RBD. Research Brief — Q2 2026
Complimentary Access · Q2 2026 Research Series

AI Governance in Partnership-Structured Professional Services: What Changes When the Profit Pool Unifies

A structural analysis of why partnership governance models uniquely constrain enterprise AI deployment, and what changes when firms consolidate from federated to unified profit structures.

Megan C. Starkey | Q2 2026 | RBD. Intelligence Center
14 Sources 3 Industries 3 Research Programs
Executive Summary

The global professional services market exceeds $1.1 trillion in legal services alone, with the combined legal, consulting, and accounting sectors representing one of the largest concentrations of proprietary expertise in the global economy. These firms are simultaneously among the most sophisticated producers of knowledge work and among the slowest to deploy AI at enterprise scale. The gap between experimentation and operational deployment is larger in partnership-structured firms than in any other organizational category.

The binding constraint is governance. Generative AI can draft contracts, summarize case law, analyze financial statements, and identify regulatory risk patterns today. The Thomson Reuters 2024 Future of Professionals survey found that 73% of law firms were actively experimenting with AI tools. Yet the same research identified that only 12% had achieved firm-wide deployment. The gap between experimentation and deployment is a governance gap—the structural inability of partnership organizations to authorize, fund, and sustain shared technology investment across autonomous profit centers.

Partnership structures create a governance problem that corporate enterprises do not face. Partners are autonomous revenue generators whose compensation is tied to individual or practice-level profitability. Shared AI investment requires pooled capital from people who have no structural obligation to subsidize technology infrastructure that benefits another practice group. The free-rider problem, well documented in commons economics, applies with particular force to knowledge-work organizations where the “commons” is a shared AI capability and the “riders” are equity partners with the mobility to leave if the cost allocation feels inequitable.

This brief extends the analysis in Enterprise AI Investment 2026 Outlook (RB-AI), which identified the 93/7 spending inversion—93% of enterprise AI budgets allocated to technology acquisition, 7% to organizational capacity to absorb it. In partnership structures, the inversion is even more extreme: technology investment concentrates in firm-level IT budgets, while the organizational capacity to deploy AI—practice-level workflow redesign, partner adoption, client-facing use case development—receives almost no structural funding.

73%
Law Firms
Experimenting with AI
12%
With Firm-Wide
AI Deployment
$1.1T+
Global Legal
Services Market
3–7%
Partnership Technology
Spend as % of Revenue
62%
Partners Citing Governance
as Top AI Barrier
Sources: Thomson Reuters, "Future of Professionals," 2024 · Thomson Reuters, "Legal Market Report," 2024 · ALM/Am Law Technology Survey · McKinsey, "Professional Services AI Survey," 2024
Section 01: The Partnership Governance Constraint

The Structural Features That Make Partnerships Different from Corporations

In a corporate enterprise, the CEO can mandate technology investment. The board approves capital allocation. The CIO or CTO executes. The authority chain is vertical, and dissenting business unit leaders can be overruled or replaced. Partnerships possess no equivalent mechanism. The managing partner leads by consensus, not by fiat. Technology investment above a threshold—typically defined in the partnership agreement—requires a partner vote. And the partners who must approve that investment are the same people whose individual profitability will be reduced by the cost allocation.

Three structural features of partnership governance uniquely constrain AI deployment. First, profit distribution is tied to individual or practice-level performance. Partners in most professional services firms are compensated through a combination of base draws, performance-based distributions, and equity stakes. The Am Law 100 data consistently shows that partner compensation is the single largest expense category for law firms, typically consuming 35–45% of gross revenue. Any shared technology investment that reduces distributable profit faces direct opposition from the compensation mechanism itself.

Second, partner vote requirements create a collective action problem. Major capital expenditures in most partnerships require approval by a supermajority of equity partners. Each partner evaluates the investment against their own practice economics, not the firm’s aggregate return. An AI system that transforms contract review delivers enormous value to the litigation and transactional practices but near-zero value to the tax or regulatory advisory groups. The partners who receive no direct benefit have a rational incentive to vote against the investment, or to demand that costs be allocated only to the practices that benefit—which fragments the shared infrastructure required for AI to function at scale.

