AI and HR: what’s really changing in the workplace

Could you tell us about your background and explain what led you to take a particular interest in artificial intelligence?

I’ve had all sorts of jobs over the years! (laughs) I used to be a singer and then a driving instructor. Now I’m a lawyer, with a Master’s degree in law from the University of Lausanne. I practised for nearly six years at a notary’s office in Lausanne, mainly in property and commercial law: drafting deeds, legal opinions and handling cases, some of which were highly technical.

My interest in artificial intelligence stemmed from a concern: in the summer of 2024, I became firmly convinced that a significant proportion of my tasks would soon be capable of being carried out by AI. Yet I know that when I feel afraid, it is usually because I do not fully understand the subject. So I decided to learn more about it, precisely so that this unknown would cease to worry me.

I discovered a tool that really sparked my interest. Today, I design comprehensive websites, create bespoke business tools for SMEs for whom such tools were previously financially out of reach, and have developed several smartphone apps, one of which, Travel With Me – Family Trips, is available on the official app stores. Above all, I founded All-in Solution, which supports legal professionals in auditing their use of AI, training their teams and deploying locally hosted solutions, ensuring that confidential data never leaves the practice.

As a lawyer, how is AI actually integrated into your work today?

I distinguish between two uses, depending on the nature of the information I am processing.

For my research into legal doctrine and case law, as well as for keeping up to date with developments in my areas of interest, I use an online AI tool, which I have extensively customised. It saves me a considerable amount of time when identifying relevant sources, tracking developments on a particular issue or structuring an initial analysis, which I then systematically verify.

As soon as any private or confidential data appears, I switch to a second tool, which is installed locally and custom-configured. It runs on my own server, so no information leaves my working environment. This is essential for reconciling the use of AI with professional confidentiality and data protection.

It is this dual approach that I recommend to legal professionals: an online tool for everything relating to public information, and a local tool for everything relating to clients, all perfectly tailored to each practice or firm. This structure enables optimal implementation and, above all, a sensible and judicious use of AI in day-to-day practice.

In your view, what are the legal tasks where artificial intelligence can add the most value?

AI excels in any situation where a lawyer needs to absorb, sort through or rephrase a large volume of information. Research and monitoring are prime examples of this: it identifies relevant legal doctrine, case law and legislative developments in a matter of minutes, where it used to take several hours, or even, in some complex cases, several days.

She also excels at document analysis. Summarising a voluminous file, comparing two versions of a contract, checking the consistency of names, amounts and dates in a document running to several dozen pages: these are all painstaking checks that she carries out with great consistency. Furthermore, she works equally well in several languages, a valuable asset in a multilingual country such as Switzerland and in cases with an international dimension.

Drafting is another major benefit. Using my own templates, the AI produces first drafts of clauses, letters or legal opinions that adhere to my structure and style, allowing me to focus on the substance. It also helps to rephrase a legal analysis into plain language for the client.

In any case, the AI prepares the work and the lawyer approves it. It is this division of labour that makes it an asset rather than a risk.

What skills does a legal professional need to develop today in order to work effectively with AI tools?

Even before we talk about effectiveness, one question takes precedence over all others: where does the transmitted data go? Legal professionals must always know, before any use is made of the data, where the information is processed and stored, which intermediaries it passes through, whether it may be retained or used to train a model, and which laws it is therefore subject to. The example of Mistral is telling: this French company, often presented as the European alternative, hosts data within the European Union by default, but uses subcontractors – including major US infrastructure providers – and provides for data to be temporarily transferred outside the EU. However, a US service provider may be compelled by its own national law to disclose data, even if it is stored in Europe. For a profession bound by professional secrecy and subject to data protection regulations, this is not a technical detail: it is a matter of responsibility and credibility.

Next comes an understanding of what an AI tool actually is: it does not seek the truth, but produces the most probable answer. It may therefore be wrong with great confidence, cite a non-existent judgement or misinterpret a provision. Legal practitioners must be aware of these limitations in order to exercise systematic critical scrutiny.

