Integrating AI in the company with legal certainty: the questions that order an in-house lawyer's work

Noticias Juridicas (Aranzadi) has published a free dossier on integrating artificial intelligence into the company with legal certainty, aimed squarely at in-house counsel. The value of the document is that it turns an abstract debate into an ordered work plan: what to inventory, which obligations to calendar and how to document decisions so they survive judicial review.

Two expert voices, one transversal model

The dossier features two contributors. Noemi Brito Izquierdo, head of intellectual property and new technologies at KPMG Abogados, analyses how to build a transversal governance model, assign responsibilities and convert the obligations of the AI Act and the Digital Omnibus into applicable controls and decisions. Eduardo Ortega, partner in the labour department of Ortega Condomines Abogados, addresses AI in employment relationships: selection, performance evaluation, algorithmic transparency, human oversight and the protection of workers' rights.

The questions that order the legal work

The document structures its analysis around six questions that work as a checklist for any company legal service: why integrating AI requires a transversal governance model; how to identify and classify the AI systems used in the organisation; which obligations and deadlines the in-house lawyer must know to integrate AI in line with regulation; how to organise decision-making and the assignment of responsibilities; what risks AI poses in employment relationships; and what the legal department must do to guarantee responsible use from the labour perspective.

System classification is probably the step that generates the most bottlenecks in practice. The European AI Act organises obligations according to each system's risk level, and personnel management systems (selection, promotion, evaluation) are among the high-risk ones, with requirements of human oversight, technical documentation and data management that admit no improvisation. Without a prior inventory of the tools the company actually uses, including those embedding AI without advertising it, any compliance plan limps from the start.

From theory to the case file

Current case law adds pressure to this preventive work. This very week we covered on our front page that the Madrid High Court has opened Spanish doctrine on the collective redundancy with AI involved: judgment 864/2026 requires the cause of the collective dismissal to be documented independently of the tool, and leaves the algorithm as evidence of how the decision was made, not as the ground. That is exactly the kind of reading a well-built governance model anticipates: if the company documents its automated decisions from the beginning, the file withstands judicial review.

On the sanctioning side, it is also worth recalling what we published about fake AI-generated case law and court sanctions: the professional use of these tools demands verifiable human controls, inside and outside the company.

Why it matters to law firms

For firms advising SMEs without their own legal department, the dossier points to a concrete service line: acting as an external governance function, with a system inventory, an obligations map and documented decision protocols. For the in-house lawyer, the message is that AI Act compliance is not a one-off project but a permanent decision structure mixing technology, labour law and data protection.

We will keep following the practical application of the AI Act in our legal AI section and its effects on law firm organisation in processes.

Source: Noticias Juridicas. This piece is informative and does not constitute legal advice.