Everyone is talking about Clawdbot.

The whole Internet is on it. There are endless videos showing how it can do EVERYTHING!

Ready to jump in and be part of the revolution?

Well, I say: not yet!

This is very, very early days.

This tool is still in innovator territory. It is simply not safe.

Let me tell you exactly what it is so you're totally clued in.

Then I'll tell you why we shouldn't use it right now, and what we should do instead.

What is Clawdbot?

Clawdbot is an AI agent that runs locally on your machine or in a hosted virtual machine.

Not a subscription you sign up for like ChatGPT or Copilot, it's your own one.

Being local, it has direct access to your system, it can do anything you can do.

It is highly autonomous. No guardrails in place.

If it doesn’t know how to do a thing, it writes a little utility to help itself, installs the utility and runs it.

If it needs more support, it goes online to find it.

For example, one person set it up and sent it a voice message through WhatsApp. The bot received the message, realised it can’t listen to it, found a voice-to-text utility, convert the message, read it, and replied.

So what do people do with it?

Someone set it up with their Meta glasses to take photos of receipts, send them to WhatsApp, read them and post them as expenses.

Someone set it up to run through messy files and fill tax forms they’d been putting off for ages.

Another got it to manage their home entertainment systems. For devices that don't have a ready integration, Clawdbot just wrote the software integration, installed it, and ran it.

That's amazing, right?

People say it's the iPhone moment of AI because we're now seeing what working with autonomous AI is going to look like.

I do agree.

Note: Clawdbot has been renamed a couple of times: first to Moltbot, now to Openclaw. You might see any of these names online, but they're all the same tool.

Why shouldn't we use it?

I don’t think we should use it yet.

Why?

Well, it is so powerful because it has uncontrolled access to your world. That is phenomenal power, phenomenal risk.

Would you hand over your laptop to a genius software developer you’ve just met and let them run your life?

It's an alpha version of a tool. Deep in innovator territory. Not even early adopter.

It is far from enterprise grade.

And unless you are quite techy, not even safe for home.

One scan, 900+ unsecure setups

Within weeks of the Clawdbot launch, security researcher Jamison O'Reilly scanned the internet and found over 900 Clawdbot installations that were completely open.

What could he access?

People’s private conversations across WhatsApp, Telegram, and Slack. Passwords and login credentials.

In some cases, he could run commands on people's computers as if he was sitting right in front of them.

There were reports of cryptocurrency wallets being drained.

Why?

Because Clawdbot was given all of the tools but none of the guardrails.

So what should we do?

Watch the stories, of course, because it's interesting and insightful.

I love watching the innovators so that I know what's coming.

But do not get tempted to play with these things unless you are an innovator and willing to manage the risks.

For most of us, it's not how we run our businesses.

Clawdbot is showing us where things are going to be maybe in a year or two. AI with real autonomy, that thinks ahead, solves problems, and takes action that you didn’t directly ask for.

What's working today in real firms

Smart businesses focus on what’s mature in business apps today.

Microsoft Copilot, Google Gemini, and other enterprise grade AI tools have a level of agency.

We're not going to be able to do all the amazing things people do with Clawdbot today, not right now.

But we’re also not risking our client data and our internal systems.

So for me the choice is clear. Watch the hype for entertainment and to see what’s coming. Focus on what’s proven and secure for actual implementation.

And there's plenty we can do right now

In the Microsoft environment, which is often the last to introduce new features because of their security requirements, we can already create agents that specialise in specific tasks.

We can connect tools in controlled ways.

We can manage agents’ access to data and control who can use them.

When I run AI Agent workshops in-house, groups end up with anywhere between 30 and 80 ideas for agents that they start implementing right away.

I'm seeing accounting firms and law firms today that have dozens of agents running around doing work for them: reviewing tax returns, preparing meeting briefs, having a first pass on emails.

They don’t work as autonomously as rogue Clawdbot agents.

They are doing work that matters every day in their workflow, in enterprise grade, secure, and manageable ways.

Want to get started?

I'm running free webinars on what you need to know in AI in 2026, and a specific AI Agents workshop where we'll create agents together, brainstorm, and share ideas. Everybody goes home with dozens of ideas and something already done.

Go to inbal.com.au/events and I'll see you there.

—

Inbal Rodnay

Guiding Firms in Adopting AI and Automation

Keynote speaker | AI Workshops | Executive briefings | Consulting CIO

Want to receive these updates straight to your inbox? Click here: www.inbal.com.au/join

When you are ready, here is how Inbal can help:

Transform your firm in 30 Days with the 30days to AI Program — Bring your entire team on the AI journey in just 30 days. This program is designed to give your team a solid foundation in using generative AI in responsible and impactful ways. Inbal helps you choose your AI tools, create an AI policy and train your team.