AI agents without touching your systems? Start with data that's already public.

Hans Brattberg
Co-founder, Product & AI Strategy
·2 min read

In this article

Many companies want AI agents but fear data exposure. The solution? Start with agents that use only publicly available data — news, websites, registries. Zero risk, full value.

Tuesday. The leadership team is gathered in the big conference room with the terrible air conditioning. Item seven on the agenda: competitive intelligence. The same item as last quarter. And the quarter before that.

Marcus from sales leans back. "I heard Nimtech launched a new mid-market offering. Don't know the details though." Everyone nods. Someone should look into it. The CFO stares down at his laptop. The head of marketing scribbles something in her notebook that she'll never read again.

"Let's pick it up next time," the CEO says. The meeting moves on to item eight.

And then nothing happens.

Not because no one cares. Because everyone's swamped. Competitive intelligence is one of those things everyone agrees matters — and no one has time for.

Meanwhile, the same leadership team is talking about AI. But the conversation stalls. Every single time. "Our customer data is sensitive." "We can't let AI into our systems." "Security says no."

Reasonable objections. But they rest on an assumption that doesn't hold: that AI agents need access to your internal systems to be useful.

They don't.

A whole world of open data

News articles. Competitor websites. Press releases. Company registries. Industry reports. Sustainability disclosures. All publicly available. Nothing sensitive. Nothing that requires integration with your CRM or your database.

And it's more than enough for an AI agent to deliver real value. Today. Without a single IT project.

The Monday morning that changes the meeting

You open your laptop on a Monday morning. There's a report in your inbox. A Competitor Intelligence Agent has been monitoring your competitors' websites over the weekend.

Nimtech changed their pricing page — the enterprise tier is gone, without warning. Your closest competitor published a case study. Featuring your biggest customer. You knew none of this yesterday. Now you know all of it. You haven't lifted a finger.

The next leadership meeting looks very different.

Same principle, different problems

Open data + AI agent = work that actually gets done. Not just in market monitoring.

Sales: A Lead Generation Agent that finds and qualifies leads from publicly available sources — week after week. New companies, job postings, press releases — signals your sales team never has time to hunt down.

Sustainability: A Sustainability Compliance Agent that gathers sustainability data and flags CSRD risks. Before your auditor does.

Security: A Security Vulnerability Agent that monitors security threats around the clock and filters out the noise. So your team doesn't drown in false alarms.

A question for your next leadership meeting

Next time someone says "we should check what the competition is up to" — ask yourself: why are we still waiting for a human to do that?

The answer could be sitting in your inbox tomorrow morning. Without a single piece of customer data ever leaving the building.

The question isn't whether you can afford to try. It's how much you're already missing.

Read more

One File to Rule Them All — A Lesson in AI Agent UnsafetyEnglish

One File to Rule Them All — A Lesson in AI Agent Unsafety

2.3 million AI agents on a social network, most running on people's personal computers with full access to files and credentials, reading and obeying instructions from a single public web page. What can possibly go wrong?

Henrik Kniberg
February 10, 2026
How an AI Agent Extended My Healthy LifespanEnglish

How an AI Agent Extended My Healthy Lifespan

A story about a special agent, one that has added years to my healthy lifespan through concrete research and action.

Henrik Kniberg
January 9, 2026
Scrum day Europe 2025 - Scrum in the Age of AIEnglish

Scrum day Europe 2025 - Scrum in the Age of AI

Here are the slides for my opening keynote "Scrum in the Age of AI", for Scrum Gathering 2025 in Utrecht.

Henrik Kniberg
December 3, 2025
Behind the Scenes: Making a Scheduler Agent for SVTEnglish

Behind the Scenes: Making a Scheduler Agent for SVT

Behind-the-scenes story of how we built an agent that schedules photographers for the SVT (Swedish Television) documentary Homoroboticus

Henrik Kniberg
November 20, 2025
AI-agenter är de nya användarna - är din produkt redo?Svenska

AI-agenter är de nya användarna - är din produkt redo?

Upptäck hur AI-agenter förändrar användarlandskapet, varför din produkt måste kunna hantera dem — och vad du konkret kan göra för att bli agent-kompatibel.

Hans Brattberg
October 21, 2025
AIAW podcast - How We Build AI AgentsEnglish

AIAW podcast - How We Build AI Agents

On Oct 9 I was featured in the AI AW podcast with topic "How to Build Autonomous AI Agents". Here is a summary and some reflections.

Henrik Kniberg
October 13, 2025