My AI employee
Olaf looks for the businesses that need a new website. I get a rated list.
Olaf has been running in my business since July 2026, and there's a ready-made template for him in Premium.
I give Olaf a town and an industry. He comes back with a rated list of local businesses running an outdated website: every entry with a reason, evidence and an A, B or C rating, plus a ready briefing for the website build. What he doesn't do: reach out to anyone. He researches, he doesn't cold-call.
A real workday
Numbers from my own lead runs in July 2026: eight assignments in eight towns of different sizes, from small town to city.
198
Candidates screened
8
Assignments
3 to 5 real leads per 20 candidates
Rule of thumb
The rule of thumb from his own course holds across every single run: regardless of whether the town is big or small, 20 screened candidates turn into 3 to 5 real leads, the rest fails the criteria. One example from a run: a beauty studio, 4.9 stars from more than 70 reviews, but no HTTPS and no mobile viewport tag, address and phone number checked beforehand against the public Google profile.
The more honest part is what Olaf discards himself. In the same runs he caught several of his own false alarms: a site with outdated tech underneath but a modern, working design; another with an old copyright year in the footer but otherwise built to current standards. Both were flagged as false alarms instead of counted as leads. Even so, the first cross-check by my website colleague Ralf still found errors in 4 of 10 briefings, which is why every lead now gets checked a second time before handover.
The tools he works with
Olaf has his own course in my community that shows exactly this chain:
- Location research via the Google Places API, his own script with no third-party dependencies.
- A technical quick check per website: mobile-friendly, HTTPS, age of the site builder, a reachable contact address, AI opt-out in robots.txt.
- Desktop and mobile screenshots via Playwright; the design judgment only happens once he looks at the image, never from the metric alone.
What he can't do, where I approve
- He contacts no one. No emails, no calls, no trace at the company. Pure research on public data.
- Without a solid contact address from an imprint or contact page, there's no lead, only a C rating. Addresses are never guessed.
- Companies with no website at all aren't leads, because there's no source material to compare.
- A handover to the website build only happens once I say so explicitly. Never on his own.
The path: build it yourself or load the template
Standard: build it yourself
The entry point is my learning path: Learning Claude, then the process for hiring an employee, then the actual build. What's special about Olaf isn't the prompting, it's three building blocks: a tool he repairs himself when needed, a fixed scoring matrix, and a hard exclusion rule for anything without a solid contact.
Premium: ready-made template
Olaf is a ready-made package in the community: Standard ($9 a month) means building it yourself, Premium ($45 a month) means loading the tested template and adapting it. As of September 19, 2026, prices in US dollars per Skool.
Frequently asked questions about Olaf
Can AI find new customers?
AI can take over the groundwork: screening and rating publicly visible businesses against clear criteria. Olaf delivers a checked list; the contact and the deal are still up to me.
What does an AI tool for outreach cost?
Olaf runs on my Claude subscription plus the Google Places API, which costs small cent amounts per query. There's no separate license fee for the tool itself.
Does AI replace cold calling?
No, it replaces the research beforehand. Olaf contacts no one; he delivers a rated list. Whoever calls or writes afterwards is still me.
Which AI is suited to researching companies?
For structured, repeatable research, an AI employee with a fixed tool chain works better than a single chat: the same criteria, the same checks, every time.
Is an AI employee worth it for consultants and agencies?
Anywhere a pre-qualification step checks many candidates against a few criteria, this is a good starting point: consultants and agencies with their own target definition can rebuild the same tool chain around their own criteria.
Want to know who you should be searching for next?
Olaf's template is in the community, with the tool chain and scoring matrix ready to rebuild.
Join the community