I demo’d our AI team member to 20+ innovation leaders...

Hi, and happy Tuesday.

We've been building an AI agent powered by OpenClaw at PreScouter over the last few months.

The idea is simple: instead of AI that sits in a browser tab waiting for you to ask it something, what if we had AI that just... worked? On its own. Like a team member would.

It’s not yet a finished product - but I demoed it live last Friday to a few of you in this newsletter community. It was an informal MS Teams call to share where we think things are headed, to get initial impressions.

You can find the recording here:

→ AI as a Team Member (27 min video)

Here are 5 things that seemed to resonate with the room:

1. It lives where you already work.

Most AI tools ask you to go somewhere: Open a new tab. Find the login. Switch contexts. Our AI shows up in your existing communication layer - Google Chat for us, MS Teams for most of our clients - as just another contact. You message it like a colleague. It messages you back.

While this sounds small, it’s a huge psychological shift from an AI tool to an AI as a colleague.

2. It was already working while we slept.

Before the demo even started, the AI had been monitoring packaging industry news overnight. It had categorised findings, flagged the most relevant developments, and pinged me that morning with an update: unprompted.

Even through the demo, it continued to update us on its ongoing work.

We've all used AI reactively - where you ask questions and it answers. This is AI acting on your behalf, with your interests in mind, before you know you needed it.

3. It collapses the tool stack.

During the demo, I asked the AI to find the right contact at a packaging company, search LinkedIn, draft an outreach email, and update the relevant CRM record. 

In our current world, that's four different tools, four different logins, and probably 20 minutes. Our AI rendered a single screen showing all of it - drafts of the work it would do - so I can review and make adjustments, before it updates the relevant tools.

You may have heard of the “SaaS apocalypse” - blunting the value of companies such as Salesforce and Monday.com. This was that in action: One agent can orchestrate across your existing tool stack, collapsing the value of individual SaaS products.

4. It's building its own knowledge - not just searching.

This one is subtle but important: When you ask it something, it's not doing a Google search. It's drawing on a growing knowledge base it's been building since it was first activated - entity relationships, context, connections between things it's observed over time.

We asked it to generate a competitive positioning map. It used everything it had learned since it started running 1.5 weeks earlier - using its accumulated understanding.

5. OpenClaw (and soon Copilot) only provides the blank canvas

Microsoft will soon give you an agentic Copilot. But an agent platform without skills is like hiring someone with a great résumé and no training - they can talk, they can search, they can summarize, but they don't know how your business works.

The value isn't the platform. It's what's loaded onto it: the extraction pipelines that know which industry developments matter to you, the monitoring skills that track your competitive landscape while you sleep, the operational knowledge that turns "we should follow up on that" into an actual follow-up. Platforms are the blank canvas. Skills are the painting. That’s where we are focusing.

The brain in a jar - the point I forgot to make

During the demo, I flashed the picture above - of a person at a desk, next to a brain in a jar (labelled AI) that is operating a computer. For me, this captures the essence of this new generation of AI: a technology that can do anything your or I can do, with access to various data and tools provisioned in the same way we provide those to staff.

But I forgot to drive home the key point:

• Humans live in the real world. We have eyes, experiences, instincts, relationships. 

• The AI doesn't. It’s like an isolated “intelligence living inside a jar,” operating “blindly”.

The implication? This “brain in a jar” scales the thinking of the human sitting next to it.

I don't think this means the end of jobs. I think it means the beginning of something different: the one-person “team”.

An expert in a function - competitive intelligence, sales outreach, market research - can now scale their work, with an AI team member handling the volume.

The human still drives the objectives. This is not a replacement -  it’s leverage.

The practical takeaway

If you're a leader at a large in any organization, after watching this demo, here's what you’ll likely takeaway from the session:

1. Agentic AI is real and it works today - perhaps not perfectly - but well enough to matter.

2. The organisations building the domain-specific layer now will have a head start when the enterprise-grade version arrives.

3. The one-person team isn't a threat. It's an opportunity.

Watch the full demo:

→ AI as a Team Member (27 min video)

Right now, we’re working on implementing this in every part of our organization - with a view to providing this as a service as demand emerges. 

I’m planning to shift this newsletter and related content to document how we’re doing it - the wins, the mistakes and the unexpected moments. I’m also planning to highlight similar work from others, so we can all learn together.

If you have questions or thoughts, let me know, so we can factor those in.

Best,

Dino