AI for law firms, with a source behind every answer.
Legal work does not tolerate guesswork. The systems behind Lovguiden, which follow Danish legislation, case law and parliamentary work around the clock, can be adapted to a law firm's practice areas, cases and internal knowledge.
- 3M+
- cross-references · Lovguiden
- 2,000+
- public sources · Lovguiden
- 35
- MCP tools · Lovguiden
- 138K+
- verdicts & decisions · Lovguiden
- 01
New rulings and decisions have to be found by hand across courts and appeals boards.
- 02
Legislative changes and bills are noticed late, often only once they already matter to a case.
- 03
Research memos take hours, and every source has to be checked afterwards.
Monitoring per practice area
New rulings, decisions, bills and consultations in the chosen areas of law are collected, summarised and sent as a report several times a day.
Research with source references
First-draft memos in which every statement is tied to a provision, a ruling or a preparatory work. References that can't be looked up are removed before the memo is shown.
An AI assistant on the firm's own documents
Search and an AI assistant across the firm's own memos, contracts and case files, run on local models when confidential information must not leave the building.
Danish law inside AI tools
Lovguiden's MCP server gives Claude, ChatGPT and Copilot access to Danish legislation and case law through 35 tools, so answers rest on the law as it stands.
- 01
Conversation
The work that takes too long is described, followed by an honest assessment of whether AI and automation can take it over, and roughly how.
- 02
Prototype
A first version on real data, so the result can be seen working before anything larger is committed to.
- 03
Build
The full system, built in short iterations with regular demos, so progress can be followed all the way.
- 04
Operate
Operated and monitored as a service, or handed over with code and documentation to an in-house team.
What about confidentiality?
Confidential documents can be processed by local models on the firm's own infrastructure, so they are never sent to external AI services.
How reliable are the answers?
Answers are built from retrieved source material, not from the language model's memory. References that can't be looked up are removed, and gaps in coverage are stated plainly.
Can it connect to existing systems?
Usually, yes. Data can be read from and delivered to case and document systems through APIs, and results can be made available in the AI tools the firm already uses.
The same approach, for law firms.
A short description of the task is enough to start the conversation.