Third, lateral partner mobility functions as a structural constraint on investment policy. Partners in professional services firms are not employees. They are mobile capital. If a firm’s cost allocation feels inequitable, high-performing partners can move to competitors—and take their client relationships with them. ALM data shows that lateral partner movement in the Am Law 200 has averaged over 3,000 moves per year since 2019. This mobility creates an implicit upper bound on shared investment: any cost allocation that pushes profitability per partner below competitive benchmarks risks triggering departures that unravel the economics of the investment itself.

Exhibit 1

Corporate enterprises and partnerships have fundamentally different authority structures for technology investment decisions

Technology Investment Authority: Corporate vs. Partnership Structure CORPORATE ENTERPRISE Board of Directors CEO — Mandates CIO/CTO — Executes BU A — Complies BU B — Complies VERTICAL AUTHORITY Dissent is overruled PARTNERSHIP Managing Partner Partner Vote — Approves/Blocks Practice A Own P&L Practice B Own P&L Practice C Own P&L Lateral exit Lateral exit DISTRIBUTED AUTHORITY Dissent is departure AI requires shared infrastructure. Partnerships require distributed authority. These are structurally opposed.
Source: RBD. structural analysis of professional services governance models, 2026.

The free-rider problem manifests with particular clarity in AI investment. AI infrastructure—training data curation, model fine-tuning, integration with practice management systems, prompt engineering for domain-specific applications—is inherently shared. The value compounds across practices: a document analysis capability trained on one practice’s work product becomes more capable when it ingests another’s. But the cost must be allocated somewhere. And in a partnership, “somewhere” means specific partners whose distributable income is reduced. The economic logic of AI demands unified investment. The governance logic of partnerships demands distributed authority over spending. These logics are in structural opposition.

Section 02: What Federated Structures Make Harder

Verein Partnerships and the Jurisdictional Data Problem

If partnership governance creates a collective action problem for AI investment, federated partnership structures compound it. The verein—a Swiss-law association that links legally separate national partnerships under a single brand—has been the dominant organizational form for the world’s largest law firms for two decades. Firms including Baker McKenzie, DLA Piper, Norton Rose Fulbright, and Dentons have used verein structures to achieve global scale without requiring profit-sharing across borders. Each national member firm maintains its own profit pool, its own partnership agreement, and its own technology infrastructure.

The implications for AI deployment are severe. A verein is a collection of independent firms that share a brand and a referral network—separate legal entities with separate profit pools, separate data, and separate regulatory obligations. The London partnership and the New York partnership are distinct organizations in legal fact. An AI system trained on London client data is legally inaccessible to New York—constrained by GDPR, attorney-client privilege doctrines that vary by jurisdiction, and conflict-of-interest rules that prevent commingling of client information across entities that are, structurally, separate firms.

Baker McKenzie’s 2024 decision to dissolve its verein structure and consolidate into a single global partnership is the definitive precedent case. The firm, with approximately 4,700 lawyers across 74 offices in 45 countries, announced the dissolution following a vote of its global partnership. The stated rationale included the need for coordinated technology investment, unified data governance, and a single profit pool that could fund enterprise-scale capabilities. The subtext was explicit: the verein structure had become an impediment to the kind of integrated technology deployment that a $3.3 billion-revenue global firm required to remain competitive.

The Baker McKenzie restructuring is instructive precisely because it was not driven by financial distress. It was driven by the recognition that federated governance and enterprise-scale AI are structurally incompatible. When your AI strategy requires unified training data, consistent model deployment, and coordinated workflow redesign, and your governance model prohibits data sharing across entities, prevents coordinated capital allocation, and maintains separate technology stacks by jurisdiction, the governance model must change before the technology strategy can execute.

Cross-Border Data Constraint

Attorney-client privilege, GDPR data residency requirements, and conflict-of-interest rules create jurisdictional data silos that no amount of technology can bridge without governance reform. In a verein structure, client data from the London office may be legally inaccessible to an AI system operated by the New York office. Even within a unified firm, cross-border data use for AI training requires jurisdiction-by-jurisdiction legal analysis. Professional services firms face data governance constraints that technology companies, manufacturers, and financial institutions do not—because the data itself is the client relationship, and the client relationship is the asset that generates partner compensation.

