AI will force law firms, management consultancies, and accounting firms to change
AI will create a barbell effect in professional services firms: commoditization of low-skill services, while revenue and income will accrue to higher-value services
After I posted my recent piece AI will not kill McKinsey, a reader requested that I explore how AI will affect professional services firms; this post is that exploration. By “professional services firms,” I am referring to law firms, management consultancies, and accounting & audit firms. This is US-biased but my guess is that it maps well to other advanced economies.
AI does not simply automate bills by the hour. It attacks, and sometimes amplifies, the three economic pillars that make professional services firms valuable:
Information asymmetry (the expert knows more than the client)
Risk transfer (the expert carries reputational or legal liability for advice), and
Coordination leverage (the partnership marshals rare talent at scale).
As large language models, predictive analytics, and agents mature, they shrink pillar 1, complicate pillar 2, and magnify pillar 3. The net effect is a barbell: low complexity, repeatable work becomes software; ultra-complex, high-stakes work becomes even more partner-driven, and more profitable because AI lets a leaner senior team orchestrate larger problem spaces. Below is a first principles tour of how that dynamic plays out for law firms, management consultancies, and accounting firms, followed by practice-level forecasts and contrarian opportunitites.
First principles decomposition
AI’s comparative advantage sits in the left half of the barbell; the right half (judgment, persuasion, liability) remains human-centered but now enjoys higher leverage.
This implies that costs will shift around:
Fixed costs decrease: research associates, paralegals, junior consultants, and audit staff become software seats.
Capital expenditure increases: secure private model hosting, synthetic data pipelines, indemnification insurance, and specialized GPUs.
Marginal cost per matter declines, but variance increases: hallucinations and model drift create tail risk events that require senior partner intervention, pushing firms toward outcome-based or risk-sharing fee structures.
Sector-specific responses
Law firms
Document-heavy practices commoditize first: e-discovery, contract review, due diligence checklists. Many AM Law firms are already bundling these as embedded AI paralegal offerings and pricing them with fixed + success fee combos.
Litigation and regulatory advisory bifurcate: routine motions drafted by GPT-style models; bespoke oral advocacy becomes more valuable because judges and regulators distrust machine output.
Ethics and privilege: secure on-prem LLMs or vendor models hosted under strict SOC-2 / ISO 27001 regimes; partner sign off required to preserve attorney-client privilege on AI-generated drafts.
Liability hedge: emergence of AI malpractice riders in professional indemnity insurance; some firms spin out captive insurtechs.
Management consultancies
Deck factory automation: slide narration, data visualization, and spreadsheet sensitivity (“what if?”) scenarios are generated in minutes. This erodes the billable hours of associates, pushing McKinsey-type firms toward:
Subscription dashboards (continuous insight vs episodic reports)
IP licensing of proprietary industry models, and
Asset-backed consulting (e.g., digital twin of a supply chain delivered as SaaS)
Field experimentation at scale: Reinforcement learning agents can run thousands of A/B tests inside synthetic markets before pilots hit the real world, allowing consultants to offer confidence intervals on operational plans and share upside.
Talent model: fewer generalist MBAs, more ML engineers and industry operators embedded in client teams; senior partners focus on C-suite politics and cross-sell.
Accountancy
Continuous audit: AI agents ingest real time ERP streams, flag anomalies, and refresh fair value estimates daily instead of annually. Big 4 move from annual opinions to audit-as-a-service subscriptions.
Tax structuring: generative code writes jurisdiction-specific filings; strategic value shifts to international risk modeling and regulator negotiation.
RegTech arbitrage: firms that master synthetic data for SOX or ESG reporting sell assurance APIs that smaller CPA firms consume under white label deals.
Regulatory moat grows: because AI increases the speed of fraud, regulators tighten audit requirements; only large firms can afford the compliance tech stack, reinforcing Big 4 dominance unless antitrust forces break them up.
Practice-level forecasts (2025-2035)
Contrarian opportunities and risks
New revenue pools
Model validation and red-teaming: Firms certify clients’ AI for bias, privacy, and safety: the legal opinion letter of the LLM era.
