Legal research with citations
Research federal and state case law, the U.S. Code and federal regulations, with a pinpoint citation for every statement.
Legal AI for US law firms
Take on more matters.
Keep the partner standard.
Gotham helps US law firms research federal and state law, review agreements across a data room and draft in Word, in one workspace. Every answer cites the passage it relies on, so associates move faster and partners can check the work.
Primary valueMore matter capacity and consistent work product
Control pointA licensed lawyer stays responsible for the work
Data locationStored in the United States
The buying problem
Faster drafts are not the goal on their own. Work product has to pass partner review, cite real authority, protect privilege and client confidences, and fit the firm's own drafting and approval steps. Courts have sanctioned lawyers for filing AI-written briefs with fake citations.
The business case also differs by practice. Saved time can mean faster turnaround, better margins on fixed fees, or more capacity, but it only counts once the firm can use that time well. Measure each workflow on its own instead of trusting a single productivity number.
Where Gotham can enter the work
Contract review and redlines
Review agreements against your playbook and mark them up with tracked changes in Word.
Due diligence across a data room
Put a data room of Delaware, New York and California-law agreements in one grid, with each cell linked to its clause.
Litigation chronologies
Build a chronology from the record, with each event linked to the page it came from.
Drafting
Draft memos, letters, side letters and board papers in your firm's style, and export them as Word files.
Firm controls
Give each person a role, keep matters separate, apply ethical walls and run the same workflow on every matter.
What to require in evaluation
Correction time
Measure how long partners and associates spend fixing the output, not only how fast it appears.
Citation checks
Require real sources, pinpoint citations and a check that each case is still good law.
Professional duties
Check the tool against your duties of competence, confidentiality and supervision, and ABA Formal Opinion 512 on generative AI.
Client terms
Confirm you can meet client limits on AI use, data location, retention and subprocessors.
Access and ethical walls
Test who can see each matter, precedent and playbook, and how screens work across teams.
Daily tools
Try it in Word, your document system and your export formats on real work.
Align the buying committee before the pilot
Questions to resolve
- Which partner owns quality and decides when to expand?
- Which client engagement terms limit AI use, data location or subprocessors?
- How will risk, IT, security and knowledge management sign off?
- What will finance accept as real value, not estimated hours?
Rollout controls
- Start with trained reviewers and a few matter types.
- Require citation checks and record material corrections.
- Keep matter permissions, confidentiality and client rules in place.
- Expand only after quality and adoption targets are met.
Good first pilot candidates
A useful pilot has bounded inputs, a known reviewer and an output whose quality can be scored. Consider:
- Review one agreement type against an approved playbook.
- Research a bounded question against authorities a senior lawyer already knows.
- Build a chronology from a closed-matter record.
Start with the workflow, not the licence count.
Use a private assessment to choose the first test and its success measures.