Banner artwork by RerF_Studio / Shutterstock.com
This series explores how senior legal professionals are navigating artificial intelligence adoption across a large global enterprise, offering practical perspectives for in-house teams of any size.
Manuel Morasch is VP Legal and leader of ABB’s Legal & Integrity Transformation Program, helping shape a human-led, technology-enabled future for more than 450 lawyers.
His pragmatic framework begins with data, adds governance, and scales toward an “AI first, Legal always” operating model.
Where to start
Spiwe L. Jefferson: You spend a great deal of time evaluating the legal AI market. What’s your overall assessment of where it stands today?
Manuel Morasch: Legal artificial intelligence has moved into a meaningful adoption phase. Major law firms and corporate legal departments are deploying solutions, and organizations that build capability now may be better positioned as the market develops.
Q: With hundreds of vendors competing, how do you make sense of all the options?
Morasch: The market remains broad and fragmented. No single vendor yet provides an end-to-endsolution, so many legal teams combine specialized tools for contracting, research, litigation, and data management. Some work out of the box; others require investment in templates and playbooks but may offer greater long-term flexibility.
Q: If you had to identify the single most important prerequisite for effective AI deployment, what would it be?
Morasch: Data quality, without question. Centralized, well-managed contract repositories provide a critical foundation. When contracts are scattered, inconsistently named, or lack metadata, reliable output becomes much harder to achieve.
Centralized, well-managed contract repositories provide a critical foundation.
Q: That sounds like a significant undertaking for a global legal function. Where do you start?
Morasch: A practical starting point is a cross-divisional contract data source. Prioritize high-value contracts, standardize key metadata such as governing law and assignment rights, apply version control, and remove duplicates before migration. The aim is simple: bring what the organization knows into one dependable place.
Governance and data-security guardrails
Q: Once the data foundation is in place, what comes next?
Morasch: Governance. Clear data boundaries and vendor data-use policies can help legal, compliance, security, and risk teams align before deployment. One safeguard remains firm: vendor models should not train on ABB data without explicit consent. We use closed models and clear contractual controls.
Clear data boundaries and vendor data-use policies can help legal, compliance, security, and risk teams align before deployment.
Q: Is that a realistic demand to make of vendors?
Morasch: Yes. For enterprise legal work, a clear non-training commitment is an important condition of trust and a practical test of vendor suitability.
Centralized core, decentralized execution
Q: How does ABB’s organizational structure shape your AI strategy?
Morasch: “The ABB Way” combines decentralized businesses and embedded legal teams with a central core. Our strategy reflects that balance through three group-level pillars:
- Contract repository: a single, cross-divisional source of contract data that supports M&A diligence, counterparty management, and consistent analysis.
- Shared risk and playbook framework: unified risk thresholds and decision frameworks ensuring consistent legal standards across all divisions.
- AI governance policy: clear policies on data boundaries, vendor relationships, and model-training restrictions.
Divisions and their respective commercial legal groups retain autonomy in tool deployment but operate within these guardrails.
The phased roadmap from foundation to continuous improvement
Q: Can you walk us through how you sequence this work?
Morasch: We think about it in four phases.
- Foundation. Establish the contract repository, clean the data, and define governance. Investing here can make later deployment more reliable.
- Pilot programs. Test low-stakes, high-frequency uses such as routine agreements, transcription, and research, with human review and integration in mind.
- Scale and optimize. Add divisional legal-technology owners, a lightweight group council, consistent vendor terms, and broader deployment where pilots demonstrate value.
- Continuous improvement. Build verification practices, safe testing environments, leadership support, and clear ownership for each tool.
Disclaimer: The information in any resource in this website should not be construed as legal advice or as a legal opinion on specific facts, and should not be considered representing the views of its authors, its authors’ employers, its sponsors, and/or ACC. These resources are not intended as a definitive statement on the subject addressed. Rather, they are intended to serve as a tool providing practical guidance and references for the busy in-house practitioner and other readers.
