Here is what shows up in almost every conversation DonnaPro has with founders about executive support.

They use AI. They like AI. They are not going back to doing everything manually. But somewhere in the last two years, they realized AI handles the volume of work and the human handles everything else, and that everything else is the most important part.

Security comes up constantly in founders' conversations as they start to recognize what AI cannot protect. An email that disappears into a filter and never resurfaces. A scheduling conflict that AI resolves without asking which meeting matters more.

The founders closest to AI are the clearest about what it cannot be accountable for. And that is changing how they hire.

DonnaPro is a European premium executive assistant agency built exclusively for CEOs and founders. It was founded by Filip Pesek on an observation he repeats often: if you don't have an assistant, you are an assistant. The company claims that every assistant passes a six-stage selection process, trains for two weeks on AI tools before any client work, and operates from more than 2,500 pages of operational frameworks. This can help clients reclaim more than 60 hours a month.

The balance that works for founders

McKinsey estimates that generative AI and related technologies could automate work activities that absorb 60 to 70 percent of employees' time today. An email that took ten minutes now takes one. A meeting summary writes itself, and so much more is efficient with the help of AI. The balance that works for founders is AI handling volume and speeding up the process, while human judgment handles the quality.

The work that can be automated has largely been automated. What is left is the work that requires context, accountability, and the ability to decide. That is the work many founders want delegated now. It is the judgment and follow-through that matters for founders and C-suite staff, not so much the inbox management and calendar scheduling.

Why security comes before price

Security matters more than price for many founders. Not because they distrust AI, but because they understand what AI cannot protect.

Founders who have given AI access to their inbox describe a specific anxiety: something important disappears and they never find out. The concerns around security may outweigh the possibility of AI doing a job badly.

With DonnaPro, the security model is different. Assistants work through delegated access, not password sharing. In Google Workspace and Microsoft 365, a delegate can read, send, and delete email on a founder's behalf, but cannot change the password or access account settings.

Assistants operate on the founder's own domain email, and access can be withdrawn at any time.

The company says that every assistant is based in the EU, which means GDPR compliance may already be built into the contract. These agreements cover GDPR, intellectual property, and confidentiality. Some clients also ask their assistant to sign their own NDA, and the assistant can do that personally.

The founders closest to AI

One pattern worth noting. The founders who often understand AI best are the ones most likely to hire human support anyway.

One founder whose own company builds AI search tooling still hired a human assistant. Another runs on automate, delegate, eliminate and automate first. He may still be hired.

What they share is not doubt about AI. All of them use it daily and expect it to keep improving. What they question is whether a tool can be accountable for an outcome. When something important slips, the tool does not notice. The founder does. That is the gap many hire for.

AI can handle what it can measure. The work that matters most is often the work that cannot be measured. The judgment call. The relationship. The follow-through that no algorithm tracks.

What founders are actually telling DonnaPro

DonnaPro has clients in more than 20 countries, including the EU, UK, and US, and in their conversations with founders considering executive support, a few things come up repeatedly.

AI came up on many calls, usually raised by the founder. Most of them already used it. One portfolio executive had found DonnaPro by asking an AI assistant for a recommendation. The same executive described where it falls short. Joining a call late from a hotel, he said: "Stuff is organized currently by AI, and then it is back-to-back and not enough space in between. The logistics are not considered."

AI can draft, summarise and schedule. It cannot notice that a back-to-back day in two cities does not work, or decide which of two clashing meetings matters more to the founder this week.

"AI handles the volume. Human judgment handles the quality," says Filip Pesek, founder and CEO of DonnaPro. DonnaPro assistants are trained on AI tools before they start work with a client, so founders get both.

The structure that makes both work

The question for founders is not whether to go with AI or humans. It is how to structure both so each does what it does best. AI handles the volume of the work and a person owns the outcome.

"AI can draft almost anything," as the team at DonnaPro puts it. "It cannot own the outcome. A draft still needs someone to decide whether it should be sent, to whom, and when. The assistant is not replaced by AI. They run it, and own the outcome the tools cannot."

Filip hired his first assistant at 19 and spent the next eight years making nearly every hiring, training, and delegation mistake possible before building a system that actually worked. He combined that experience with what he learned running his own marketing agency, Spark, and launched DonnaPro in 2024. Assistants are trained to take on more than 50 different types of tasks, from inbox management and calendar coordination to research, project follow-ups, and stakeholder communication. That breadth means founders are not necessarily hiring a generalist. Instead, they are likely building a support structure that fits how they actually work.

At DonnaPro, this is built into how assistants start. Two weeks of training before client work, including the AI tools assistants use daily for notes, drafting and research. More than 2,500 pages of operational frameworks, so each assistant works from documented methods rather than improvising. Written processes for every client, built as the assistant takes work over, so the founder's way of working is recorded rather than held in one person's head.

AI raised the floor. The job moved up. What founders often want now is accountability, judgment, and someone who notices what the algorithm misses.

The founders who are already there made one decision early. They stopped thinking about AI and humans as competing options. They started thinking about how to use both.


VentureBeat newsroom and editorial staff were not involved in the creation of this content.