Vibe Legal: Describe the Intent. The System Drafts the Contract.
The first draft of your next significant contract will take four days to produce — and every hour of those four days is spent retrieving and assembling knowledge your lawyers already possess. Thomson Reuters' 2025 Legal Market Report found that first-draft preparation accounts for 34% of associate time at mid-size European law firms. That figure has barely moved despite a decade of investment in document automation, contract management software, and AI copilots. A decade of tools. A persistent 34%.
Each generation of legal software made retrieval and formatting faster. None changed the assembly burden — the work of translating a deal's intent into 40-section contract structure using clauses the firm has already approved, in language the firm has already decided is acceptable. That translation is what four days buys. And it's knowledge-retrieval, not legal judgment.
AI copilots launched since 2022 haven't changed this. They make the typing faster, suggest the next clause, predict the missing term. The lawyer still describes what they want, reviews the output, rewrites the clauses that missed the firm's standard positions, and corrects the governing law. A lawyer who spent 90 minutes correcting a GPT-generated non-disclosure agreement that used the wrong jurisdiction and omitted the firm's standard carve-outs has a legitimate reason to be skeptical of the entire category. That experience is common. The frustration is valid.
The Problem Is Translation, Not Speed
Most legal AI tools are built on a mistaken diagnosis. The bottleneck in contract drafting isn't speed — it's the translation from intent to standard positions. A lawyer who knows exactly what a shareholders' agreement should achieve still has to translate that knowledge into 140 clauses, applied consistently, using language the firm has approved across prior engagements. The knowledge exists. The assembly is the burden.
Vibe Legal is built on a different model: encode your firm's or company's standard positions once, and the translation becomes automatic. When a new contract need arises — described in plain language: "service agreement with a US software vendor, our standard limitation of liability, Luxembourg governing law, three-year term" — the system assembles a complete draft reflecting your actual institutional positions. Not a generic template. Your positions, assembled correctly.
Prompt-and-review legal AI is structurally different from this. Harvey AI, the US legal AI startup valued at $3B, operates on a prompt-and-review basis: lawyers describe what they want, the system generates an output, lawyers rewrite what doesn't match their standard positions. Vibe Legal encodes the positions as inputs. When the firm's decision about limitation of liability has already been made and encoded, the system doesn't need to be told again — it applies the firm's position consistently, every time.
What Actually Changes
A Luxembourg law firm with 28 lawyers deployed Vibe Legal through the Leeloo Framework on its corporate transaction practice and cut first-draft time from four days to under four hours for standard mergers and acquisitions documents. The senior lawyer review step was preserved — nothing in the process removed the judgment layer. What changed was where associates' time went. Instead of spending 60% of their engagement time on first-draft assembly, they now spend it on substantive review, negotiation preparation, and the client communication the firm actually bills for.
In-house legal teams see the difference in deal velocity. A Brussels-based in-house legal team at a €400M logistics company encoded its standard supplier agreement positions and now responds to supplier contract requests within 24 hours — compared to a prior average of six working days. The team didn't grow. The backlog cleared because the first draft stopped being the constraint. Legal became an accelerant in supplier onboarding rather than a six-day approval gate.
Complexity is where the counterintuitive advantage lives. Simple contracts — non-disclosure agreements, standard service terms, employment agreements — are already template-driven and require minimal drafting effort. Vibe Legal's return is highest on complex, bespoke agreements: joint ventures, shareholder agreements, complex licensing deals where first-draft burden is measured in days and where inconsistency risk compounds across a portfolio. Encoding positions for a shareholders' agreement covers 120–180 decision points — governing law, drag-along thresholds, pre-emption structures, exit mechanics, deadlock provisions — and applies those positions consistently across every subsequent agreement without requiring a senior lawyer to re-decide each one.
The Position Encoding Phase — What It Requires
Encoding positions for complex documents is not a quick process. A shareholders' agreement with jurisdiction variations and investor-type variations may require 3–4 weeks of structured sessions with the firm's senior lawyers before the first automated draft runs cleanly. That upfront investment is real and should be named clearly.
Two rewards come from the encoding work beyond the automation itself. First, the firm's positions become explicit, documented, and pre-approved — so when a counterparty pushes on a specific clause during negotiation, the response is "our position on this is X and our fallback is Y" rather than "let me check with the team." Second, encoding surfaces internal disagreements about standard positions that were previously hidden by individual lawyer variation. Firms that go through position encoding report that the clarification work has value independent of what it enables technically.
EU AI Act provisions on AI-generated legal documents — requiring explainability, human oversight documentation, and audit trails — take effect progressively through 2026. Vibe Legal deployed on the Leeloo Framework includes complete audit trails: which encoded position generated each clause, which lawyer approved the encoding, and which human review step occurred before the document was sent. Firms using public AI tools to generate contract drafts and clean up the output will need to retrofit those compliance controls under regulatory deadline pressure. Building the position-encoding model now satisfies the requirements by design.
Running the Numbers
Law firms see the ROI as a capacity question. If 15 associates spend 34% of their time on first-draft assembly at €1,200/day fully-burdened cost (salary, benefits, office, and overhead combined), that's roughly €1.5M per year in time spent on knowledge-retrieval that a position-encoding system handles automatically. Vibe Legal converts most of that drafting time to billable review and negotiation work at the same rate — increasing effective capacity without increasing headcount. The margin improvement is structural: senior lawyers' positions are applied consistently by the system rather than approximated by junior staff whose work requires senior correction.
Companies with in-house legal teams see the benefit differently. A €400M company that closes supplier contracts 15 days faster on average, across 200 annual agreements per year, generates measurable working capital improvements and competitive advantage in supplier relationships. Legal speed becomes a commercial asset. Legal teams that operate as deal accelerants get resourced differently by management than legal teams that operate as approval queues.
Legal risk also accumulates through contract inconsistency in a way that manual drafting conceals. Each agreement drafted by a different associate applies slightly different language to the same concept — "best efforts" in one contract, "reasonable endeavours" in another, "commercially reasonable efforts" in a third. Across 400 supplier agreements, those inconsistencies create a liability portfolio no one has mapped. Position encoding eliminates the accumulation mechanism by applying approved language to each concept automatically, every time.
What the Legal Function Looks Like After
Picture the General Counsel whose team turns around a complete first draft for any standard commercial agreement within four hours of receiving the brief. Lawyers who spend engagement time on negotiation preparation, counterparty analysis, and the judgment calls that actually require legal training. A documented position library that answers "what is our standard position on X?" with a reference rather than a phone call. An audit trail that satisfies regulatory scrutiny without a manual reconstruction effort.
Your lawyers weren't trained to assemble known clauses into standard positions. They were trained to recognize when the standard position isn't good enough for this deal. Give the assembly to the machine and the judgment back to the professionals — and watch the legal function become the part of the business that makes deals move faster.