The Structural Contradiction

AI Requires Unification. Partnerships Require Autonomy.

The contradiction can be stated precisely. AI at enterprise scale requires three things: unified data (to train and fine-tune models on the full breadth of firm expertise), shared infrastructure (to deploy, maintain, and govern AI systems consistently), and coordinated investment (to fund capabilities that benefit the firm as a whole, not individual practices in isolation). Partnership governance requires three different things: autonomy (partners control their practice economics), distributed authority (major decisions require collective approval), and practice-level accountability (compensation reflects individual contribution, not firm-level outcomes).

These requirements are not merely in tension. They are structurally opposed. Every AI deployment decision in a partnership is simultaneously a governance decision about who pays, who benefits, and who has authority over shared resources. When a firm deploys an AI-powered contract review system, the decision is not “which AI tool should we use?” The decision is: Who funds the implementation? How is the cost allocated across practices? Who governs the training data? Which partners must adopt the system? What happens to the associates whose work the system partially automates? And who has authority to make these decisions in an organization where authority is, by design, distributed?

McKinsey’s 2024 research on generative AI in professional services estimated that 20–30% of billable hours in legal, consulting, and accounting could be augmented or automated by current AI capabilities. But “could be” and “will be” are separated by the governance gap. The technology exists. The business case is compelling. The constraint is the organizational structure through which the investment decision must pass—and in partnerships, that structure was designed to preserve autonomy, not to enable coordinated investment.

The Thomson Reuters finding that 73% of firms experiment while only 12% deploy firm-wide reflects a governance outcome, not an adoption lag. Experimentation requires one partner’s initiative and a modest budget. Firm-wide deployment requires collective agreement on cost allocation, data governance, workflow redesign, and change management—precisely the coordinated decisions that partnership structures are designed to avoid.

The Research Question

If AI requires shared infrastructure and partnerships require distributed authority, what governance model allows both?

Where the Evidence Converges

Three Evidence Streams, One Governance Insight

No single research program connects partnership governance to AI deployment failure. Technology analysts study adoption curves. Governance scholars study partnership structures. Management consultants study organizational transformation. But when these three independent evidence streams are placed in sequence, a convergence pattern emerges that none of them state individually.

Stream 1: AI Deployment Data in Professional Services

Thomson Reuters’ 2024 Future of Professionals survey established the 73%/12% gap: near-universal experimentation, minimal firm-wide deployment. Gartner’s AI governance research identifies the absence of centralized decision-making authority as the primary barrier to enterprise AI scaling. The ALM/Am Law Technology Survey shows that legal technology spending averages 3–7% of firm revenue—substantially below the 8–12% technology-to-revenue ratio typical of corporate enterprises. The data shows that professional services firms invest less in technology, deploy less at scale, and have a wider gap between experimentation and operational use than any comparable industry segment.

Stream 2: Partnership Governance Research

The academic literature on partnership governance, spanning Oxford’s Saïd Business School research on professional service firms to Gilson and Mnookin’s foundational work on law firm economics at Harvard Law Review, identifies the same structural features this brief describes: distributed authority, compensation tied to individual performance, and partner mobility as the enforcement mechanism. McKinsey’s Professional Services practice reports that 62% of partners in surveyed firms cite governance complexity as the primary barrier to AI investment—not technology readiness, not cost, not talent availability. The governance literature explains WHY the deployment gap exists: partnership structures distribute authority in ways that prevent the coordinated investment AI requires.

Stream 3: Firm Restructuring Evidence

Baker McKenzie’s 2024 verein dissolution is the most significant governance restructuring in global professional services in a decade. The firm explicitly cited technology investment coordination as a driver. Deloitte’s separation from its consulting practice (completed 2025 in several jurisdictions), PwC’s ongoing structural consolidation, and EY’s abandoned but instructive “Project Everest” split all reflect the same underlying dynamic: firms are restructuring governance to enable coordinated technology investment. The restructuring evidence shows WHAT changes when governance unifies: the structural barrier to coordinated AI investment is removed, but new governance challenges emerge around centralized authority in a culture built on partner autonomy.