Synthetic precedent libraries: Curated corpora of adjudicated, bias-screened model outputs licensed to in-house counsel or controllers.
Outcome bank financing: Shared savings or litigation funding vehicles where the firm takes equity-like positions, priced via AI risk models.
Existential threats
Disintermediation by platform companies: If Microsoft or Thomson Reuters bundle best-in-class AI copilots directly into MS 365 or Westlaw, mid-sized firms lose differentiation.
Liability cascade: High profile AI drafting error triggers plaintiff bar; insurers jack up premiums; small partnerships exit market.
Regulatory squeeze: Bar associations and PCAOB may limit AI usage to preserve public trust, raising compliance costs and delaying ROI.
Strategic plays
Own the data: secure client-side private moat datasets that fine tune models uniquely.
Productize judgment: embed partner heuristics into decision support sofrware and sell it separately from labor hours.
Portfolio talent: retain a small bench of polymath partners, contract specialists on demand, and orchestrate with AI project managers.
Practical steps firms are taking today
Implications for clients and professionals
Clients will demand explainability certificates and indemnities for AI-assisted work; they will also expect faster turnarounds and variable pricing.
Young professionals face a skill inversion: rote analytical work is automated, so career paths skip straight to higher-order judgment, which makes mentorship and tacit knowledge transfer a bottleneck.
Partners who master AI-augmented persuasion (courtroom storytelling, board facilitation, regulator negotiation) will command outsized fees, because the relative scarcity of trusted authority rises as raw information becomes abundant.
Bottom line
AI hollows out the middle of professional services while expanding both ends: fully automated solutions at commodity prices and elite human judgment command premium fees. Firms that productize their intellect, own privileged data, and shoulder novel forms of risk will thrive; those that cling to the billable hour pyramid will watch the pyramid invert.
What follows are some third-party resources I relied on to synthesize everything above.
Law firms
Top tier firms are productizing routine work with LLMs: Allen & Overy provides access to legal AI tech Harvey to 3,500 lawyers.
Legal tech consolidation is accelerating as incumbents buy AI IP: Thomson Reuters acquires Casetext.
E-discovery and first pass review already run on generative AI: Relativity’s aiR generative review product line.
Bar regulators now require AI competence and confidentiality controls: ABA issued an opinion about lawyers’ use of generative AI.
Insurers are writing, or excluding, coverage for AI drafting errors: ABA Journal article about professional libaility policies for AI mistakes.
Judges are sanctioning hallucination-based filings, raising tail risk: Lawyers have been fined for submitting AI-generated hallucinations (fake precedents) to court.
Firms are standing up AI governance practices: Clifford Chance’s dedicated AI & Tech risk management practice.
Management consultancies
Consultancies partner with model labs to embed GenAI in client work: Bain & Co’s partnership with OpenAI.
Deck factory tasks are being automated by in-house agents: McKinsey has developed a tool, dubbed Lilli, that searches firm knowledge and drafts analyses in seconds.
AI is shrinking analyst staff while boosting outcome-based deals: McKinsey cut its headcount and now uses thousands of AI agents.
Consultancies are repositioning as implementation/IP platforms: BCG’s research on scaling AI value and needing new business models.
Outcome-based pricing frameworks are diffusing across B2B services: Paper on the rise of outcome-based pricing in SaaS and AI-enabled products.
Accounting firms
Big 4 have deployed AI-first audit platforms:
Regulators are rewriting standards to accommodate AI-assisted testing: PCAOB statement urging guidance on AI and data analytics in audits.
Continuous-audit vision (streaming EPR data, anomaly flags) is mainstream: KPMG’s description of its tool Clara performing real-time client collaboration.
Together, these sources confirm that the forces I outline above—commoditization of low-complexity work, concentration of high-stakes judgment, and a surge in governance and risk services—are not speculative; they are already visible in firm press releases, regulator dockets, and market transactions.
If you enjoy this newsletter, consider sharing it with a colleague.
Most posts are public. Some are paywalled.
I’m always happy to receive comments, questions, and pushback. If you want to connect with me directly, you can:
> “the relative scarcity of trusted authority rises as raw information becomes abundant.”
This might end up being one of the most important themes of the next five years.