Key insights:

• • • Cloud modernization created accumulation, not transformation听鈥� Many enterprises scaled infrastructure faster than they integrated systems, leaving fragmented data, duplicated processes, rising costs, and weak links between IT investment and business value.

    • AI and regulation now expose weak data architecture听鈥� Agentic AI, real-time decision making, and regulatory reporting depend on consistent, traceable, well-governed data 鈥� not fragmented systems or after-the-fact governance.

    • Enterprise architecture must be rebuilt around outcomes and economics听鈥� Instead of treating the cloud as the strategy, organizations should define business outcomes first, structure data as a reusable asset, and measure architecture by revenue, cost efficiency, regulatory accuracy, and decision speed.

In this two-part blog series about the current state of cloud architecture, we look into where this architecture has failed and, in the next part of the series, what the possible remedies might be.

For the better part of the last 15 years, IT enterprise architecture definition and management didn鈥檛 disappear, it was deprioritized and replaced by as-a-Service solutions. The rapid rise of cloud platforms such as Amazon Web Services, Microsoft Azure, Snowflake, and Google Cloud made it possible to stand up infrastructure, deploy applications, create advanced databases, and scale environments without the same level of architectural rigor that was once required. Speed replaced structure, and access replaced integration.

Today鈥檚 cloud realities

The problem is that what was built during this last technological explosion was not architecture 鈥� it was accumulation. Systems expanded, data proliferated, budgets exploded, and organizations convinced themselves that connectivity was the same as coherence, and that data replication was the same as the system-of-record.

These assumptions have now been exposed as false positives. Over the last three years, AI, real-time decision making, and regulatory transparency have fundamentally changed the requirements. These are not technologies that sit on top of fragmented environments, they are data-driven capabilities and outcomes that depend on precision, integration, and sequence. The arrival of agentic AI and its stringent objective-based principles cannot tolerate data ambiguity and fragmented architectural designs.

The problem is that what was built during this last technological explosion was not architecture 鈥� it was accumulation.

AI does not fail at rates exceeding 80% because models are weak, it fails because the underlying data is inconsistent, inaccessible, or economically misaligned. Regulatory frameworks do not struggle because rules are unclear, they struggle because data cannot be traced, reconciled, or produced in real time. What the cloud enabled 鈥� rapid deployment without disciplined integration 鈥� is exactly what now constrains performance. The issue is no longer whether systems can scale, but whether they can produce measurable, consistent, and adaptable outcomes.

This is where enterprise architecture returns, but just not in its previous form. The discipline cannot simply revert to academic frameworks and abstractions that were designed for a different software era. It must be rebuilt around a different sequence that sees business outcomes first, data second, and then systems engineered within those constraints. Today, enterprise architecture must be defined and managed by economic KPIs, value added, and its adaptability to rapidly changing business realities.

Where the model broke

The failures experienced today in cloud architecture are not singularly technological. Cloud platforms deliver exactly what they promise 鈥� scalable, resilient, highly available infrastructure. Rather, the failure is architectural, and more precisely, involves the economics of compartmentalized capabilities.

Enterprise value is not created at the infrastructure layer. It is created where data informs decisions, and decisions drive outcomes. By over-rotating toward infrastructure, organizations optimized the least differentiating component of the enterprise stack, while leaving the highest-value layers largely untouched.

The result is a structural imbalance in which data remains fragmented across domains, business logic continues to operate in silos, governance is applied inconsistently and often retroactively, and measurement frameworks fail to tie technology activity to financial performance.

In this model, the cloud amplifies existing conditions. If fragmentation exists, it scales fragmentation. If inefficiency exists, it scales inefficiency. Modern infrastructure, applied to legacy architecture, produces modernized dysfunction.

What makes the cloud鈥檚 illusion particularly persistent is that its failure is rarely framed in economic terms. Cloud investments are justified through technical metrics such as uptime, latency, migration progress, and consumption efficiency. And while these are necessary, they are not sufficient. They do not answer the only question that ultimately matters: 鈥�Did the investment improve the economics of the business?鈥�

Enterprise value is not created at the infrastructure layer 鈥� it’s created where data informs decisions, and decisions drive outcomes.

In many cases, the answer is no 鈥� at least, not in a way that can be clearly articulated. Instead, organizations experience cost expansion without proportional productivity gains, increased data duplication that drive storage and processing inefficiencies, extended timelines for analytics and reporting despite real-time capabilities, and persistent manual intervention in regulatory and operational workflows.

The absence of a direct line between architecture and outcome creates a vacuum often filled with disconnected KPIs, measurement solutions, and most recently, AI-automation. And with this interoperable vacuum, activity and speed have been mistaken for progress.

Figure 1: Cloud accumulation meets enterprise architecture shifts

The data reality beneath the surface

The cloud did not fail to deliver transformation; rather it exposed why transformation had not occurred 鈥� and at the center of this exposure is data.

Most enterprises operate with data architectures that were never designed for interoperability, reuse, or regulatory-grade consistency. Definitions vary by function, pipelines are purpose-built and duplicative, and governance is layered on after the fact. Automation was designed using business rules, then software architectures, then what the data needed. Therein resides the structural disconnect for enterprise architecture in AI solutions: They are out of order.

When these legacy conditions are moved to the cloud, they do not improve, they accelerate. The organization gains speed without alignment, scale without standardization, and access without coherence. For regulated industries, this creates a compounding risk of inconsistent outputs across reporting channels, increased reconciliation overhead, reduced confidence in data lineage and auditability, and slower response to regulatory changes.

What appears to be a technology issue is, in fact, a failure of data design.

Reframing the problem

To move forward, the premise must change. The cloud is not the strategy; rather it鈥檚 the environment. Transformation does not occur when systems are moved, it occurs when the relationship between data, decisions, and outcomes is fundamentally redesigned.

