Twenty-two years, five languages, one notary stamp: how a senior Zurich practitioner stopped resisting AI

In our In Practice series we sit down with the lawyers who actually use Lexity day-to-day. Not to demo features, but to ask what changed in their work, what they trust, and what's still missing.

Episode four: Giuseppe Mongiovi, attorney and notary public in a small Zurich law firm, twenty-two years in practice, advising clients in German, English, Italian, French, and occasionally Spanish. Notary, multilingual, careful: the profile of the senior practitioner most legal-AI pitches bounce off. We wanted to ask him what changed.

Executive summary

• Giuseppe is a senior practitioner with one of the highest professional-secrecy obligations in legal work. His first reaction to AI was "it will never replace a lawyer", and the second was that the publicly available tools were not safe to upload client data into.
• The case-assessment work that used to take "between four and five hours" now takes "15 to 20 minutes" with the Client & Case Data Analysis Clickflow™. That is the entire premise of his change in posture, in his own words.
• The unlock wasn't speed. The unlock was a Swiss-hosted, ISO-certified, attorney-vetted environment that let him stop spending the time saved on anonymising documents before each query.
• In a multilingual practice, the in-platform Translation sits inside the same safe boundary as everything else, "I get a 40-or-50-page Swiss court decision in German, I send it to the client in their language, accurately, the same day, without leaving the safe area."
• His honest critique: outputs in nuanced legal matters still need careful human verification. He is comfortable saying that out loud, and so are we.

Meet Giuseppe

Giuseppe is an attorney and notary public in a small Zurich law firm, twenty-two years in practice. The work splits across corporate, contract, inheritance, and real-estate law. The client base is multilingual: German, English, Italian, French, sometimes Spanish. So are the documents that arrive on his desk under deadline.

Two things matter for understanding why we wanted him on camera. First, the notary stamp: a notary public's confidentiality obligations are among the strictest of any legal role. Second, the model: a small firm without an army of associates, operating across languages, often under time pressure. "Sometimes clients send documents in different languages. It's not that easy on the time pressure to get a summarised picture of the situation. With AI it gets easier."

Giuseppe is the kind of lawyer who, by everyone's expectation, shouldn't be the early adopter. He is.

First contact: the careful sceptic

Asked for his honest first reaction to AI, he didn't hedge:

What changed wasn't a demo. It was a different question (how many years am I going to be working in this profession?) and the calm decision to learn the new world properly. "I attended workshops here in Zurich, online and in person. I watched videos. I spoke with colleagues. I quickly realised the better the input, the better the output."

The trigger for actually using it was, in his words, exactly what we hear from practitioners under deadline pressure:

He addressed the billable-hour question head-on, without prompting:

That is one of the more honest answers we have heard on camera about how the economics of senior-practice work might shift. We didn't push it further. He was clear.

Before Lexity: publicly available AI wasn't enough

Giuseppe tried the obvious tools first.

The deeper problem, however, wasn't precision. It was data. He couldn't put the kind of documents he handles into an open tool without anonymising them first, and anonymising them first erased the time saving. That is the universal trap senior lawyers fall into with general-purpose models, and it is the moment most stop using them entirely.

A new case on the desk: four hours to twenty minutes

We asked Giuseppe to walk us through one specific task. He picked the one most senior lawyers reach for first when a new matter arrives: getting a clean overview of the facts, the timeline, and what's missing.

The Clickflow™ he uses for this is Client & Case Data Analysis, the new-case orientation flow. He prompted it to produce a timeline and a key-facts overview, in German because the documents were German. The output came back structured, in his language, with citations to the relevant documents, and then he asked it the question senior lawyers always ask: what's missing?

That is, in our view, one of the most underrated uses of legal AI. Not summarising what the file contains. Identifying what the file doesn't contain, and converting that into a client instruction. It's the move from a passive overview to an actionable intake.

For the same case, when he wants to dig in, he opens the chat at the bottom of the Clickflow™ output and asks specific questions about specific documents. "This is something I usually start with."

His total time on a recent matter:

The multilingual unlock: translation inside the safe area

A part of Giuseppe's practice that is invisible from the outside: he routinely translates between German, Italian, French, English, Spanish for clients across Europe. The Swiss courts hand down forty-to-fifty-page decisions in German. His clients don't all read German.

Until the in-platform translation existed, this was either expensive (outsourced sworn translation) or risky (publicly available AI on confidential court material). It is now neither.

What changed about the work itself

We asked Giuseppe the standard series question, not what feature changed, what about the work changed. His answer was specific to the senior, small-firm role:

That second half of the quote is the part that we think most legal-AI marketing misses entirely. Senior lawyers in small firms have very few substitutes, particularly in multilingual or specialised practices. The bench, in any meaningful sense, doesn't exist. AI doesn't replace bench depth. It substitutes for the kind of bench depth a small firm couldn't have built anyway.

On trust: the part that made it possible

We asked Giuseppe directly: what made him willing to upload? His answer is the section we'd most like every legal-AI buyer to read:

He wasn't being theoretical. The whole "15 to 20 minutes" claim above is conditional on this paragraph. If the platform required him to anonymise client material before upload, the way public AI does, the time saving disappears entirely. The trust posture isn't a sales bullet at the bottom of the page. It is the precondition that lets the speed gain be real.

He came back to the same point when asked what he would tell a colleague:

That trio, safe environment, attorney-vetted Clickflows™, navigable surface, is the answer we keep getting unprompted from senior practitioners. It's what makes the platform usable for the lawyers who, by training and obligation, are the most cautious adopters in the market.

Honest critique

We asked what frustrates him. His answer was characteristic of his profile: measured, but real.

The honest answer: the human review will not go away in a high-stakes profession, and we don't think it should. What will change is the time spent on it. More automation of more tasks, more precision in which tasks are automated, more integrations to gather context from where the work actually lives. The reviewer remains. The review shrinks.

What changed about the work itself, the second-order version

We closed by asking Giuseppe what AI does for him that nothing else has. He went somewhere we didn't expect: back to the language work.

That's the version of "AI for senior lawyers" we believe in. Not autopilot. Not replacement. A faster, cleaner path to the part of the work where a senior lawyer's judgment, and twenty-two years of language fluency, is the actual asset.