Swedish legal tech company Aloi has built a system that captures and deploys a law firm's institutional judgment across new matters. CEO Johan Häger told Artificial Lawyer the approach moves beyond simple document retrieval to identify decision patterns in past transactions - addressing a capability critics have long said AI could not handle.
"There's been one thing that has been problematic and that's really utilizing your larger data sets and larger unstructured data sets that sits in all law firms," Häger said. The company's technology combines several layers to surface how specific risks were approached and mitigated in previous deals.
How the judgment layer works
Aloi's system rests on four interconnected components. A model spec defines the structure of legal documents - teaching the system to recognize dependencies and rules within contracts. Metadata fields, sometimes numbering in the hundreds, are applied to each document. A judgment graph learns the context of entire transactions, recognizing that a single clause in a share purchase agreement cannot be interpreted in isolation from the 50 to 100 other documents in an M&A deal.
Version handling forms the fourth pillar. A transaction might generate 20 or 30 versions of a document across both sides of a negotiation. Tracking what changed between versions reveals how lawyers identified and mitigated specific risks. "All of this together comes to a conclusion and a recommendation for the lawyer," Häger said. "We see the decision patterns throughout the data."
DMS search versus judgment retrieval
A document management system can find past clauses or precedents with reasonable precision. It cannot surface how decisions were made. Häger drew a sharp distinction: "It will not see decision patterns in the data. It will not have that sophistication level that you actually need for sophisticated drafting, sophisticated negotiation." Law, he said, centers on risk identification and mitigation. Knowing how a firm approached a particular risk across dozens of transactions provides value that keyword search cannot match.
The system responds to lawyer prompts with context-aware results. A lawyer drafting a merger provision for a specific client can ask how the firm handled similar clauses for that client in the past - or narrow further to deals involving IP-heavy businesses or seller-friendly terms. The precision depends partly on data volume. Häger noted that the jump from 10 to 100 precedent documents matters significantly, while the difference between 1,000 and 10,000 shows diminishing returns. Smaller firms with 30 to 40 lawyers still get strong results from the baseline product.
Where Aloi fits in the legal tech stack
Häger described a three-layer architecture. DMS platforms sit at the bottom, handling document storage and filing. Data processing companies like Aloi occupy the middle layer, structuring that data for intelligence retrieval. Application-layer tools with workflow and drafting interfaces sit on top, facing the end user. Aloi integrates with both levels - ingesting from DMS systems and feeding structured data to the applications lawyers already use.
Deployment ranges from a baseline product that onboards in hours to customized implementations for larger firms. A recent engagement sent five engineers to a Norwegian law firm to set up an MCP connection integrating Aloi with another product in the firm's stack. Ingestion of millions of documents typically completes within days.
For legal professionals looking to understand how AI is reshaping document-heavy work, AI Learning Path for Paralegals covers the tools and techniques emerging in this space. The skills translate directly to the kind of data-supported drafting Aloi enables.
What this means for associates and the law firm model
Häger predicted the technology will compress the learning curve for junior lawyers. A first-year associate can draw on the crystallized experience of senior partners across thousands of transactions. "A lawyer will flow into your desktop," he said. "It's sort of the experience of all the senior partners that you have available at your fingertips every day."
He also forecast pressure on pricing for low-end work. When institutional knowledge becomes instantly retrievable, charging for simple documentation becomes difficult to sustain. The value shifts to the data itself - the firm's proprietary repository of past decisions - and to the higher-value advisory work that data supports. Firms that build this capability will develop stickier client relationships, Häger argued, because they will learn client preferences systematically rather than relying on individual partner recall.
Why this matters for legal professionals
The judgment layer approach changes what it means to develop expertise inside a law firm. Instead of spending years absorbing institutional knowledge through apprenticeship, associates can query it directly. The technology does not replace lawyer judgment - Häger was clear that the system provides recommendations while the lawyer and client make the final call. But it shifts the profession from individually dependent knowledge toward data-driven advice backed by the full weight of a firm's transaction history. Firms that ignore this capability risk watching clients gravitate to competitors who can demonstrate they have actually learned from every deal they have ever done.
Your membership also unlocks:
