Annecy’s bar association embraces generative AI and paperless law—while insisting lawyers stay in charge

In Annecy, a lakeside city in France’s Haute-Savoie region, lawyers are rapidly folding digitization and generative artificial intelligence into everyday practice—from document management to legal research—without moving toward a model where attorneys disappear.

As 2026 brings faster, more capable AI tools, the Annecy Bar (the local professional association for attorneys) is trying to capture productivity gains while tightening controls around confidentiality, the quality of legal filings, and professional liability—responsibilities that remain tied to a lawyer’s advice and signature.

The Annecy Bar is organizing law firms’ shift to digital workflows

In Annecy-area firms, the rise of digital tools is no longer an experiment. Going paperless, building shared case files, and circulating drafts more smoothly are changing how lawyers work, including in small practices. Routine tasks—filing, indexing, and hunting for exhibits—are steadily moving into digital systems that are more traceable and faster.

Adoption of generative AI is often incremental. Firms start with lower-stakes uses—rewriting, summarizing long documents, formatting—before moving toward more sensitive work, such as helping orient litigation strategy or preparing a draft legal instrument. Lawyers quoted in local coverage describe a pragmatic approach: test on simple matters, measure the time saved, then set guardrails. The logic is straightforward: the law doesn’t tolerate approximation, and mistakes can directly trigger professional liability.

In Annecy, as in other French bar associations, the challenge is also organizational. Rolling out new tools requires internal rules: who can access what, which data can be processed, how records are archived, and how versions are controlled. Lawyers emphasize that traceability is becoming a quality benchmark—being able to explain how information was produced, checked, and then integrated into legal reasoning. In that sense, digitization isn’t just modernization; it reshapes how information moves inside a firm.

Economic pressure is part of the story, too. Clients want shorter timelines, simpler exchanges, and remote meetings when a case allows. Digital tools can meet that demand for responsiveness without eliminating face-to-face time. In sensitive matters—divorce, criminal cases, and high-stakes litigation—lawyers stress that interviews and a nuanced understanding of a client’s personal situation remain central to the job, something tools can’t replicate.

For many firms, the stated goal is to save time on mechanics and reinvest it in analysis and strategy—areas where attorneys continue to act as decision-makers, educators, and protectors of their clients’ interests.

AI is mostly used for research and summarizing—under human supervision

The most common AI uses in law offices involve finding information and summarizing it. In a case file running hundreds of pages, a tool can propose a timeline, pull out points of friction, or suggest questions to investigate. Lawyers describe it as help with direction, not an automated decision-maker. The line is clear: the tool can speed things up, but it doesn’t decide. Responsibility for the reasoning—and then the filing—remains with the attorney.

That matches a theme in French legal-trade coverage: AI can assist without replacing. Lawyers describe a practical benefit—freeing time from initial research to focus on higher-value work such as legal characterization, strategy, and negotiation. Many compare it to earlier shifts—databases, search engines, and practice-management software—that changed daily work without eliminating the need for a responsible professional counterpart.

The sensitive point is reliability. AI systems can generate errors, approximations, or even nonexistent citations depending on how they’re used. Firms are building verification habits: checking sources, cross-referencing, and demanding citations. In some situations, the tool is limited to draft functions—a working text that will be revised, corrected, and then validated. For lawyers, time saved only matters if it doesn’t degrade the quality of legal work.

A second sensitive point is data. Feeding a tool a confidential document immediately raises questions about attorney-client secrecy, hosting conditions, and reuse of information. Firms favor solutions that guarantee a secure perimeter, or they impose rules—anonymization, removing identifying details, and banning the insertion of certain documents. Confidentiality becomes a deciding factor, alongside performance.

In this phase, AI resembles a technical assistant. It can suggest, rewrite, and organize, but the final decision depends on a professional who can apply context. A case isn’t just a stack of text; it’s a story, a relationship of trust, and a human and financial stake. Lawyers interviewed emphasize that law is practiced in the real world—with imperfect evidence, emotions, risk, and sometimes heavy consequences.

Professional ethics set hard limits around secrecy and liability

Integrating AI runs into a non-negotiable principle: professional secrecy. The moment a lawyer entrusts sensitive information to an external system, the issue stops being purely technical and becomes an ethics question. Communications between attorneys, medical records, financial details—everything that makes up a case file can’t be processed without strong guarantees and real control over data flows.

Professional bodies cited in specialized coverage have stressed that any use must remain compatible with core principles: independence, loyalty, confidentiality, and competence. In practice, that means internal rules about what information can be entered, which platforms are allowed, and what retention policy applies. The aim is to avoid a “black box” effect where no one knows where a case file’s content goes—or who can access it.

Liability is the other guardrail. Even if a tool proposes a line of reasoning, professional civil liability and duties of diligence remain attached to the lawyer. That reality slows the most ambitious uses. A firm may accept help drafting an outline for court submissions, but it can’t delegate legal and factual analysis to a system that will never bear the consequences of an error. In litigation, even a small approximation can weaken a strategy or expose a party to additional motions.

Still, the shift is moving forward. Some lawyers note that not using AI can mean falling behind, especially when clients compare turnaround times and service levels. The balance many firms are seeking is to use AI as an accelerator without giving up human control. That also means documenting practices, training teams, and maintaining consistent standards—steps that come with costs and tradeoffs.

In Annecy, the accounts converge on one idea: a lawyer’s job isn’t simply producing documents. Lawyers listen, advise, warn about risk, negotiate, argue in court, and protect. Those functions require human judgment, a relationship, and an ethical framework. AI can optimize part of the work, but it doesn’t turn a tool into the holder of a client mandate—or the bearer of professional responsibility.

Client relationships and courtroom hearings remain hard to automate

Digital change shows up in daily routines: scheduling, sharing documents, secure messaging, and video meetings when a case allows. But some moments resist automation. The client relationship remains central, especially when a situation is emotionally charged or the financial stakes are high. In those cases, clients aren’t looking for a standard answer—they want someone who can explain, reassure, prioritize, and take a position.

Court hearings are another area that resists automation. Lawyers say advocacy and case management depend on reading the room, responding to opposing arguments, and adapting to a judge’s questions. Even with powerful tools, surprises happen: a new element, a contradiction, a disputed exhibit. The work remains real-time, where experience, strategy, and composure matter.

AI can help with preparation. It can summarize a file, propose scenarios, or generate checklists. But preparation isn’t execution. Lawyers point out that decisions are built with non-textual elements: a witness’s credibility, the coherence of a narrative, the dynamics of negotiation. Those factors don’t reduce neatly to structured data. The law firm remains an interpreter between legal rules and lived reality.

In work with companies, another factor emerges: compliance demands. A client may require internal procedures, audits, and confidentiality guarantees about the tools being used. Firms then have to prove they control their processing chain. That pushes the market toward more tightly controlled hosting and stricter configurations. Modernizing legal service also means being able to account for how work is done.

Digitization is also changing competition. A responsive firm that can quickly deliver a structured memo may win a matter. But the advantage isn’t just the tool—it’s the method, quality control, and the ability to explain risk and propose strategy. In the reporting cited, AI is described as leverage, not a substitute, with the core mission remaining the defense of concrete interests carried by real people.