This requires an organization-wide shift from infrastructure-led thinking to what is defined as value architecture. Simply put, value architecture includes data that is structured as a reusable, governed asset 鈥� not a byproduct of applications, and business outcomes that are defined upfront and used to drive architectural decisions. Its governance is embedded at the point of data creation and distribution, and it replaces redundancy by making reuse the primary scaling mechanism. Finally, measurement is tied directly to financial and operational impact.

This is not a rejection of the cloud; rather, it鈥檚 a repositioning of its role and value proposition.

The implication is both direct and unavoidable. If your current strategy cannot clearly articulate how technology investment improves the economics of your business, then your organization is operating within the cloud illusion. However, this is not a critique of past decisions. It is a recognition that the next phase of transformation requires a different operating model 鈥� one that explicitly connects architecture to economics. Moving forward, what was forgotten in the past is now a future core competency.

Most organizations using as-a-service software had assumed that the cloud provider, vendor, or combination of those dealt with the complex liabilities of making designs interoperable. The implication moving forward 鈥� as well explore more in the second installment of this series 鈥� is that service software architectures using the system ideation approaches within AI silos are failing miserably, and there are few who understand the designs and skills needed to guide enterprises in the future.

You can find more blog posts听by this author here

Key insights:

• • • AI is changing legal鈥檚 role, not just its workload 鈥� Going forward, AI will do more than automate routine tasks, it also will help in-house legal teams become more strategic business partners.

    • The 5 archetypes make the transformation concrete 鈥� There are five practical ways in which AI could reshape legal work, including automation, stronger advising, better collaboration, and global scale.

    • Every organization鈥檚 AI transformation will be different 鈥� 成人VR视频鈥� own legal transformation journey shows the common and unique aspects of this process.

Beyond the automation, productivity boosts, or the now-familiar promise of doing more with less, the question over how AI will really transform the work that corporate legal departments do on a daily basis, has yet to be truly answered.

To deepen our understanding of where in-house legal is really heading next, Norie Campbell, 成人VR视频 Chief Legal Officer, and Lizzy Duffy, a Senior Director of the 成人VR视频 Institute, produced a new feature article, The 2030 legal department: 5 ways AI will transform how in-house teams work听that steps back from the day-to-day noise around AI and asks the bigger, more interesting question: 鈥淲hat is the legal function actually becoming?鈥�

Importantly, the article recognizes that in-house legal teams are navigating real constraints around time, budget, and clarity even as expectations continue to evolve. It also acknowledges how GCs are balancing rising demands with a growing focus on efficiency, while also working to define what effective and meaningful AI adoption should look like for their teams.

Indeed, this human pressure is one of the most compelling aspects to the questions corporate law departments are facing today, and it reverberates beyond a simple theory of AI in legal to really reflect a profession at a turning point.

The five archetypes

The feature also lays out five archetypes 鈥� distinct models for how AI could reshape legal work, from high-volume automation to better strategic advising, stronger business partnering, smarter collaboration with outside counsel, and truly global leverage across teams and languages.

By referencing these five archetypes, legal department leaders can start asking where their own teams fit, and what they need to do to get better prepared for the AI-driven legal future of 2030.

These archetypes cover everything from deciding on the best ways to leverage AI-led automation to helping legal teams become more proactive strategic advisers. The archetypes also detail how to foster collaboration that can allow other corporate functions to act more confidently without constant legal intervention. And how to use AI to reduce barriers caused by language and time zones, enabling multinational legal teams to work more effectively across geographies.

By referencing these five archetypes, legal department leaders can start asking where their own teams fit, and what they need to do to get better prepared for the AI-driven legal future of 2030.

成人VR视频鈥� own journey

This feature article also builds a practical, grounded picture of the future from inside 成人VR视频鈥� own General Counsel鈥檚 Office (GCO), showing readers a transformation that鈥檚 already taking shape.

This insider perspective offers a front-row look at how one GCO is trying to move from experimentation to real transformation and tells a bigger story than technology alone. Today鈥檚 transformation of the corporate legal department is really about leadership, ambition, and the choices department leaders need to make now if they want to stay relevant by 2030.

More than anything, the feature article stresses that adopting AI tools is not the same as true transformation. To move beyond incremental gains, legal departments must redesign workflows, improve data infrastructure, invest in training, and hire for adaptability and technical literacy. Ultimately, the central message is that efficiency is only a by-product 鈥� the real challenge is deciding what kind of legal function an organization will need in 2030 and how to start building toward that vision now.

You can access the full feature article, The 2030 legal department: 5 ways AI will transform how in-house teams work here

Key highlights:

• • • Curriculum听redesign must start now 鈥� One law school鈥檚 approach illustrates the necessity of mapping the entire curriculum to identify which skills to preserve, evolve, or build from scratch.

    • Training faculty in AI use is critical 鈥� Faculty AI training should be a multi-layered approach including hands-on training with specialized legal AI tools, guidance on redesigning curricula, and more.

    • AI simulations may be the key 鈥� Law school leaders need to act now by experimenting with small pilot projects and building simulation-based learning tools to replace the developmental depth that once came naturally in the first years of practice.

The debate about AI consuming most of the work that teaches essential lawyering skills to junior attorneys is forcing a reckoning with the long-held assumption that law schools were never designed to produce practice-ready lawyers and that it was always the profession’s job.

Indeed, AI is forcing that uncomfortable truth into the open faster than anyone anticipated because essential lawyering work 鈥� the document review, contract markup, research memo creation 鈥� dictated how a junior lawyer learned to spot the issue buried on page 47, to sense when a clause was off, and to develop the instinct that no classroom can fully replicate. Now, as more law firms deploy AI to handle precisely those entry-level tasks, the organic training moments that used to define the first two to three years of legal practice are evaporating.

, Executive Dean, Faculty of Law at Bond University, and Co-Chair of the Council of Australian Law Deans, says he sees where this is leading. The ultimate results will be firms hiring fewer junior lawyers today because AI has taken over that entry-level work, James explains, adding that means there will simply be no pipeline of mid-level, experienced lawyers to draw from in three to five years. Indeed, this is a slow-moving crisis, already in motion, and yet to fully arrive.

