Your lawyers spend too much time on work that doesn't require legal judgment
Liorant builds AI assistants for legal teams — on Microsoft 365, Google Workspace, or Claude — that handle the repetitive work so your lawyers focus on the judgment that clients actually pay for.
Four problems that slow down every legal practice
Legal operations don't fail because of strategy. They fail because of friction — the same friction, every day, at every level of the firm.
Contract review takes hours, not minutes.
Reviewing SaaS agreements, NDAs, DPAs, IP clauses, and AI provisions clause by clause consumes senior lawyer time on work that follows predictable patterns. The judgment is legal. The scanning is not.
Internal knowledge is invisible.
Memos, opinions, and standard clauses from past matters sit in shared drives that no one searches effectively. Every new matter starts from scratch. Work gets duplicated. Advice becomes inconsistent.
Client intake is unstructured.
Queries arrive as emails, calls, or chat messages — without scope, jurisdiction, timeline, or risk signals. Someone has to extract that information manually before the matter can even be qualified.
Due diligence runs on pressure.
IP, data contracts, open-source licences, privacy exposure, and AI risk reviewed under M&A deadlines means red flags get missed. The data room is large. The window is short.
How we work — activation first, then ongoing operations
Activation
We pick your highest-friction workflow, configure your existing AI platform, and deliver a working assistant with documentation and governance guardrails.
Expand
We add one new workflow per month from an opportunity backlog we build together. Each new assistant deploys faster because we already know your systems.
Operate
We monitor, optimise, and report every month. Outcome reports state hours recovered and matters accelerated — not system uptime.
What this looks like in practice
Activation (weeks 1–6)
We start with a scoped conversation — 30 minutes, no slides. We identify the workflow with the highest friction-to-fix ratio, typically contract review, research synthesis, or client intake. We then configure an AI assistant on the platform your firm already runs (Microsoft 365 Copilot Studio, Google Gemini Enterprise, or Claude), test it against real matter samples, document the governance guardrails, and hand it over. At the end of 4 to 6 weeks, your team has a working AI assistant, not a pilot report.
Expand (ongoing)
During the activation phase, we build an opportunity backlog — a ranked list of additional workflows that are ready to automate. We then implement one per month. By month 6, most legal teams we work with have 4 to 6 active AI assistants covering distinct parts of the practice.
Operate (ongoing)
We monitor every deployed assistant for prompt drift, platform updates, connector failures, and compliance changes. We report monthly on outcomes in hours and euros. We review the backlog quarterly so there is always a concrete next priority — and your team always knows what they are getting.
Eight use cases we have built for legal teams
Contract review
Bottleneck: Reviewing SaaS agreements, NDAs, DPAs, licences, and AI/IP clauses clause by clause takes hours per document.
Result: The assistant marks deviations from firm playbook before a lawyer opens the document. First-review time drops. Focus shifts to judgment, not scanning.
AI assistant · Contract deviation detectionLegal research
Bottleneck: Manual search across legislation, GDPR guidance, AEPD, CNMC decisions, BOE, and prior internal memos — then synthesising into a memo under time pressure.
Result: Research and memo cycles accelerate. The assistant surfaces relevant sources, internal precedents, and regulatory updates in a single query.
AI assistant · Research synthesisInternal knowledge
Bottleneck: Prior memos, opinions, and standard clause positions are buried in unstructured folders. Work gets duplicated. Advice is inconsistent across matters.
Result: An "Ask the Firm" assistant retrieves precedents, clause positions, and internal know-how in natural language. Knowledge compounds instead of being lost.
AI assistant · Knowledge retrievalClient intake and scoping
Bottleneck: Queries arrive as unstructured emails or calls. Scope, jurisdiction, parties, and timeline need to be chased before a matter can be qualified or assigned.
Result: Intake queries are converted into structured matter briefs automatically. Faster qualification. More professional first response to high-value prospects.
AI assistant · Matter brief generationBusiness development intelligence
Bottleneck: The firm wants more tech-sector clients but has no systematic method for tracking regulatory shifts, AI developments, or client signals.
Result: A weekly briefing covers regulatory trends, AI legislation, SaaS sector news, and cybersecurity developments. Partner conversations become proactive.
AI assistant · BD intelligence briefingPrecedent drafting
Bottleneck: Lawyers search for old templates or adapt generic documents that don't reflect the firm's current standard positions — under time pressure.
Result: The assistant drafts from approved templates and intake responses. First drafts arrive before the matter meeting. Language matches firm standards.
AI assistant · Template-based draftingDue diligence
Bottleneck: IP, data contracts, open-source licences, privacy exposure, and AI risk reviewed under M&A deadline pressure. Data rooms are large and windows are short.
Result: A red-flag triage assistant maps the data room against a DD checklist and surfaces critical issues. Faster triage. Fewer omissions. Better partner leverage.
AI assistant · DD red-flag triageBilling narratives
Bottleneck: Time entries are incomplete, inconsistent across email threads, research sessions, calls, and documents. Billing is under-recorded and time-consuming to reconstruct.
Result: The assistant drafts time narratives from calendar entries, matter notes, and documents. Compliant, consistent, complete — without reconstruction work.
AI assistant · Time narrative generationFrom one AI assistant to a practice-wide AI operation
Phase 1 — Setup
- Use case selection and data architecture
- Initial configuration on Microsoft Copilot / Gemini Enterprise / Claude Cowork
- First AI assistant built and tested
- Governance documentation and handoff
Phase 2 — Progressive build
- One new AI assistant per month from the opportunity backlog
- Iteration on deployed assistants based on real usage
- Prompt tuning, retrieval tuning, adoption support
Phase 3 — Ongoing operations
- Monthly monitoring (platform health, connector checks, compliance)
- Outcome reporting in hours and euros
- Quarterly backlog review with next priorities
- EU AI Act governance updates as obligations phase in
Client data is processed on enterprise-grade infrastructure. Your systems. Your controls.
How Liorant compares
| Feature | Liorant | In-house AI hire | IT project | SaaS tool |
|---|---|---|---|---|
| Time to first working AI | 4–6 weeks | 3–6 months | 3–4 months | Immediate, rarely adopted |
| Legal-specific workflows | 8 proven use cases | Generic | Generic | Fixed features only |
| EU AI Act / USA / El Salvador compliance | Built in, no extra cost | Depends on the hire | Extra engagement | |
| Runs on existing tools (Microsoft / Google / Claude) | Yes | Yes | Sometimes | Adds new vendors |
| Human oversight designed in | Standard | Varies | Rarely | |
| Ongoing optimisation | Monthly | At salary cost | Project ends at delivery | |
| Outcome reports in hours and euros | Every month | Rarely | ||
| Scales across practice areas | One new workflow/month | Limited by one person's capacity | New project required | Limited by product |
| Multilingual delivery | Rarely | No | No |
We build to the AI regulations that apply to your firm — EU, USA, and El Salvador
Liorant operates across three jurisdictions with distinct and evolving AI regulatory frameworks. Every system we deploy is aligned to the rules that apply where your firm and your clients operate.
Every AI deployment includes
- Risk classification per applicable regulation — before any system is deployed
- Human-in-the-loop checkpoints at every decision point where legal judgment applies
- Audit logs and prompt version control so every AI action is traceable and explainable
- Data residency compliance and access controls matched to your firm's security policy
- Ongoing governance review as obligations evolve across all three jurisdictions
Lawyers remain responsible for every legal judgment. The AI handles what does not require that judgment. That is the principle we design to, not the exception.
