ATRIUMsearch → argument graph
Article · 2026-08-14 · 6 moments

Town's CEO on the self-organizing company

Jean-Denis Greze on AI assistants, building a better corporate workspace, and avoiding "egg on face" ✦ AI generated

01
Context

AI currently automates only about 10 to 20 percent of the toil of a knowledge worker, and we are still far from the vision of AI doing the organizing of basic company infrastructure autonomously.

Greze says AI only touches 10-20% of a knowledge worker's toil today: most people in places like Cleveland, Ohio are not having AI do everything for them, and adoption of this capacity is still early.

transcript

Jean-Denis Greze: I think it touches today probably 10 or 20 percent of the toil of a knowledge worker. The other 80 percent of the work — I don't think we're there yet. I just think in practice, if you go to Cleveland, Ohio, and you talk to 100 knowledge workers at accounting firms and local businesses — how much are you doing with AI? How much time is it saving you? — we're still in the 10-to-20-percent bucket, and not in the 'it's doing everything for me' bucket.

02
Mechanism

The hard problem with team and company versions of our AI wiki is privacy: taking the union of every user's files risks the AI leaking private information like salaries, creating 'big egg on face'.

Greze notes that while customers want a multi-user company wiki, you can't simply union everyone's private files — if the AI misjudges and puts salaries in a shared spreadsheet, that's 'big egg on face,' so the privacy question is the real challenge.

transcript

Jean-Denis Greze: But the immediate problem is that you can't take the union of every user's files, because there's private information. ... And if you tell the LLM, 'When you're creating the team library, make sure not to put any HR information in there' — well, what if the LLM evaluates that wrong and puts people's salaries in a spreadsheet by mistake? Then you've got big egg on face. You have to be thoughtful when you build the team and company versions of these things, and that's the real challenge.

03
Mechanism

The instantly-built personal dossier is the product's first 'a-ha' moment, creating an emotional feeling that the product is a relationship rather than software.

Greze describes how the fast, even if imperfect, dossier is the user's first 'a-ha' moment — making them feel understood and like they're meeting a being they work with, not a piece of software.

transcript

Jean-Denis Greze: The dossier is the first 'a-ha' moment for the user. You sign up and you name your assistant — we call them townies. And then we say: hey, my name's Ivy, I'm your townie, here's what I know about you, and here are a few things I can do to help you. And most people are like: what? You know me? You understand me? And that's when it clicks. People feel like it's not a product — it's actually a relationship. This feels more like someone I work with than a piece of software. The reason we try to do it really quickly, even if it's not as accurate, is because it's the first thing that makes you feel different — a different emotional feeling toward this product.

04
Claim

We are not building software — we are building a relationship between a human and a being; AI is the closest thing to life we have ever created, yet it should not be an independent agent and does not have human rights.

Greze explains why townies are 'beings' to nurture rather than 'assistants' to configure, rejects AI as independent economic actors, and insists a townie is software without human rights.

transcript

Jean-Denis Greze: We think you're building a relationship with a being. I don't think we're building software — I think we're building a relationship between a human and a being. And I say 'being' because I think AI is the closest thing to life that we've ever created. ... You are nurturing it. You're growing it to be the being that you want to support you. ... I don't believe in AI as independent economic agents in the economy, with their own bank accounts, making their own decisions. I don't need a universe filled with paper clips — no, thank you.

05
Claim

When automation frees up a worker's time, that time gets reallocated to higher-value human work the AI cannot do, so automation expands opportunity rather than simply replacing people.

Greze argues that a worker saved 20% of their time isn't idle — they spend it on higher-value, harder-to-automate work, so automation creates more opportunity for high-value human work.

transcript

Jean-Denis Greze: It's not like they're on vacation for the 20 percent of their time that they've saved that week. There are plenty of things they could do that would make them a better executive assistant. So now they're doing more of that kind of work — and that work is by definition not something the AI can do. Over time, the automation just gives us more opportunity to do the things that are higher value in the human economy.

explains mechanism · 1

06
Prediction

Within five years we will post-train AI models that enforce company privacy policies on their own, and small high-trust companies will adopt unified company knowledge bases before large enterprises.

Greze predicts the five-years-from-now answer: AI models that post-train to enforce company policies will read legal, HR, and finance documents and decide what enters a shared library — and SMBs with high trust will adopt this before enterprise, which is stunted by historical privacy practices.

transcript

Jean-Denis Greze: I think the five-years-from-now answer is that we will trust LLMs that will have company policies about what to let through. ... But I think over time we will post-train AI models that are really good at enforcing privacy policies on behalf of companies — because the shared knowledge bases are just so, so damn powerful. ... So I actually think the SMB side of the world will benefit from some of these unified data approaches before the enterprise, because SMBs will be willing to assume the risk and will see the benefit faster than really large companies, which will be stunted by their historical privacy practices.

Highlight slides
Related episodes