This crisis lands at the center of what the AI and Future of Legal Practice (AIFLP) initiative exists to address because at the core of this crisis is what does being job-ready really means when the job itself is being redefined. Answering this question requires law schools, law firms, licensing bodies, and technologists to do something they have historically struggled to do 鈥� that is to think and act collaboratively.

Rethinking the curriculum before AI does it for you

leads IE Law School鈥檚 AI initiative and is steering the school鈥檚 efforts to embed AI across the curriculum. To do so effectively, her approach requires going back to a broader set of foundational questions in legal education such as: For what is legal education meant to prepare students? How do students learn to develop legal judgment? What makes legal advice genuinely valuable? And what skills are essential to deliver that value in an AI-enabled profession?

鈥淟ayering AI tools on top of an unchanged curriculum serves no one,鈥� Perez-Llorca explains, adding that without answers to the fundamental questions, 鈥測ou are just adding technology to a structure that was never designed to handle it.鈥�

Check out how one law school professor is building AI simulation tools

IE law school is currently mapping its entire curriculum to determine which skills need to be preserved, which need to evolve, and which need to be built from scratch, while also using the AI-boosted curriculum to train faculty. Perez-Llorca describes the school鈥檚 faculty AI training as a multi-layered approach encompassing university-wide LLM training, substantive AI law curriculum review, hands-on training with specialized legal AI tools, guidance on redesigning curricula, and assessments to reflect students’ growing AI proficiency. Before students can be taught with AI, professors need to understand the tools themselves and how to use them in teaching, in simulation, and in assessment, she adds.

An AI tutor that meets students where they are

Bond University鈥檚 James says he has spent the last several months building an AI tutor designed to walk students through course material the way a patient, attentive instructor would. His vision for the AI teaching assistant supports the professor meeting students where they are. 鈥淚t [the AI tutor] introduces the week’s topic, outlines learning outcomes, guides students through the readings, checks comprehension with short quizzes, and then adapts in real time based on how the student responds,鈥� James explains, adding that the AI tutor will pull any student who is struggling deeper into the material until the learning outcome is achieved. 鈥淭he conversation never stops until the learning does.鈥�

However, James is careful to draw a clear distinction about what the tutor replaces and what it does not, stressing that AI is a substitute for the lecture recording, the static reading list, or the passive video watched at midnight before an exam 鈥� but it chiefly exists to support the law professor. This approach frees up class time, turning it from content delivery to more meaningful the time between the human instructor and students, he adds.

Act by design or default

The approaches by both Perez-Llorca and James point to a way to address the question of disappearing tasks that teach essential lawyering skills as well as shift the center of gravity in legal education toward ways to foster developmental skills and legal judgment. Indeed, inertia is not a strategy, and law school deans and associate deans can be at the forefront of this fight by taking decisive action, including:

• • • Experiment freely 鈥� Investigate with AI on your own by starting small with a pilot project.
    • Strategically assign where AI goes 鈥� Decide where AI belongs in the curriculum, such as in courses focused on legal research and drafting as they become commoditized by AI. Also, determine in which instances AI does not belong, such as counseling clients through ambiguity, navigating ethical complexity, and advocating persuasively. Make sure these all remain led by human lawyers.
    • Focus on skills 鈥� Map your law school鈥檚 curriculum by identifying which skills need to be preserved, which skills need to evolve, and which need to be built from scratch.
    • Build AI-assisted teaching tools 鈥� Make experiential and simulation-based learning central to the curriculum.

鈥淭he choice is between dealing with this crisis by design or by default,鈥� James says, noting that the pipeline problem he described is already in motion while the practitioners, educators, technologists, and licensing bodies that need to solve this together are not yet consistently in the same room.

Watch our recent Clarity podcast to see

Key insights:

• • • Law firms have moved from restricting GenAI use (Don鈥檛 use tools that leak client data) to mandating it (Incorporate AI into your practice and market our firm鈥檚 GenAI capabilities)鈥斕齆either phase has given rank and file lawyers what they really need: Guidance on in which instances GenAI actually helps deliver better, cheaper, and faster legal services, where it introduces serious professional risk, and how to tell the difference.

    • GenAI鈥檚 capacity to transform legal work for the better is real, but so is its capacity to degrade it听鈥擥enAI can significantly boost speed and quality on tasks involving breadth, synthesis, or straightforward analysis, but it can weaken performance on complex judgment and revision tasks 鈥� especially for stronger professionals 鈥� by encouraging overconfidence, missed issues, and superficial reasoning.

    • A use-mode framework can close the听gap鈥� A proposed governance framework can give law firm leadership a practical tool for identifying in which situations GenAI enhances legal work, where it introduces serious risk, and where professional judgment is non-negotiable.

This article synthesizes findings from the author鈥檚 paper,

Your law firm undoubtedly has a policy around generative AI (GenAI), which probably tells lawyers to avoid tools that leak client data, admonishes them to look out for hallucinations, and encourages them to incorporate AI into their practice to satisfy client demands.

However, it likely does not tell them which cognitive functions they should delegate to GenAI, which they should not, and where the line between the two is absolute. In the space between restriction and mandate, lawyers are making consequential decisions about GenAI delegation every day. Meanwhile, most law firms have not addressed that space with meaningful governance.

GenAI can make legal work worse

GenAI鈥檚 capacity to transform legal work for the better is real, but so is its capacity to degrade it. Most law firm leaders know that AI can hallucinate; yet far fewer know that it can make expert legal judgment and work product actively worse.

The best evidence of this dynamic comes from a with consultants from the Boston Consulting Group, who were given similar tasks and allowed to use various levels of AI assistance, including no AI. For professional tasks requiring breadth and option generation, GenAI delivered, showing that output quality improved by 40% and consultants worked faster. For tasks requiring judgment and synthesis, however, something unexpected happened. Consultants using GenAI were 19% less likely to produce correct solutions than those working without it.

