Actually Useful #25
AI tools that actually work at work | Tuesday, May 26, 2026 | 4 min read
One person is managing 77 social media channels across 10 languages — without a developer, without a large team, and without losing their mind. That’s the workflow worth understanding this week.
Tool of the Issue
How one non-developer manages 77 social channels in 10 languages using Buffer’s API
Best for: Social media managers and ops teams handling multi-channel publishing
Buffer published a case study on Atena, a social media manager running 77 channels across 10 languages using a no-code automation setup built on n8n and Buffer’s API. The short version: instead of logging into dozens of accounts manually, Atena built automated workflows that route content to the right channels at the right time — no engineering background required.
What makes this worth paying attention to is the specificity. 77 channels and 10 languages isn’t a vague “scaled our social presence” claim — it’s a number you can hold up against your own situation and ask: how are we doing this today? If the answer involves a lot of tabs, copy-pasting, and someone staying late on Fridays, this approach is worth exploring.
n8n is a workflow automation tool similar to Zapier but with more flexibility for complex routing logic. Buffer’s API is well-documented and stable. Neither requires you to write code to get started.
The honest trade-off: Buffer published this case study themselves, so there’s an inherent promotional angle. There’s no independent breakdown of time saved or ROI, and we don’t know how long the setup took to build. But the core claim — that a non-developer can automate multi-channel publishing at real scale using these tools — is credible and testable. If you manage more than 10 channels, this is worth an afternoon of investigation.
Bottom line: Try it. Start with your highest-volume channels and one language, then expand.
ClickUp laid off hundreds of employees and says it replaced them with AI agents
Best for: Managers and executives thinking about team structure and AI adoption
ClickUp — the project management platform used by millions of teams — made headlines recently after a mass layoff, with the company stating it replaced hundreds of employees with thousands of AI agents. TechCrunch covered the announcement, and it’s been circulating in operations and HR circles ever since.
This matters because ClickUp isn’t a startup running a pilot. It’s a well-funded company with a large customer base making a real operational decision at scale. That puts it in a different category from most AI-replacement stories, which tend to be either speculative or anecdotal.
What we don’t know is significant: which roles were replaced, what “thousands of AI agents” actually means in practice, and whether those agents are genuinely performing the work or primarily augmenting a smaller remaining team. “AI agents” can mean anything from a fully autonomous workflow to a slightly smarter autocomplete. The lack of operational detail in the announcement makes it hard to evaluate.
What we do know is that this is the direction several large software companies are moving. ClickUp’s decision is less a blueprint and more a signal — the question isn’t whether AI will change headcount decisions in your industry, but when and in what form.
Bottom line: Don’t copy this move without understanding what it actually involved. Do use it as a prompt to audit which repetitive, high-volume tasks on your team could be partially automated — before someone else makes that decision for you.
Workflow Win
The regulatory window is open. It won’t stay open forever.
Something interesting happened in the last few weeks that most teams aren’t tracking closely enough. Big Tech successfully shaped Trump’s AI executive order to minimize oversight requirements — a significant policy win for companies that want to move fast on AI deployment without formal guardrails. At the same time, Pope Leo XIV released a manifesto calling for robust AI regulation and criticizing what he described as a “culture of power” in the technology industry. Europe’s startup ecosystem is growing, partly fueled by AI, and Huawei announced a chip design breakthrough that signals China is closing the hardware gap faster than many expected.
These four things look unrelated. They’re not.
What they add up to is this: US companies are operating in the most permissive AI regulatory environment they’re likely to ever have — and the pressure building from outside that environment is real, organized, and accelerating.
The counterargument from tech companies is reasonable: light-touch regulation enables faster innovation and keeps the US competitive against China. Market incentives already push companies toward responsible behavior. The Pope’s manifesto reflects institutional caution, not practical governance. And European startups are succeeding despite regulation, not because of it.
All of that may be true. But the geopolitical picture complicates it. Huawei’s chip progress means the US hardware advantage is narrowing. Europe’s AI startup surge means the innovation advantage isn’t guaranteed either. And when religious institutions, international competitors, and alternative regulatory models are all pointing in the same direction, the window for “move fast, figure out compliance later” tends to close faster than expected.
Here’s what this means if you’re running an ops, marketing, or product team: the AI tools and workflows you’re building right now are being built in unusually permissive conditions. That’s not a reason to slow down — it’s a reason to build thoughtfully. Products and processes that would survive stricter data handling requirements, clearer disclosure rules, or international compliance standards are worth more long-term than ones that depend on the current environment staying exactly as it is.
The teams that will be fine when the rules change are the ones that didn’t wait to be told what the rules were.
Skip This One
Cox Media’s “Active Listening” ad targeting — not worth touching
An ad-tech firm claimed it could target users based on phone microphone data. It couldn’t. The FTC noticed.
Cox Media Group marketed a product they called “Active Listening” — the pitch was that it could detect consumer intent by analyzing audio from users’ device microphones and use that data for ad targeting. It was a striking claim. It was also false.
The FTC took enforcement action and issued fines. Cox Media later admitted the marketing overstated what the product could actually do. The core capability — secretly listening to users for ad targeting — wasn’t real. And even if it had been real, it would have been a significant privacy violation.
The lesson here isn’t complicated: don’t market capabilities you don’t have, especially ones that involve user data and surveillance. The short-term attention from a bold claim isn’t worth the regulatory exposure, the reputational damage, or the legal liability that follows. Customers and regulators will eventually test the claim, and “we overstated it in the marketing” is not a defense that holds up well.
This one isn’t worth revisiting. The product category — covert behavioral targeting based on ambient audio — has no legitimate path forward regardless of whether the underlying technology improves.
If your team is building AI workflows right now, ask yourself one question: would these hold up if the rules got stricter next year?
Tested by AI. Approved by people who just want to get work done.
Know someone who said “we should be using AI more” in a meeting last week? Forward this.
On Substack? A quick Recommend helps more teams find us — takes 2 seconds.
Want the skeptic’s take on what’s real vs. hype?
Overfitted cuts through AI noise with data and zero patience for buzzwords. Mon/Wed/Fri.
Build your own?
The Autonomous AI Newsletter Pipeline Guide: GumRoad or Dodo
How we source, classify, synthesize, and publish at scale.