Spotlight: Awish — automations you edit by chatting, not rebuilding
Most automation tools make you build the workflow once, in a visual editor, and then treat it as a fixed thing you go back and edit with clicks. Awish's pitch is that you should be able to talk to your automations the way you'd talk to a coworker: tell it what you want in a sentence, let it wire up the apps, and then just keep talking to change it later.
The example on their site is a good stand-in for the whole product: "Send my new leads to Slack, notify me on WhatsApp, and create a weekly report in Google Sheets." Awish's agents parse that, pick tools out of a claimed 500+ app catalog, and propose a plan before anything connects. Once it's running, you don't open a dashboard to pause it or change the schedule, you just say "change trigger to every Friday" in the same chat thread and it updates. That's the interesting decision here: they're positioning the chat interface itself as the ongoing control panel, not just the onboarding step.
The other thing worth noting is where they've chosen to live: WordPress, Slack, and Telegram, explicitly, in the tagline. That's a narrower and more deliberate wedge than "connects to everything," and it makes sense as a go-to-market: those are three places where non-technical people already spend their day, so the automation shows up inside a tool they've already got open instead of asking them to adopt a new one.
On permissions, they're upfront that integrations only happen with explicit approval, which matters given how much trust you're handing over when an AI agent is choosing which apps to connect on your behalf. There's no detail on their site about what data it retains or how it handles credentials beyond that one line, so if you're evaluating this for a business with real compliance requirements, that's a question to ask directly rather than assume.
Who should try it: solo operators and small teams doing the boring connective-tissue work of a business, things like routing leads, posting notifications, or assembling a recurring report, who don't want to learn Zapier's logic or hire someone to maintain it. If you already run your operations through Slack, Telegram, or a WordPress site, the fit is more obvious because that's literally where the tagline says it lives. It's also aimed at people who find visual workflow builders more intimidating than a chat box, which is a real audience even if it doesn't get talked about much in this space.
Who should skip it: anyone with a complex, branching workflow that needs precise conditional logic, error handling, or auditability. Natural language is great at expressing intent but vague on edge cases, and it's not clear from the site how Awish handles the automation doing something you didn't quite mean. Teams already deep in Zapier, Make, or n8n with hundreds of existing zaps probably have more to lose than gain by migrating. And if your business runs on tools outside the WP/Slack/Telegram trio, you're likely testing the edges of what they currently support.
Where this goes next, as I read it: the real product isn't the automation, it's the conversation layer that manages it. If that holds up in practice, the natural next step is expanding beyond those three surfaces into more of the apps in that 500+ catalog as first-class citizens, not just endpoints an agent can technically reach. The harder problem will be trust: the more autonomy you give an agent to choose apps and adjust triggers on its own, the more you need visibility into exactly what it changed and why, and that's usually where "just talk to it" products start needing a proper audit log whether they planned for one or not.
Try Awish: awish.ai
See the launch: Awish on welaunch.sh