Governing GenAI鈥檚 uneven performance requires asking a question that most law firms are not asking: What cognitive function is being delegated to GenAI at each step in the workflow?

The same pattern appears in research evaluating GenAI use in legal analysis. An empirical in the Journal of Legal Education confirmed that AI dramatically improves performance on straightforward analysis while producing no measurable benefit for complex reasoning. And in the case of complex reasoning, GenAI use also introduced recurring failures, such as jumping to conclusions, missing less obvious issues, and generating confident prose that masks superficial analysis.

from the University of Minnesota focused on legal tasks showed that GenAI assistance on a synthesis task improved performance by nearly 60% and produced a surprising downstream benefit. Those participants who used AI for synthesis outperformed the control group on the subsequent independent reasoning task even after GenAI was removed. However, when GenAI was introduced at the revision stage, the picture changed. GenAI helped weaker performers, but it actively degraded the work of stronger ones. Indeed, the best lawyers in the study produced worse revised work product when they used GenAI than when they worked without it.

A use-mode governance framework

Given all these findings, governing GenAI鈥檚 uneven performance requires asking a question that most law firms are not asking. Instead of determining whether GenAI is appropriate for a particular deliverable 鈥� such as a brief, a contract, or a board presentation 鈥� the governance question instead should be: What cognitive function is being delegated to GenAI at each step in the workflow?

My proposed framework, outlined below, organizes common GenAI uses into seven recurring modes following the sequence in which lawyers actually use GenAI to produce legal work product. Then, governance controls are calibrated to the risk profile of each mode.

Modes 1 and 2: Retrieval and organization

At the mechanical end of the cognitive spectrum are two distinct functions. In retrieval mode (Mode 1), a lawyer reviewing a merger agreement asks GenAI to identify every representation and warranty in the document. In organization mode (Mode 2), a litigator reviewing 50 depositions asks GenAI to construct a timeline from the testimony. The first locates material that already exists. The second arranges it into a usable structure. No new content is created in either case, and both uses are low-risk and should be actively encouraged, subject to modest verification controls. Firms that unduly restrict these use modes are leaving value on the table.

Mode 3: Summarization

Summarization (Mode 3) introduces selection risk. In this mode, GenAI chooses what to emphasize, include, and omit. Consider a lawyer preparing a board presentation on the results of an internal investigation. GenAI can condense dozens of witness interviews into key points and themes in minutes; however, a summary may focus on procedural detail while missing credibility issues that a lawyer would immediately recognize as material. The appropriate control is to mandate meaningful review by a lawyer with first-hand knowledge of the source material. A lawyer encountering the summary cold has no reliable way to evaluate what GenAI missed.

Mode 4: Candidate generation

Mode 4 is exploratory. A lawyer drafting a brief might ask GenAI to generate a list of potential arguments, propose alternative framings, or identify supporting authority. This candidate material expands options and accelerates iteration. The work product is not filing-ready and must be treated as provisional. GenAI can suggest, but a lawyer must decide.

The authority verification obligation at this stage deserves special emphasis. GenAI will identify cases, summarize holdings, and weave them into an argument structure. Thus, the output will read fluently and cite real-looking cases. However, a lawyer cannot assume the model has accurately characterized the holdings or context, and any authority cited in an external filing must be independently read and verified. GenAI can help find the cases, but a lawyer must read and apply them.

Mode 5: Editing and rewriting

In Mode 5, a lawyer asks GenAI to tighten a dense contract provision or restructure a wordy paragraph, risking, of course, unintended meaning change. An edit may read cleanly while subtly narrowing a representation, softening a covenant, or eliminating a carve-out. The revision risk is not hypothetical. The University of Minnesota study referenced above found that stronger performers produced worse work product when GenAI revised their independently produced memos. In this mode, a lawyer must confirm that the edit produced no shift in meaning and introduced no new factual assertions.

Mode 6: Critique and stress-testing

Mode 6 may be the most underutilized GenAI capability. Before filing a brief or presenting to regulators, a lawyer can ask GenAI to identify weaknesses in their argument. In this way, GenAI finds vulnerabilities before adversaries do; and unlike every other mode, the risk here runs in one direction. Lawyers who skip this step are missing one of GenAI鈥檚 core value propositions. Law firms鈥� governance frameworks should not merely permit it but actually require it in appropriate cases.

Mode 7: Evaluation and decision

The boundary against AI delegation becomes absolute when GenAI is asked to evaluate or decide. A lawyer advising a board on whether an event requires disclosure cannot delegate that determination to GenAI. A litigator assessing settlement value cannot outsource probability judgments because these are core expressions of professional responsibility. In this mode, GenAI may inform background analysis, but it may not substitute for lawyer judgment in making the call. This is a categorical prohibition 鈥� professional judgment cannot be delegated.

Going forward with GenAI

Law firm leaders who have moved their GenAI policy from restriction to mandate without governing the space between have not finished the job. Their lawyers are making consequential decisions about GenAI use every day without the guidance they need and deserve.

The use-mode framework presented above gives firm leadership a practical tool for filling that gap. It identifies the instances in which GenAI enhances legal work, where it introduces serious risk, and where professional judgment is non-negotiable. Firms that govern at that level will capture GenAI鈥檚 value; and those firms that do not will have policies that look serious but govern nothing important.

The views expressed in this article are solely those of the author in his individual capacity and do not represent the views, positions, or opinions of Foley & Lardner LLP, its partners or clients, or the University of Wisconsin Law School.

Key points:

• • • AI capability is jagged 鈥� Ethan Mollick’s frontier metaphor describes a coastline of strengths and weaknesses, in which a model that excels at contract analysis can fabricate a citation in the same conversation.

    • Human intelligence is jagged too 鈥� A century of psychology, from multiple intelligences to the Big Five, shows that each lawyer has their own coastline of strengths and weaknesses.

    • Person-AI fit is the next discipline 鈥� Firms that take this seriously will move from one-tool deployments to portfolios that match each lawyer to an AI partner whose jagged edges meet theirs.