The convergence insight: The 12% firm-wide deployment rate is primarily a governance outcome, not a technology limitation. Partnership structures distribute authority in ways that prevent coordinated AI investment. Firms that unify their profit pool remove the structural barrier—but create new governance challenges around centralized technology authority in a culture built on partner autonomy. The question is not “which AI to deploy” but “who authorizes shared investment when no one is structurally obligated to share.”

Exhibit 2

Three independent research streams converge on a governance insight that none of them articulate individually

Three Evidence Streams: Partnership Governance and AI Deployment STREAM 1 AI Deployment Data 73% / 12% experiment vs. deploy STREAM 2 Partnership Governance 62% cite governance as top AI barrier STREAM 3 Firm Restructuring Baker McKenzie verein dissolved 2024 CONVERGENCE INSIGHT The 12% firm-wide deployment rate is a governance outcome, not a technology limitation. Partnership structures prevent coordinated AI investment. Firms that unify profit pools remove the structural barrier—but create new challenges: centralized technology authority in a culture of partner autonomy.
Source: RBD. synthesis of Thomson Reuters AI deployment data, Oxford/Harvard partnership governance research, and Baker McKenzie restructuring evidence.
Emerging Patterns

Five Governance Models for Partnership AI Investment

Synthesizing the deployment data with governance research and firm restructuring evidence, five distinct models emerge for how partnerships are attempting to resolve the structural contradiction between AI’s requirements and partnership governance. Each model reflects a different trade-off between investment coordination and partner autonomy.

Model 01
The Practice-Led Model
Each Practice Funds Its Own AI
Individual practices or partner groups fund AI experimentation from their own budgets. Adoption is fast in high-revenue, high-margin practices (M&A, capital markets, restructuring) and effectively zero in lower-margin or volume-driven practice groups. The result is internal technology inequality: the same firm operates at two different capability levels depending on practice profitability. No shared data, no shared infrastructure, no scale economics. This is the default model in most firms today and the primary reason the 73%/12% gap exists.
Model 02
The Firm Assessment Model
Mandatory Technology Levy Across Partners
The firm levies a mandatory technology assessment on all equity partners, creating a shared investment pool for AI infrastructure. Enables enterprise-scale deployment but faces persistent partner resistance on two fronts: ROI attribution (which practices benefit?) and cost fairness (why should tax partners subsidize litigation AI?). The ALM data shows firms that adopt this model typically set the assessment at 2–4% of distributable profit—enough to fund infrastructure, insufficient for the practice-level workflow redesign that adoption requires.
Model 03
The Hybrid Investment Model
Firm Funds Infrastructure, Practices Fund Use Cases
The firm funds shared AI infrastructure (platforms, data governance, security) through the technology budget, while individual practices fund practice-specific use cases from their own P&L. This emerging consensus model resolves the fairness objection—shared costs fund shared capabilities—but creates coordination complexity. Who decides which infrastructure capabilities to build? How are priorities set across practices with competing needs? The governance challenge shifts from “who pays?” to “who decides?”
Model 04
The Client-Funded Model
AI Investment Embedded in Client Billing
AI development costs are embedded in client engagements, either as explicit technology fees or as components of value-based billing arrangements. This model works for bespoke, client-specific AI deployments—a custom due diligence tool for a major M&A transaction, for example—but fails for shared capabilities that benefit the firm broadly. Client-funded AI also raises questions about intellectual property ownership: if the client paid for an AI capability developed during their engagement, who owns the resulting model? Gartner identifies IP governance as a top-three concern for professional services firms deploying client-funded AI.
Model 05
The Unified Authority Model
Centralized Technology Office with Mandate Authority
A firm-level Chief Technology Officer or Chief AI Officer is given mandate authority over AI investment, deployment, and governance—authority that does not require practice-level partner approval for each initiative. This model requires governance reform: typically, profit pool unification and amendment of the partnership agreement to grant technology investment authority to a central function. It produces the highest capability and the highest cultural resistance. Baker McKenzie’s restructuring is a prerequisite for this model; the verein dissolution created the unified governance structure that makes centralized technology authority possible.
Horizon

When This Unfolds

Phase 1 · Now – Q4 2026
The Experimentation Plateau
The 73% experimentation rate has plateaued. Most firms have completed pilot programs, proof-of-concept deployments, and innovation committee assessments. The conversation shifts from “should we experiment with AI?” to “who authorizes the investment required to move beyond experimentation?” Governance conversations begin as AI costs scale from six-figure pilots to seven-figure enterprise deployments. Firms that have not resolved the investment authority question begin to feel competitive pressure from the 12% that have.
Phase 2 · 2027
The Governance Reckoning
Firms that have not resolved the governance question begin losing talent to firms that have. Associates choosing between two offers will increasingly select the firm with AI infrastructure—not as a technology preference, but as a work quality indicator. Lateral partners evaluating moves will factor AI capability into their assessment of a firm’s competitiveness. Client RFPs begin requiring evidence of AI-enabled service delivery. Talent arbitrage and client demand drive governance reform faster than any internal technology strategy initiative could.
Phase 3 · 2028+
The Structural Separation
Firms with unified governance structures deploy AI at enterprise scale. Federated firms remain stuck at practice-level experimentation. The competitive divergence becomes measurable in client satisfaction scores, associate retention rates, revenue per lawyer, and profit per equity partner. The separation is not between firms that chose the right AI vendor and those that did not. It is between firms that reformed their governance to enable coordinated investment and those that preserved partner autonomy at the cost of technological capability.
External Factors

Catalysts and Barriers

Catalysts
Client Pressure for AI-Enabled Services
Corporate general counsel and CFOs are requiring AI-enabled delivery in RFPs. Gartner reports that 45% of large enterprise legal departments have included AI capability requirements in law firm selection criteria since 2024. Client demand creates external pressure that bypasses the internal governance impasse.
Associate and Talent Competition
Top-quartile law school graduates increasingly evaluate firms on technology infrastructure. NALP data shows that technology capabilities rank among the top five factors in associate firm selection for the first time in 2025. Firms without AI infrastructure face a talent disadvantage that compounds annually.
Profit Pool Unification Trend
Baker McKenzie’s verein dissolution is the leading indicator. Multiple Am Law 100 firms are reportedly evaluating structural consolidation to enable coordinated investment. The Big Four accounting firms are simultaneously restructuring partnership economics to accommodate technology-intensive service delivery.
Barriers
Partner Compensation Models
Compensation structures that reward individual origination and billing performance create direct incentive misalignment with shared AI investment. The Am Law 100 average compensation-to-revenue ratio of 35–45% leaves limited margin for technology levies without compressing partner income—the one outcome that reliably triggers lateral departures.
Lateral Partner Mobility
Partners can leave. This is the enforcement mechanism that caps shared investment. If a firm’s AI assessment reduces a partner’s effective compensation below market rate, the partner moves. High-performing partners are the most mobile and the most sensitive to cost allocation, creating a structural paradox: the partners whose adoption would generate the most AI value are the ones most likely to resist the cost.
Jurisdictional Data Regulation
GDPR, attorney-client privilege, and conflict-of-interest rules create jurisdictional data silos that are legal constraints, not technical ones. Cross-border AI deployment in global partnerships requires jurisdiction-by-jurisdiction legal analysis. No technology solution resolves a governance problem rooted in regulatory compliance.
Implications

