Dormakaba: A Traditional Industrial Company in the AI Era

The dormakaba Group is a leading provider of access and security solutions with over 150 years of corporate history. With products and markets on every continent, the company faces a classic challenge: global contract volume, small legal team.

When dormakaba recognised the challenge, the legal team opted for AI support in contract review in order to delegate it to the business.

At a glance:

• 9 lawyers (7.8 FTE) plus 3 contract managers
• All continents, a wide range of jurisdictions and languages
• More than 10,000 contracts per year, from simple NDAs to complex SaaS agreements
• 150+ years of corporate history

The Challenge: Many Contracts — Few Resources

The legal department cannot possibly review every contract with the same high quality and consistency. The business has to be able to provide first-line legal support itself, without compromising quality or compliance standards.

1. Capacity bottleneck: 9 lawyers for 10,000+ contracts per year — an equation that simply doesn't add up without technological support.

2. The articulation gap: Experienced lawyers intuitively know what to watch out for — but this implicit knowledge is hard to scale. "I know it when I see it" is not a process.

3. Playbook creation as a bottleneck: Manually capturing clause rules in Excel, team alignment and qualification processes took weeks to months — and had to be repeated for every new contract type.

4. Consistency across borders: Different jurisdictions, languages and local business units demand uniform quality standards — hard to guarantee with manual processes.

5. Strategic overload: When lawyers are tied up with routine reviews, capacity for strategically valuable, complex matters is missing.

The Solution: Agentic Legal AI in Three Phases

Together with Legartis, dormakaba implemented a three-stage agentic approach that systematically translates legal expertise into AI instructions. It then scales autonomously.

Phase 1 – Knowledge Acquisition: AI-Assisted Knowledge Transfer
A Playbook Agent in the Legartis Legal Workspace systematically captures company-specific requirements per contract type through dialogue. The implicit knowledge of the lawyers is formalised.

Phase 2 – Playbook Generation: Automated Contract Playbooks
Playbook Creation Agents turn the captured knowledge into complete contract playbooks: precise clause definitions, mandatory/optional classification, risk prioritisation and recommended actions — in hours instead of weeks.

Phase 3 – Autonomous Review: Validated AI Contract Review
The system validates itself via F-scores using test cases. An F-score of 0.95 guarantees 95% precision. Business units can now review contracts independently — with legal quality assurance running in the background.

Results: What Has Really Changed

−67% time saved on 100-page contracts
Review time reduced from 1.5 hours to 30 minutes. Lawyers focus on truly critical clauses instead of low-risk ones.

Eliminated risk of oversight
Consistent review across all contracts. No routine errors due to fatigue. Full logging for end-to-end traceability.

Global scaling to business units
Contract review delegated to local teams. Junior colleagues review complex contracts with playbook support — the legal team takes over strategy.

F-score 0.95 — self-validating AI precision
The system validates whether company requirements are being applied correctly.

Playbook creation in hours instead of weeks
What used to be a months-long manual process now takes hours. Dormakaba calls it a "quantum leap".

Strategic re-positioning of the legal department
Freed-up capacity flows into strategic legal work. The legal team becomes a driver of business decisions — no longer just a reviewer.

"Agentic Legal AI is coming — it's only a matter of when we jump on board. Anyone who doesn't integrate AI into their daily work simply isn't working as efficiently as possible."

— Cedric Ruepp, General Counsel Europe, Africa/APAC, dormakaba Group

Implementation Learnings: What Cedric Ruepp Tells Other General Counsels

Trial and error is normal
The path is rarely linear. Recognising pain points, testing solutions, iterating — that is the real transformation process, not the vision from the slide deck.

Cultural change takes time
Old legal working habits don't disappear overnight. Self-reflection helps: in contracts from earlier years, you often find improvement points today — the AI is more honest than your own memory.

Engagement is not optional
Agentic Legal AI is like a sports car — it needs someone to take the wheel. An internal driver, especially during the build-up phase, makes the difference between potential and result.

The right partner matters
Flexibility across contract types and jurisdictions, true multilingual coverage (no machine translation), dynamic product development and a Success Manager with best practices — these criteria separated the field of Legal AI providers. Dormakaba chose Legartis.

The Evolution of Legal AI: Where Dormakaba Stands — and Where the Journey Is Heading

Stage 1 – Assistive AI (past)
Text analysis, summaries, clause search, proofreading. AI as a passive tool — helpful, but not scalable.

Stage 2 – Agentic Legal AI (today · dormakaba)
Independent contract review, risk classification, automated revision suggestions, global consistency monitoring.

Stage 3 – Autonomous Systems (near future)
Automated contract negotiation, direct document adjustments in real time, proactive risk management before contract conclusion.