There’s a funny paradox: we live in a world where the AI in your meeting might “attend” more reliably than your human teammate - even if your teammate wanted to come. As The Washington Post put it: “Robots often outnumber humans in virtual meetings.” The Washington Post
But here’s the thing: the point of AI in meetings (or anywhere at work) shouldn’t be to replace people. It should be to uncover the things people miss, to amplify voice not silence it, to make hidden friction visible so humans can step in and fix it.
📌 Why the notepad replacement is a red herring
AI notetakers already do a credible job at transcription, summarizing, and surfacing action items. (Zapier’s recent list of 9 top AI meeting assistants is proof that this space is crowded - and improving.) Zapier
But meeting transcripts alone don’t tell you:
- Who didn’t speak and why.
- How teams actually operate: Who dominates discussion, who stays silent, how leadership dynamics evolve, and whether group structure helps or hinders outcomes.
- Whether follow-through was vague or meaningful.
- Whether the energy or engagement decreased mid-call.
These are the signals behind the transcript. That’s where real insight lies.
🏛️ Laws catch up: seeing AI, demanding accountability
Let’s peek at what's happening globally:
- Italy passed Law 132/2025, which mandates that AI used in the workplace (hiring, evaluation, monitoring) must be transparent, explainable, and overseen by humans. (nortonrosefulbright.com)
- The law also establishes a national observatory to monitor AI effects on work, promotes training for employers/employees, and prohibits interpretative AI in decision-making without oversight. (aoshearman.com)
- Meanwhile, emerging legal debates (e.g. about Otter.ai) show risk exposure: recording private conversation, using transcripts for model training, or lack of user notice may invite lawsuits. (workplaceprivacyreport.com)
These laws aren’t about banning AI. They’re about regulating it: ensuring people know when it’s used, can question decisions, and live in workplaces that respect dignity, not just efficiency.
🤝 What a people-first AI approach must do
If AI is going to replace the notepad, it must not replace the values:
- Human oversight must always exist. Algorithms assist, but people decide.
- Transparency, not mystique. Workers deserve to know: what data was used? What criteria? What was the threshold?
- Prepare your people. Rolling out AI without training or communication is asking for resistance, confusion, or worse.
- Focus on insight, not surveillance. The goal is to spot friction, not feed paranoia.
- Embed review and recourse. If someone asks “Why was I flagged low participation?” there must be a pathway for human review.
🔍 Why this matters - and where Ulla fits in
We don’t think AI should micromanage people’s time. But when meetings change how people work, we have to see that change.
At Ulla, we agree with these legal principles. That’s why our onboarding walks teams through what is measured, why, and how. We don’t pretend metrics are neutral - they are choices. We give you room to opt-in, to explain, to challenge.
So when AI replaces your notepad, it's not the end of human voice. It's the beginning of listening better - making the invisible visible, so humans can do what humans do best: decide, adapt, empathize.
If you ever want to test a meeting’s hidden signal (not just the text), we’re here.❤️ - Ulla Technology