Four Priorities for Partnership Technology Leaders

01
Separate the governance decision from the technology decision
The most common error in partnership AI strategy is to begin with technology selection. Which AI vendor? Which use cases? Which practice goes first? These are second-order questions. The first-order question is: who has authority to authorize shared AI investment, and through what governance mechanism? Until the investment authority question is resolved, technology selection is academic. Firms that spend 18 months evaluating AI platforms without resolving who can approve a firm-wide deployment have optimized the wrong variable. Resolve governance first. Technology follows.
02
Design AI investment governance around existing compensation structures
The surest way to activate partner resistance is to propose an AI investment model that reduces individual partner distributions without a direct, attributable return. Effective partnership AI governance maps investment obligations to the compensation structures partners already understand. If the firm uses a modified lockstep system, the AI assessment should be proportional to lockstep position. If the firm uses an origination-driven model, AI investment should be embedded in origination credits. Do not work against the incentive system. Design around it.
03
Treat data governance reform as a prerequisite, not a parallel workstream
In corporate enterprises, data governance can be reformed alongside AI deployment. In partnerships—especially global partnerships with multiple jurisdictions—data governance reform must precede AI deployment. Client data ownership, cross-border data movement, privilege implications of AI-assisted legal work, and conflict-of-interest considerations are governance questions that must be resolved at the partnership agreement level before any enterprise AI system can ingest firm data. Firms that deploy AI first and resolve data governance later will discover that the data governance constraints invalidate the deployment architecture.
04
Use client demand as the forcing function for governance reform
Internal advocacy for AI investment faces the collective action problem described throughout this brief. External client demand bypasses it. When a top-10 client requires AI-enabled delivery as a condition of engagement renewal, the governance conversation changes. Partner resistance decreases when the alternative is client attrition. Technology leaders in partnerships should systematically document client AI requirements, quantify the revenue at risk from AI capability gaps, and present the governance reform as a client retention imperative, not a technology aspiration.
Decision Support

Partnership AI Governance Readiness Diagnostic

The following diagnostic helps technology leaders in partnership-structured firms assess whether their governance model is prepared to support enterprise AI deployment. It is organized around the Four Capability Bands from The Intelligence Organization™, applied to the specific challenge of AI investment governance in distributed-authority organizations.

Exhibit 3

Technology leaders can assess partnership AI governance readiness by scoring across four structural dimensions

Capability Band Governance Readiness Question Score 1–5 If Score <3
Band 1: Right-Fit Technology Is there a defined governance mechanism for approving firm-wide AI investment that does not require a full partner vote for each initiative? ___ Establish a technology investment committee with delegated authority below a defined threshold. Partner votes for every AI initiative create a structural veto that prevents enterprise deployment.
Has the firm completed a jurisdiction-by-jurisdiction data governance analysis that identifies which client data can legally be used for AI training and deployment? ___ Commission a cross-jurisdictional data governance review before any enterprise AI system ingests client data. Privilege, GDPR, and conflict-of-interest rules vary by jurisdiction and cannot be assumed.
Band 2: People & Purpose Is AI investment cost allocation mapped to the existing partner compensation structure in a way that partners perceive as fair relative to benefit received? ___ Redesign the AI cost allocation model to mirror the firm’s compensation logic. If partners cannot see the connection between what they pay and what they receive, resistance is rational, not irrational.
Has the firm defined how AI deployment will affect associate staffing ratios, partner-to-associate billing dynamics, and the career development model? ___ AI that automates associate-level work without a corresponding workforce model adjustment will face adoption resistance from partners who depend on associate staffing ratios for profitability.
Band 3: Operational Integration Are AI deployment plans structured as practice-level workflow redesigns with measurable impact metrics, or as technology rollouts with adoption targets? ___ Reframe AI deployment as workflow transformation, not technology adoption. Partners respond to practice economics, not technology features. Define success in terms of realization rates, cycle times, and client satisfaction.
Does the firm have a mechanism for sharing AI-generated insights across practices without violating client confidentiality or conflict rules? ___ Cross-practice AI value requires cross-practice data governance. Without a mechanism for ethical knowledge sharing, AI capabilities remain siloed in the same practice structure that created the governance problem.
Band 4: Adaptive Governance Is there a single person or body with authority to make binding AI governance decisions (data use, model deployment, ethical guidelines) without requiring consensus of all equity partners? ___ Appoint or create a technology governance authority with defined scope and delegated decision rights. Consensus governance and enterprise AI deployment are structurally incompatible.
Has the partnership agreement been reviewed and, if necessary, amended to accommodate centralized technology investment authority and cross-practice data governance? ___ If the partnership agreement does not contemplate centralized technology investment, any AI strategy that requires it operates on informal authority that can be challenged or revoked by partner vote.
Source: RBD. analysis. Framework aligned with The Intelligence Organization, Four Capability Bands (Starkey, 2026).