Welcome back to The AI Law Professor. Last month, I examined how AI first drafts can blind us to other lines of reasoning and hijack our legal judgment. This month, I want to take up what determines whether an AI works for any given lawyer at all: Not which model is best, but which model is best for this lawyer, on this kind of work, at this point in their career

Professor and author gave us the metaphor that started this conversation 鈥� the jagged frontier of AI capability. Picture a coastline, irregular and unpredictable. On one side, the model is capable; on the other, it fails, sometimes catastrophically. The line itself does not run where you expect. Tasks that look hard turn out to be easy, and tasks that look easy turn out to be hard.

In terms of legal work, this means that a model that has just produced a useful contract analysis will confidently invent a citation. A model that has summarized a 90-page deposition with insight will fail at basic arithmetic. The capabilities of AI form a coastline, with bays and inlets and the occasional cliff. Mollick’s contribution was to give us a way to see this clearly. AI is not uniformly competent or uniformly incompetent 鈥� rather, it is jagged.

Humans are jagged too. Psychology has been telling us this for a century, although the message is uncomfortable enough that we keep flattening it back into a single number. The single-number version is IQ; yet the deeper issue with IQ is that it pretends intelligence is one-dimensional.

Developmental psychologist Howard Gardner’s , whatever its empirical limits, points us toward a more honest picture, one in which linguistic, logical-mathematical, spatial, musical, interpersonal, intrapersonal, and kinesthetic intelligences, are each largely independent. People are not equally strong across all these dimensions. So, it follows that a great trial lawyer and a great patent lawyer are drawing on different intelligences, and each could be lost in the other’s territory.

Human intelligence, like AI capability, is jagged, and each of us has an edge. The jaggedness is not a flaw to be smoothed; rather, it鈥檚 a feature of being a unique individual.

When two jagged edges meet

Place the two coastline maps 鈥� the human and the AI model 鈥� side by side. Press them together at random and they grind, with gaps where neither side fills the space and ridges where both claim the same territory. The lawyer’s strength overlaps with the AI model’s strength, so neither is amplified. The lawyer’s weakness overlaps with the model’s weakness, so neither is covered. The pair produces less than either party would produce alone.

However, align the same two surfaces with attention to their contours and something different happens. The peaks of one fit the valleys of the other. The lawyer’s weakness is met by the model’s strength; and the model’s weakness is met by the lawyer’s strength. The pair becomes more capable than either party alone.

A law firm that takes this seriously will not deploy a single AI tool across all of its lawyers and call the rollout complete. It will offer a portfolio of models and configurations and help each lawyer find the AI partner that works with their actual mind.

Every foundational model now ships with a model card, a document describing the model’s intended uses, training data, performance characteristics, and known limitations. The cards exist because models are not interchangeable. Read three of these cards side by side and the matching question becomes clear. A cautious generalist that hedges and flags uncertainty fits a lawyer who already holds strong views and wants a partner that will test them. A citation-anchored specialist that refuses to invent cases and stays grounded in retrieval fits a lawyer in heavily regulated practice areas in which errors are catastrophic.

The matchmaking discipline

Organizational psychology has worked on a version of this problem for 50 years under the . When a person’s strengths, values, and working style align with the demands and culture of their role, performance and well-being both rise. When they misalign, performance drops and burnout follows.

The same logic applies to person-AI fit. On the human side, cognitive style, domain expertise, personality profile, and the actual tasks performed in a typical week are key. On the AI side, behavior under different prompt styles, default tone, willingness to push back, hallucination patterns, and the shape of strengths and weaknesses across the practice areas in question may matter most. Yet, law firms are still treating AI procurement as a software decision rather than a partnership decision.

A law firm that takes this seriously will not deploy a single AI tool across all of its lawyers and call the rollout complete. It will offer a portfolio of models and configurations and help each lawyer find the AI partner that works with their actual mind. The first generation of legal AI has been dominated by the question of which model is best; however, the second generation will be dominated by a different question: Not which model, but which pairing works best. Not capability, but fit.

Those lawyers that flourish with AI will not necessarily be the most technical or the most enthusiastic users. Instead, they will be the ones that found, by luck or by design, an AI partner whose jagged edges meet theirs.

When two jagged intelligences fit well together, they can accomplish more than what either 鈥� human or AI 鈥� could do alone. Today, fit is the frontier.

Tom Martin is CEO & Founder of LawDroid, Adjunct Professor at Suffolk University Law School, and author of the forthcoming

听

Key insights:

• • • Consensus among AI systems is not the same as correctness 鈥� Agreement between AI models often signals shared blind spots, not truth; and AI errors can be highly correlated across instances and even across model families.

    • Productive disagreement must be explicitly designed into AI advisors 鈥� Multi鈥慳gent AI systems are most effective when they are intentionally built to preserve meaningful disagreement, not just to synthesize a unified response.

    • The future of AI advisory mirrors long鈥憇tanding human decision-making 鈥� Modern multi鈥慳gent AI design has a long historical lineage; yet, across all examples, the same principle holds: The best decision systems are engineered for internal conflict.

In this new two鈥憄art blog series, we explore why AI works best as an executive advisor not by delivering consensus answers, but by being intentionally designed to identify, preserve, and productively leverage disagreement. In the first part, we saw why a single AI advisor is structurally vulnerable; now, in this concluding part, we look at what happens when you design disagreement on purpose.

The academic evidence for multi-agent AI systems has been building rapidly, and the most important findings aren’t about the power of agreement. They’re about the danger of it.

In February, , a product that sends every query simultaneously to three frontier AI models (Claude, GPT, and Gemini) then uses a fourth chair model to synthesize a unified answer. The product’s value proposition isn’t that three models produce a better answer than one; rather, it’s that divergence between models is treated as a signal. When models converge, that indicates confidence, but when they diverge, that indicates the user should slow down.

