless of whether audiences ever warm up to the format.
Hooks are being deprioritised
For years, the formula in advertising was simple: craft a punchy three-second opener, grab attention before the skip, and engagement would follow.
It doesn’t work like this anymore, Smailys cautions. The three-second hook was a response to an algorithm that rewarded early click-through, but the algorithm now rewards different things.
TikTok’s and Instagram’s algorithms have shifted their heaviest weighting to watch time and completion rate. A video that gets skipped after a hook scores worse than a slower video that holds attention all the way through.
According to Sprout Social’s social media trends report, the brands generating genuine engagement are the ones building recognizable characters and brand worlds that audiences return to.
“Everyone learned the same three hooks from the same ten creators,” he said. “Audiences have seen all of them. The hook that worked in 2025 is now the signal that this video is not worth finishing. You cannot shortcut your way to watch time.”
As algorithms evolve and audiences grow more skeptical, Smailys says marketing is shifting away from pure speed and scale, and back toward trust, originality, and recognizable human identity.