The EU AI Act is the world's most comprehensive AI regulatory framework. It phases in obligations for general-purpose AI models and mid-to-high risk AI systems through 2026–2027. For legal firms operating in Spain and the EU, this covers AI used in client-facing workflows, document analysis, and any system that informs a decision with legal consequences. Liorant builds EU AI Act compliance — risk mapping, traceability, human oversight, and documentation — into every engagement by default.
The United States regulates AI through a combination of federal executive guidance, the NIST AI Risk Management Framework, sector-specific rules (FTC, SEC, FDA), and a growing body of state-level legislation including Colorado's AI Act and California's AI regulations. For legal firms serving US clients or operating through US partners, Liorant aligns deployments to the NIST AI RMF controls, documents model behaviour and risk acceptance for sector-specific compliance, and structures human review checkpoints to satisfy emerging state-level obligations.
El Salvador passed its AI Promotion Law (Decree 234) in March 2025 and established ANIA, the national AI agency. The regulatory framework prioritises AI adoption with responsible governance — creating an environment where firms that deploy AI with clear documentation, human oversight, and risk controls are ahead of the compliance curve. For legal firms operating in El Salvador or advising Salvadoran clients, Liorant builds systems that align with ANIA's published governance guidance and positions your firm well as the regulatory framework matures.
If your firm advises tech clients on AI compliance
We offer a joint AI Act Readiness Sprint — a scoped engagement where Liorant provides technical instrumentation (system inventory, risk mapping, traceability setup, documentation pack) and your firm provides regulatory interpretation and legal sign-off. Applicable across EU, US, and El Salvador regulatory contexts. The result is a co-delivered compliance-readiness service you can offer to SaaS, AI startups, fintech, and legaltech clients.
Contact us to discuss how this worksWhat makes Liorant different
We deliver working software, not strategy decks.
Every engagement ends with an AI assistant your team can use on day one. We build, test, document, and hand over. There is no separate implementation phase billed separately.
We build on what you already run.
No new vendor contracts. No shadow IT. No licence negotiations. We configure Microsoft 365 Copilot Studio, Google Gemini Enterprise, or Claude on the infrastructure your firm already pays for.
Every AI action keeps a lawyer in control.
Human oversight is not an option we offer — it is a design principle we apply to every workflow. Legal judgment stays where it belongs: with the professional.
We report in outcomes, not metrics.
Monthly written reports state hours recovered, matters accelerated, and review cycles shortened. Not API calls. Not system availability. Outcomes your managing partner can read and act on.
Common questions
How long does the first engagement take?
The activation phase takes 4 to 6 weeks. We identify one high-friction workflow, build the AI assistant on your existing platform, test it against real matter samples, and deliver it with documentation and governance guardrails.
Do we need to change our current systems?
No. We build on the infrastructure you already licence — Microsoft 365 Copilot Studio, Google Workspace Gemini, or Claude. We configure what you have; we do not add new vendors or require data migration.
What happens after the activation project ends?
The Phase 1 scope of work includes a complete roadmap. After delivery, we continue as your managed AI operations partner — one new workflow per month, ongoing monitoring, and monthly outcome reports. You can cancel with 30 days' notice.
Does this cover AI regulations — EU AI Act, USA, and El Salvador?
Yes. Every Liorant deployment for legal teams includes risk classification, human oversight design, audit logging, and documentation aligned to the applicable regulatory framework — EU AI Act for Spain and EU-based firms, NIST AI RMF and US state-level requirements for US-facing work, and El Salvador's AI Promotion Law (Decree 234 / ANIA) for regional operations. This is standard, not an add-on.
Can Liorant help us offer AI compliance services to our tech clients?
Yes. We have a productised AI Act Readiness Sprint that legal firms can co-deliver with us — Liorant provides the technical evidence pack, your firm provides the legal interpretation and sign-off.
Your next matter starts with less friction than the last one.
Start with a free 30-minute conversation. We identify your highest-value automation opportunity and explain exactly what Liorant would build — no slides, no pitch.
4–6 week activation. Fixed price. Built on your existing tools.