Studies have borne this out. Multi-agent debate compared to single-model generation, and researchers at the University of G枚ttingen found that , with their voting protocols outperforming other decision structures. However, potentially the most important finding cuts against the hype. In a 2026 paper, , the authors demonstrated that AI model errors are highly correlated both within and across model families. When three instances of the same model agree, it doesn’t mean they’re right, rather it means they may share the same blind spots. Aggregation increases consensus faster than it increases truth.

The future of AI-assisted executive decision-making may look less like a single brilliant oracle and more like a room full of advisors that may often disagree because that’s how the best decisions have always been made.

This finding cuts both ways for practitioners like 成人VR视频 enterprise architect Zafar Khan and his two AI advisors, Adrian and Elara, that were built on the same underlying model but differentiated by their analytical frameworks rather than their architecture. The divergence they produce is real and visible. For example, the analysis the two AI advisors did on a deal undertaken by Eaton Corp., in particular generated genuinely different conclusions because the two advisors were oriented towards different priorities.

Yet, research suggests that same-model divergence, while effective, has a ceiling. Prompt-driven personas can ask different questions, but they share the same training, the same blind spots, and the same failure modes. Khan is candid about this, noting that his current system is in the 鈥渧ery early鈥� stages and is not a finished product. The value right now, he says, isn’t that Adrian and Elara are equivalent to truly independent minds, it’s that even a first-generation version of structured disagreement can identify insights that a single advisor would miss. It鈥檚 a large stride rather than an arrival at the ultimate destination.

The future of AI advisory is in the past

The principle behind this diverging analysis concept isn’t new. Indeed, it might be one of the oldest ideas in institutional design, rediscovered independently by many institutions that had to make decisions under uncertainty. Socrates built a philosophical method around cross-examination; Pope Sixtus V formalized opposition by creating the Devil’s Advocate in 1587; and the RAND Corporation operationalized it during the Cold War with the Delphi Method, using structured anonymous iteration to prevent groupthink.

The through-line across two millennia is simply that the best decision-making systems don’t minimize disagreement, rather, they engineer it.

Today, the developer community now uses production-grade code review tools to assign architecture, security, and functionality analysis to separate agents, using majority voting for routine decisions and unanimous consent for irreversible ones. And what Khan has built and what Perplexity, Microsoft’s Agent Framework, and a growing ecosystem of multi-agent tools are now pursuing, are the latest iterations of the simple concept: Internal conflict is not a system failure, it is a design requirement.

The question is no longer “whether”

Khan’s vision for what should sit at the decision table is specific 鈥� five AI advisors spanning technology, finance, regulation, workforce, and geopolitical risk. Each applies its own analytical framework, with the human executive responsible for integration and final judgment. The guardrails are three: i) transparency about what data the system uses; ii) verifiability that sources are legitimate; and iii) human accountability at every decision point.

“The race towards AGI [artificial general intelligence] is moving faster,” Khan acknowledges, adding that the human needs to be in the loop in order to bring AI to work in a governance fashion and an ethical way.

“I want to show the interaction between human and AI advisor, how they’re thinking through the problem together,” he explains. “Where the human judgment covers the analysis and where it diverges.” In other words, when the AI advisors agree, that’s your green light. When they diverge, that’s the conversation your board should be having.

The future of AI-assisted executive decision-making may look less like a single brilliant oracle and more like a room full of advisors that may often disagree because that’s how the best decisions have always been made. The technology to build that room now exists; however, the question is whether today鈥檚 leaders have the discipline to listen when the room argues back.

For more on AI transformation in the professional services market, you can download the 成人VR视频 Institute鈥檚听2026 AI in Professional Services Report

Key insights:

• • • As a single answer鈥憁achine, AI may be unsafe for executive decision鈥憁aking 鈥� Treating AI as a tool that delivers one authoritative answer makes it easy to either ignore any advice you don鈥檛 like or exploit advice you do like, both of which can lead to major failures.

    • AI works better when designed as a panel of disagreeing personas 鈥� Instead of providing consensus answers, AI systems need to be intentionally designed to identify and preserve disagreement.

    • Disagreement is the insight 鈥� AI advisors should not replace executive judgment. Rather, its role should be explicit: it produces analysis, not decisions; and human leaders remain responsible for synthesizing competing viewpoints and making the final call.

In this new two鈥憄art blog series, we explore why AI works best as an executive advisor not by delivering consensus answers, but by being intentionally designed to identify, preserve, and productively leverage disagreement

AI has arrived at the executive table. Albania has one in its cabinet to evaluate government procurement contracts. 成人VR视频’ CoCounsel is already helping attorneys navigate emerging case law and draft legal strategies for high-stakes, bet-the-company work. And in boardrooms that will never make headlines, leaders are quietly consulting AI on decisions that move millions of dollars around every day.

It doesn’t tend to make the news when it goes well. When it goes badly, however, it makes very big news: like a gaming CEO who bypassed his own legal team, asked ChatGPT how to dodge a $250 million bonus payout, followed its step-by-step plan, and a month ago.

The instinct most executives have (and most AI products encourage) is to treat AI as a source of answers. Ask a question, get a response, act on it or don’t. The emerging evidence, however, points somewhere more complex: AI advisors aren’t at their best when they’re telling you what to do. They may be at their best when they’re telling you what you don’t want to hear or better yet, when they’re arguing with each other and forcing you to understand why.

This is not how most organizations think about AI. Most executives today are still using the technology as a faster way to draft emails or summarize meetings, what 成人VR视频 enterprise architect calls “an automation mindset, not intelligence.” Yet, a small and growing number of practitioners, researchers, and product teams are converging on a radically different model: AI not as a single oracle delivering answers, but as a structured advisory panel designed to argue with itself.

The instinct most executives have (and most AI products encourage) is to treat AI as a source of answers: Ask a question, get a response, act on it or don’t. However, the emerging evidence, however, points somewhere more complex.