Finally, one must master the art of phrasing a request: providing context, an objective, a framework and the expected format, just as one would brief a trainee. Lawyers, incidentally, have a natural advantage in this regard, as precision of language lies at the heart of their profession.

None of these skills can be acquired by winging it. Tools evolve very quickly, and only comprehensive training, which is regularly updated, enables them to be used effectively and safely. This is the very essence of the ‘augmented lawyer’ approach.

What role can a legal professional play in supporting businesses that wish to integrate artificial intelligence whilst complying with the legal framework?

It plays a key role, provided it is implemented at the right time – that is, before deployment and not after an incident.

In most organisations, AI is already in use, often without any framework: staff copy contracts, customer data or internal information into consumer-grade tools, without knowing where this data ends up. The legal professional’s first task is therefore to carry out a review of the current situation, identify these practices and assess the risks they pose to the company.

It must then translate the legal framework into practical rules. In Switzerland, the Data Protection Act already applies in full to AI, as the Federal Data Protection and Information Commissioner has pointed out. Companies that do business with the European Union must also take account of the EU’s AI Regulation. The role of the legal professional is to translate these requirements into operational decisions: which tools to authorise, for which data, under what contractual terms, and with what internal rules of use.

Finally, it provides ongoing support to teams through training and by keeping them up to date with the rapidly changing regulatory framework.

The lawyer thus acts as a bridge between technology and the law. This is precisely the approach I propose with All-in Solution: audit, secure, train, then roll out.

What advice would you give to a student or a young lawyer who wishes to specialise in the field of artificial intelligence?

First and foremost, do not neglect the fundamentals. AI does not replace legal reasoning: it enhances it. A lawyer with a solid grasp of the law will immediately spot an error in a generated response, whereas a less experienced lawyer might accept it without realising the mistake. The more powerful the tool, the more crucial the competence of the person using it becomes. It is therefore always essential to have a thorough understanding of the subject matter before seeking to save time by using AI.

Next, make data protection a genuine specialism. It is now the point where law and AI intersect, and this is where businesses most need lawyers who can understand both the rules and how the tools work.

Above all, don’t just read about AI: you need to put it into practice. Test the tools, compare their results, and understand their limitations through experience. You can only give good advice on something you know from the inside.

Finally, I would say: don’t be afraid. I myself started out feeling anxious, and it was by committing myself to serious study that this fear turned into a passion. Dual expertise in law and AI is still rare. Those who commit to it today in a systematic way, through comprehensive training, will have a considerable advantage tomorrow.   

In your view, which of a lawyer’s tasks will remain difficult for artificial intelligence to replace?

All those based on human relationships, judgement and responsibility.

First and foremost, a relationship of trust. A client does not simply come looking for a legal answer: they come with a story, concerns and, at times, unspoken issues. An inheritance, a separation or the purchase of a first home involves far more than just the law. Knowing how to listen, picking up on what is left unsaid and tailoring your advice to the person sitting opposite you remains a profoundly human quality.

Next, the strategic choice. AI can set out all the legally possible solutions, but choosing the most appropriate one involves weighing up interests, anticipating reactions and assessing risks that are not set out in any legal text. The same applies to negotiation and mediation – alternative methods of conflict resolution which I greatly prefer – where everything hinges on understanding the other party.

Finally, there is the question of responsibility. It is the legal professional, the solicitor or the notary who signs the documents, provides advice and takes responsibility for that advice. No tool can bear that responsibility in their stead, and it is precisely this that underpins the client’s trust.

And that is precisely the paradox of AI: by taking on the most time-consuming tasks, it frees up the lawyer’s time so that they can devote themselves to what constitutes the true value of their profession and their humanity. That is what it means to be an ‘augmented’ lawyer.

Conclusion

Artificial intelligence is opening up new prospects for the legal profession, whilst reinforcing the importance of expertise, accountability and human interaction. As Margot ALLIN points out, training and learning to master these tools are essential to keeping pace with this development and becoming an ‘augmented’ lawyer, capable of combining technological innovation with professional judgement.

OSAM extends its warmest thanks to Margot ALLIN for her account and for sharing her experience.