Gurpreet S. Bal has spent enough time in the market to be skeptical of both the doomsayers and the boosters. His take on AI in law is grounded rather than breathless: "Clients have every right to ask whether AI is reducing their bill. If the answer is no — if your lawyers are using tools that cut their time in half and billing you the same rate — that conversation is overdue." AI is transforming legal practice. It is doing so unevenly, and the clients best positioned to benefit are the ones who ask direct questions.
Bal advises on AI company acquisitions, technology transactions, and the legal infrastructure that underlies some of the most complex deals in the market. He thinks about legal AI from both sides of the equation.
The legal AI transformation that is happening today — not speculatively, but measurably — is concentrated in two areas: due diligence and contract review. In M&A transactions, AI-assisted diligence tools can process thousands of documents, extract defined terms, flag deviations from market standards, and surface issues that would previously have required dozens of associate hours. In contract management, AI review tools have made first-pass analysis of large commercial agreement portfolios a matter of hours rather than weeks. These are genuine productivity gains. They compress timelines, reduce the cost of volume work, and free senior lawyers to spend time on the judgment-intensive questions that actually matter. Law firms that have deployed these tools at scale are doing real diligence faster and more thoroughly than firms that have not. The question of whether that productivity gain flows to clients or stays in the firm's margin is a separate — and important — conversation.
The next wave of legal AI is autonomous execution — AI systems that do not just analyze but act. In transactional practice, this looks like AI agents that can manage the mechanics of a closing: tracking conditions, circulating signature pages, updating data rooms, sending notices. In compliance, it looks like regulatory monitoring agents that continuously scan for new rules, flag applicability to the client's business, and draft response protocols. These applications are not science fiction — they are in early deployment at forward-leaning firms and legal operations teams. The governance frameworks for these deployments are lagging behind the technology. Who is responsible when an autonomous agent misses a closing condition? When a compliance bot misclassifies a regulatory change? The accountability structure for AI-driven legal work is still being designed, largely by the firms and clients deploying these tools and largely without external regulatory guidance.
AI is accelerating the commoditization of routine legal work. Standard NDA review, boilerplate employment agreements, form equity documents, routine regulatory filings — these are tasks that AI handles capably and that clients increasingly expect to be priced accordingly. The associates at large firms who built their early careers on this work are facing a market that values it less. That is a real economic disruption, and the profession is not being honest enough about it. What is not being commoditized — and what cannot be commoditized in any near-term version of this technology — is judgment on novel questions. The AI tools available today are extraordinarily good at applying established patterns to new facts. They are not capable of reasoning through a genuinely novel legal question, advising on strategy in a high-stakes negotiation, or navigating the relationship dynamics that determine whether a deal closes. The lawyers who will thrive in this market are the ones who are ruthlessly honest about which category their work falls into.
Three things will remain the domain of human lawyers for the foreseeable future. First: fiduciary judgment. The decision to advise a client to walk away from a deal, to settle a case on unfavorable terms for strategic reasons, to disclose a problem that the client would prefer to ignore — these require a lawyer willing to be accountable for their advice and able to hold a fiduciary duty with legal and ethical force. No AI system is a fiduciary. Second: novel legal questions. When the law is genuinely unclear — when the right answer requires predicting how courts will respond to a new fact pattern, how regulators will exercise discretion, how a contract provision will be construed in circumstances the parties never contemplated — the answer requires human reasoning about uncertainty and human accountability for being wrong. Third: relationship and trust. Sophisticated clients retain the lawyers they trust. That trust is built through demonstrated judgment, through shared experience in difficult situations, through the confidence that your lawyer's interests are genuinely aligned with yours. AI cannot earn trust. It can only produce outputs.
Clients should be asking their lawyers — specifically, not generally — how AI tools are affecting the cost and timeline of their matters. Ask: which tasks on my matter have been accelerated by AI? How is that reflected in the fee? If the answer is that AI is not being used, ask why. If the answer is that AI is being used but the fees have not changed, ask for an explanation. The legal market is beginning to price AI productivity gains into flat fees, capped fees, and alternative fee arrangements. Clients who negotiate those structures are capturing the benefit. Clients who accept hourly billing without asking these questions are subsidizing the firm's transition to a more efficient model without sharing in the upside. This is not a hostile question. It is a basic conversation that sophisticated clients are already having, and that every client should feel empowered to raise.
Gurpreet S. Bal is a corporate partner with 16 years advising on private equity, merger transactions, and public offerings for companies and investors at three of the world's top law firms. He has represented clients in hundreds of transactions with aggregate deal value exceeding $60 billion across AI, semiconductors, fintech, and emerging technology. For more information and to get in touch, visit gurpreetbal.com.