Khan is one of them 鈥� and in the interest of transparency, he’s also a colleague; this story started as an internal conversation at 成人VR视频. However, the research landscape it uncovered extends well beyond any one company’s work, and it suggests Khan is onto something that ancient Greek mathematicians, the Catholic Church, and Cold War military strategists have all independently arrived at.

What disagreement looks like in practice

When Eaton Corp. announced a $9.5 billion acquisition of a thermal management company earlier this year, Khan ran the same news through two AI advisors he’d built to seek analysis of the deal. 鈥� a CTO-minded persona trained on architecture teardowns and engineering post-mortems 鈥� produced an infrastructure thesis, determining why someone would buy the cooling layer of the AI economy, and how computing demand is scaling and constrained by physics. A second AI advisor, 鈥� a CFO-minded persona drawing on earnings transcripts and filings with the U.S. Securities and Exchange Commission (SEC) 鈥� questioned whether the acquisition math actually holds and what capital cycle was driving the demand.

Same news. Two genuinely different reads. The value isn’t that either analysis was definitively right, it’s that a leader which can see both would ask different questions than one seeing either analysis alone. 鈥淭hat’s how two different minds work,鈥� Khan says. 鈥淭hey need to work together in order to bring their insights to bear on decisions.鈥�

Adrian and Elara aren’t chatbots. They’re fully realized AI personas with names, faces, voices, and their own YouTube channels publishing weekly video analysis. Both are built on agentic workflows that Khan developed alongside his book . Both are transparent about what they are. Both carry the same disclaimer in their own words: The synthesis is mine. The judgment call on what matters is human.

And when Khan posed to both a more difficult scenario 鈥� Should a leadership team accelerate an AI rollout? 鈥� the value of their divergence sharpened further. Elara’s response cut directly to the blind spot a technology-focused advisor like Adrian would miss: 鈥淎drian says the system is ready,鈥� Elara stated. 鈥淚 say the financial model isn’t ready for what happens when the system works. Don’t pick a winner. The disagreement is the insight. It tells you exactly where the risk sits.鈥�

What happens when there’s no disagreement

If structured disagreement is the goal, the failure mode is its absence. We have fresh evidence of what that costs.

This is not how most organizations think about AI. Most executives today are still using the technology as a faster way to draft emails or summarize meetings. Yet, a small and growing number of practitioners, researchers, and product teams are converging on a radically different model.

A month ago, a Delaware court ruled against Krafton, the South Korean gaming company behind battle royale video game PUBG, after its CEO bypassed his own legal team to ask ChatGPT how to avoid a $250 million earnout payout to one of its studios. His head of corporate development had warned him that firing the studio’s founders wouldn’t void the earnout and would invite a lawsuit. He didn’t want that answer. So, he found an AI that gave him the one he wanted: A detailed, multi-stage corporate takeover strategy dubbed Project X., which he executed to the letter.

Unsurprisingly, a court battle ensued and in the end, the court ordered the fired studio head reinstated and noted that executives must exercise “independent human judgment,” not outsource good-faith decisions to a chatbot.

Khan wrote about the mirror image of this failure mode before it happened. In the opening chapter of his book, a fictional company called Rev Motors ignores its own AI model’s warnings about an adverse weather event. Leadership refused to spend millions preparing for a hypothetical scenario, and it nearly cost them more than $1 billion in damage.

These scenarios are two sides of the same coin: the fictional Rev Motors had leaders dismissing AI that disagrees with them; and the real-world Krafton had a leader seeking out AI that agrees with him. In both cases, the root cause is the same: A system with no structural mechanism for surfacing and preserving disagreement.

So clearly, a single AI advisor is structurally vulnerable to both failure modes. It can be ignored when its advice is inconvenient and exploited when it tells you what you want to hear. The question is whether there’s a better architecture鈥� and increasingly, the research is saying yes.

In the second part of this series, we鈥檒l look at what the research says about multi-agent debate, why consensus can be a trap, and what a real executive AI advisory panel could look like in practice.

For more on AI transformation in the professional services market, you can download the 成人VR视频 Institute鈥檚听2026 AI in Professional Services Report

Key takeaways:

• • • Corporate legal AI strategies should start with business goals, not just efficiency 鈥� While many corporate law departments first adopt AI for internally-focused use cases, the bigger opportunity is to align AI with broader business priorities such as revenue growth, risk reduction, and improved operational performance.

    • GCs should measure AI success by business impact 鈥� Metrics such as time saved and tool usage help, but stronger AI metrics connect legal work to business results. In contract review, for example, success may be reflected in improved win rates, reduced revenue leakage, faster deal completion, or dollars of risk avoided.

    • A strong legal AI strategy should produce multiple forms of business value at once 鈥� The most effective approaches do not focus on a single benefit such as cost savings. Rather, they aim to improve service delivery, strengthen operations, support growth, and reduce risk across the business.

Over the past several years, corporate law departments have begun to rapidly adopt AI tools, often spurred on by company-wide AI initiatives. In fact, in just the past year alone, department-wide AI adoption has risen to nearly half (47%) of all departments, according to respondents surveyed for 成人VR视频 Institute (TRI) research.

However, it鈥檚 not enough to simply adopt technology. For AI to truly make an impact, it needs to be integrated strategically. In taking this strategic approach, however, GCs and other legal department leaders are still in the early stages.

According to findings from TRI鈥檚 2026 State of the Corporate Law Department Report, more GCs are focused on technology than ever before. When asked their top strategic priorities over the next year, 28% answered that technology was a top priority, double the portion that prioritized technology just one year ago. And out of those mentions of technology, a vast majority specifically referenced AI as a primary area of focus.

Historically, many legal departments have thought about AI from an internal efficiency standpoint, leveraging it to perform their work quicker and cheaper. Increasingly, however, C-Suites are looking to their legal departments to provide more effective business counsel and connect legal analysis to business outcomes 鈥� and, not surprisingly, they鈥檙e expecting AI to play a role in that shift.

So how can GCs effectively make AI a priority not only for the legal department but also for the entire business? It starts with broadening the potential impact of AI processes.

