Ex-Sidley Chair Bets Big on an AI-Native Law Firm: Norm Law
Mike Schmidtberger, former chair of Sidley Austin's executive committee, isn't easing into retirement. He's now chairman and head of investment funds and regulatory at Norm Law LLP, a two-month-old firm built around AI agents and "legal engineering" instead of billable hours.
This is a real-world test: can the partner who helped double Sidley's revenue wire that operating discipline directly into AI systems - and build a profitable, scalable firm for institutional clients?
Why this move matters
From 2018-2025, Schmidtberger led Sidley's executive committee, managed the New York office, and co-led investment funds. After stepping down under Sidley's age rules, a visit to Norm Ai's office convinced him to switch tracks.
The bet is simple: for funds and heavily regulated clients, consistent outcomes at scale beat artisanal drafting. AI-native workflows promise that consistency - with lawyers staying accountable for judgment and sign-off.
How Norm Law works
Norm Law was launched in November as the legal-services counterpart to Norm Ai, which builds AI agents encoding legal and compliance workflows. Lawyers design the workflows; agents generate first drafts for regulatory analysis, policy mapping, playbooked negotiations, and document production; partners and associates review, edit, and advise.
The firm is focused on financial-services and other regulated sectors. Early hires include former Brown Rudnick partner David Sorin and former Cadwalader partner Mike Rupe, plus senior associates. More partners are expected as the platform scales.
Fuel and proximity
Blackstone invested US$50 million in Norm Ai and is collaborating through its in-house legal team to shape Norm Law's offering. Other investors include Bain Capital, Vanguard, Coatue, and New York Life.
This capital and client access let Norm design agents around specific regimes and in-house policies, turning bespoke compliance know-how into reusable, auditable systems.
What's different from traditional firms
- Core production: AI agents plus lawyers generate and review work product; legal engineering is the center, not a support function.
- Pricing: Outcomes- and value-based fees tied to efficiency and defined deliverables, not time spent.
- Roles: Partners and associates design and govern workflows, supervise outputs, and focus on judgment-heavy tasks.
- Knowledge: Know-how is productised into agents that run across matters and clients, not just stored in precedent banks.
- Market position: Co-exists with Big Law by taking repeatable, rules-heavy work; frees internal and external counsel for strategy and disputes.
Why "legal engineering" is the hinge
Norm hires lawyers to encode doctrine and process into machine-readable workflows - an extension of Schmidtberger's push at Sidley to operate as a "collective intelligence."
For in-house teams, the pitch is clear: consistent decisions across regions and business lines, with audit trails, while external counsel focuses on complex structuring and contentious matters.
What institutional clients should test before buying
- Scope fit: Map your repeatable, rules-heavy tasks (e.g., marketing reviews, trading policy checks, fund disclosures) and confirm agent coverage.
- Accuracy rates: Ask for benchmark results, error types, and human-in-the-loop review gates for each workflow.
- Regulatory alignment: Validate that your specific regimes and house policies are codified (and version-controlled) in the agents.
- Auditability: Ensure full logs: inputs, prompts, sources, drafts, reviewer edits, and final decisions with timestamps.
- Privilege and confidentiality: Confirm data handling, on-prem or VPC options, retention windows, redaction, and model training boundaries.
- Model risk and validation: Look for formal testing, drift monitoring, and a rollback plan. Consider aligning with the NIST AI Risk Management Framework.
- Liability and conflicts: Clarify responsibility if an agent-assisted output is wrong; review indemnities, caps, and conflict checks.
- Metrics and pricing: Tie fees to defined deliverables, cycle time reduction, error rates, and coverage expansion.
- Change management: Demand playbooks for rollout, training, exception handling, and ongoing update cadences.
- Ethics and supervision: Ensure compliance with lawyer-oversight duties and tech competence expectations (see ABA Model Rule 1.1).
Career impact: the rise of legal engineers
Expect new roles for associates and mid-levels who enjoy workflow design, data, testing, and client implementation. The skill stack: strong doctrine, process mapping, drafting standards, prompt construction, and QA.
If you're building capability inside your team, consider focused training paths for workflow design and AI-assisted review. A practical starting point is a curated set of role-based courses: AI courses by job.
Strategic options for incumbent firms
- Build: Stand up internal legal-engineering pods; convert high-volume work into agents; price on outcomes.
- Partner: Use AI-native firms as a production layer; keep advisory and bespoke structuring in-house.
- Productise KM: Move beyond precedent banks; turn best practices into executable, auditable systems.
- Rewire incentives: Reward reuse, cycle-time gains, and accuracy - not hours logged.
The open question
Schmidtberger helped double Sidley's revenue by professionalising knowledge and management. Norm Law is the sharper test: can that same mindset, embedded in AI agents with accountable human oversight, become the core architecture of a profitable, scalable firm?
If institutional buyers say yes, repeatable regulatory and documentation work won't live inside time sheets anymore - it will live inside systems that lawyers supervise.
Who's building it
Founder John Nay says Norm was built "backwards" from conversations with existing Norm Ai users who wanted a firm architected around AI from day one - not a legacy practice bolting on tools. Early momentum plus capital and client proximity will make the next 12 months the real proof point.
Your membership also unlocks:
