Most discussions about AI in the legal field focus on algorithms, data, metrics, sources, and regulatory compliance. But there is another equally critical dimension: the human experience.

In legal practice, technology does not exist merely to “process data” it exists so that professionals and clients can interact, understand, decide, and trust.

Designing truly human-centered legal AI means asking not only “what does the system do?” but also “how do people use it?”, “what do they need?”, “what barriers do they face?”, and “how do we ensure the tool is accessible, usable, and meaningful?”

This article briefly explores three key areas:

1. Usability and access for the professional.
2. Accessibility and client experience.
3. A “human-in-the-loop” design linking AI, professional, and client in a continuous process.

Finally, we examine how this vision is applied in a tool like GenIA-L.

Usability for legal professionals: AI as an ally, not an obstacle

For lawyers, advisors, and legal teams, AI must integrate seamlessly into the workflow, not disrupt it. A recent study examining AI assistants in legal settings found that usability and adaptation to legal context weigh more heavily than raw algorithmic accuracy.

What does usability mean in legal AI?

• Intuitive interface: professionals cannot afford to get lost among complex screens or forms.
• Integration with existing tools: firms already use document management systems, CRMs, and email AI must fit in smoothly.
• Clear, structured feedback: when AI suggests a document, clause, or argument, the lawyer should see why it was suggested, what room there is for review, and how it can be modified.
• Minimal training rapid learning curve: reports show that legal teams value “how quickly I can work with it” more than “how sophisticated the algorithm is.”

Human-centered design

Research on Human-Centered Design (HCD) and Human-Centered AI (HCAI) shows that legal tools must be defined and developed around real users their work, needs, and obstacles.

Professionals need direct access to the original legal text cited by the tool, not just the generated summary.

Application to GenIA-L

In the context of GenIA-L, this means:

• The interface must allow lawyers to easily view the legal source, jurisdiction, and date of the provision.
• A “human review status” panel should be included so that professionals can document their supervision.

Accessibility and client experience: AI must also serve the end user

It’s not only the professional who matters the client (whether a company, individual, or internal department) also interacts, at least indirectly, with AI-generated results. The design must therefore consider their experience.

Accessibility and inclusion

Legal technology design should adopt accessibility standards (such as WCAG) and value-sensitive design principles that prioritize human values.

For example:

• Clear, comprehensible language: ensure the client does not receive an “AI document” that is incomprehensible.
• Responsive design: the client may review it on a phone, tablet, or computer.
• Transparent information: indicate when AI was used, which parts were reviewed by a professional, and what the system’s limits are.

Client experience as a competitive advantage

When clients see that the tool used by their advisor is both intelligent and human that it was designed for them trust grows. A study found that even when system accuracy was similar, the tool that “feels made for the legal user” is the one that prevails.

How this translates to GenIA-L

• Provide clients with an executive summary of the AI-assisted work, written in plain language, explaining what the AI did, what the advisor did, and what the next step is.
• Include a “how to interpret this document” section with a glossary and links to key regulations, so clients are not excluded from the process.

Human-centered design: humans + AI + continuous improvement

A human-centered AI is not a finished, static product. It’s an ecosystem combining participatory design, human oversight, iteration, and continuous improvement.

Participatory design

Research by Margaret D. Hagan (Stanford Legal Design Lab) emphasizes that when designing AI for legal assistance, end users citizens, clients, and professionals must be involved from the start to understand how they would want to use the tool.

This means conducting workshops, interviews, prototyping, and usability testing.

Human oversight and feedback loop

In legal AI, the human-in-the-loop model is essential: the professional reviews, corrects, and provides feedback. The system must, in turn, collect that feedback to improve its interface, usability, and outputs.

Ethics, values, and design

The Value Sensitive Design (VSD) approach stresses that human values fairness, dignity, justice, transparency must be considered throughout the technological design cycle.

For legal AI, this means asking:

• How does this tool affect vulnerable clients?
• What barriers to access does it create?
• How much effort does it demand from users?
• Does it foster technological dependence?

How we’ve implemented this in GenIA_L

In GenIA-L, a human-centered approach is not just a design promise it’s an established practice.

From the beginning, we have worked with legal and tax professionals, lawyers, advisors, compliance officers, and representative clients to identify real needs, pain points, and usage expectations before developing any feature.

GenIA-L updates go beyond incorporating new laws or case law they also include usability and UX improvements based on real user interaction data and feedback.

These efforts ensure that GenIA-L is designed around people, not data: a technology that understands professional work and reinforces client trust.

Challenges and how to address them

Challenge: resistance to change and learning curve

Although AI can be powerful, many legal professionals have little confidence or experience with such tools. The key lies in an experience simple enough to encourage use, but with depth available for “power users.”

Challenge: “technical accuracy vs. usability”

A highly capable but complex system may remain an “experiment” if it is not usable. This reinforces that usability is a determining factor.

Challenge: maintaining a human focus without losing automation

Balancing automation (efficiency) and human oversight (quality) is essential. Clear processes are required when AI suggests, when the professional reviews, when they intervene, and how it is documented.

Challenge: client accessibility

Most tools are designed for professionals, but clients interact with them too. Ensuring that clients do not feel excluded or confused is key to trust and added value.

Conclusion: innovation starts with people

Designing human-centered legal AI is both a competitive advantage and a quality imperative. It is not enough for a tool to work it must be used, useful, accepted, and understood.

When law firms and advisory practices adopt this approach, they achieve:

• higher internal adoption rates.
• more engaged and satisfied clients.
• faster processes without sacrificing trust.
• an AI vision aligned with professional values.

The GenIA-L platform, with its legal infrastructure, traceability, and responsible approach, provides a solid foundation.

The next step is to strengthen its human-centered design making it intuitive for professionals, understandable for clients, and adaptive in usability.

Because in the legal field, technology must not only solve it must accompany: the professional who advises and the client who trusts. And that human connection begins with design.