Interpreting the Score

Total 32–40: Unified Authority readiness. The firm’s governance structure can support enterprise AI deployment. The priority shifts from governance reform to execution: technology selection, workflow redesign, and change management. Focus on the Hybrid Investment Model or Unified Authority Model.

Total 20–31: Hybrid readiness. Some governance mechanisms exist, but gaps remain in cost allocation fairness, data governance, or decision authority. Identify the lowest-scoring Band and address it before scaling AI beyond pilot programs. The Firm Assessment Model or Hybrid Investment Model is the appropriate governance target.

Total 8–19: Practice-Led default. The firm lacks the governance infrastructure to deploy AI beyond individual practice initiatives. Experimentation will continue; enterprise deployment will not occur until governance reform addresses investment authority and data governance. Prioritize Band 4 (Adaptive Governance) immediately—without centralized decision authority, all other capability development is structurally blocked.

Federated Structure Application

Firms operating under verein or other federated structures face an additional diagnostic layer. The questions above assume a single legal entity. In a federated structure, each question must be answered separately for each member firm—and then a meta-governance question must be addressed: is there a mechanism for coordinating AI governance across member firms? If the answer is no, the firm faces the Baker McKenzie dilemma: the technology strategy requires governance unification, and governance unification requires structural reform that may take years to negotiate and execute.

Decision support aligned with The Intelligence Organization · Band 4 (Adaptive Governance) as the primary lever for partnership AI readiness · Governance authority as the structural determinant of AI deployment capability

AI governance in partnerships is a structural design problem, not a technology procurement decision.

This research informs our partnership governance advisory practice. If your firm is navigating the transition from experimentation to enterprise AI deployment, we should talk.

Schedule a Conversation
Sources

References

AI Deployment in Professional Services: Thomson Reuters, "Future of Professionals: Harnessing AI to Reimagine Professional Services," 2024, surveying 1,200+ professionals across legal, tax, accounting, and compliance, finding 73% experimentation and 12% firm-wide deployment rates. Gartner, "AI Governance Frameworks for Enterprise Deployment," 2024, identifying centralized decision-making authority as the primary barrier to enterprise AI scaling. McKinsey & Company, "The Economic Potential of Generative AI: The Next Productivity Frontier," 2024, estimating 20–30% of professional services billable hours augmentable by current AI capabilities. McKinsey & Company, "Generative AI and the Future of Work in Professional Services," 2024, surveying partner attitudes toward AI governance.

Legal Market and Technology Spending: Thomson Reuters, "2024 Report on the State of the Legal Market," Georgetown Law Center for the Study of the Legal Profession and Thomson Reuters Institute, reporting $1.1T+ global legal services market. ALM/Am Law, "Technology Survey and Am Law 100/200 Financial Rankings," 2024, documenting 3–7% technology-to-revenue spending ratios and lateral partner movement data. NALP (National Association for Law Placement), associate survey data on firm selection factors, 2025.

Partnership Governance Research: Gilson, R.J. and Mnookin, R.H., "Sharing Among the Human Capitalists: An Economic Inquiry into the Corporate Law Firm and How Partners Split Profits," Stanford Law Review, 37(2), 1985. Empson, L., "Leading Professionals: Power, Politics, and Prima Donnas," Oxford University Press, 2017, examining governance challenges in partnership-structured professional services. Saïd Business School, University of Oxford, research program on professional service firm governance and organizational design.

Firm Restructuring Evidence: Baker McKenzie, announcement of verein dissolution and global partnership unification, 2024, covering 4,700 lawyers across 74 offices in 45 countries. Reuters, "Baker McKenzie Votes to Unify Global Partnership," 2024. The American Lawyer, coverage of Big Four structural consolidation and professional services governance reform, 2024–2025. Ernst & Young, "Project Everest" consulting/audit separation analysis, 2023 (abandoned).

RBD. Research: Starkey, M.C., The Intelligence Organization, 2026. RBD., "Enterprise AI Investment 2026 Outlook: From Technology-First Budgets to Capability-First Returns," RB-AI, Q2 2026. RBD. cross-industry partnership governance and AI deployment synthesis, 2026.