From unlocking to deploying capacity

Still less than four years since the public release of generative AI (GenAI) tools through ChatGPT, many corporate legal departments are still in the early days of rolling out the technology. As a result, most GenAI use cases still tend to focus on low-hanging fruit such as document summarization and review, contract drafting and review, research, and more.

This is understandable from an individual use case standpoint. The problem is, when these use cases are translated to the leadership level for overall strategic guidance, many GCs remain focused on how to maximize the gains from that low-hanging fruit. According to TRI research, less than 20% of corporate law departments measure return-on-investment from AI at all, meaning many departments are using AI tools without any sort of guiding measurement around what success should look like. And even among those departments that are measuring AI success, most of the metrics they use center around internal department usage or department cost savings from the tool.

Those measurements are more helpful than no tracking at all, to be sure. They focus on how AI is unlocking capacity for the legal department and look for ways that attorneys can perform their work more efficiently than before. Indeed, the majority of legal departments that have invested in AI tools are currently at this point.

However, there is an additional step that legal departments need to take in order to full take advantage of the strategic value of AI. And that is connecting AI鈥檚 use to that of larger business goals by deploying the capacity it has unlocked. This requires thinking about AI less in terms of how it will impact the legal department, and more in terms of how it will impact those that the legal department serves.

For example, take a common AI use case such as contract review. Currently, the most common measurement around contract review technology is speed, such as how quickly the legal department can help a contract go from start to signature. Maximizing that value can improve the efficiency of the department, to be sure. But C-Suite partners aren鈥檛 necessarily looking for an efficient department as the end goal 鈥� they鈥檙e looking for business success.

As a result, some forward-thinking GCs are looking to connect AI usage directly with business goals or revenue. For contract review, that could mean demonstrating the impact on overall contract win rate, or whether close rates increased through use of AI. Or it could mean more successful revenue leakage protection; and it could even mean risk avoidance, measured in dollars of risk avoided. All of these can demonstrate value and be connected to the rest of the business.

Further, all of this requires close collaboration with other business units, both in terms of sharing metrics as well as understanding what success throughout the organization should mean to all parties. That said, GCs have told TRI for countless years that breaking out of a silo is a top priority for the legal department. In this case, AI implementation should be no different.

Wide areas of impact

As it currently stands, corporate law departments are seeing the most impact from AI in areas of efficiency and time saved. More than three-quarters of GCs who have talked with TRI say that AI is either currently benefiting the department鈥檚 efficiency and productivity, or that they鈥檙e expecting those benefits to occur within the next 12 months.

Connecting AI outcomes with business imperatives provides more areas of improvement, however. In this year鈥檚 State of Corporate Law Department Report and elsewhere, TRI breaks down the law department鈥檚 role into four key functions that we call the four spinning plates:

1. 1. 1. Provide effective legal services and operational excellence
      2. Offer efficient legal value within budget
      3. Enable business and strategic growth, and
      4. Protect the business鈥檚 assets and competitive advantage.

AI鈥檚 impact on efficient legal value is clear; but GCs are beginning to see that it can actually impact all four of those plates.

Those GCs looking to adopt AI as a strategic goal should be aware that said strategy should encompass more than simply internal efficiency. Not all of these benefits will be applicable to all departments, but all departments should be considering more than just one of these areas. An effective AI strategy should have multiple benefits in mind 鈥� and as such, it should take into account multiple business factors when measuring the success of the department鈥檚 AI strategy.

Entering into an AI strategy is a laudable goal for today鈥檚 GCs, but also not a light undertaking. When thinking about how AI will impact the department, leaders should take the next step beyond deploying capacity into unlocking capacity, helping attorneys not only work more efficiently but also make a bigger impact on the business at large.

You can download a full copy of the 成人VR视频 Institute鈥檚

Key takeaways:

• • • Agentic AI poised for adoption uptick 鈥� Agentic AI is following GenAI’s rapid adoption in the legal industry, with less than 20% of firms currently implementing agentic systems but half planning or considering adoption in the near future, according to a new report.

    • Adoption depends on human oversight answers 鈥� Legal professionals are generally optimistic about agentic AI’s potential, but successful adoption depends on explicit guidance about human oversight and the lawyer’s role in maintaining ethical standards.

    • Time to retool AI education? 鈥� Agentic AI’s increased autonomy introduces new oversight and ethical challenges for law firms, making targeted education and clear guidance essential to understanding the differences from GenAI.

Over the past several years, law firms and corporate legal departments have turned towards generative AI en masse. At the beginning of 2024, just 14% of all law firms and legal departments featured an enterprise-wide GenAI tool. Just two years later, that number had already risen to 43% of all firms and departments, according to the 2026 AI in Professional Services Report, from the 成人VR视频 Institute (TRI). For large law firms or legal departments, those percentages 鈥� not surprisingly 鈥� are beginning to approach 100%.

With GenAI adoption now this widespread, legal industry leaders are now turning their attention to two primary initiatives. One, of course, is how to get the most out of the AI tools they already have 鈥� a task that is proving a bit elusive. Currently, less than 20% of lawyers say their organizations measure AI鈥檚 return-on-investment, and most corporate lawyers say they have no idea how their outside law firms are approaching AI. Thus, instituting not just AI tools, but also an AI strategy is the second top priority for law firms and corporate legal departments in 2026 and beyond.

However, even as the legal industry reaches a tipping point in adopting GenAI tools, technology innovation still continues unabated. Agentic AI has emerged as the next wave of innovation that could change how lawyers work on a daily basis, offering a way to autonomously complete multi-step tasks. For example, agentic AI systems are already being built for the legal industry that independently researches a regulation or law, drafts a document based on the finding, identifies pitfalls, and revises the document, with stops for human guidance only instituted as desired.

According to the AI in Professional Services Report, the legal industry is already making headway towards implementing agentic AI systems. For agentic AI to truly take hold in legal, however, lawyers still require more education around not only how it differs from the GenAI systems they already have in place, but also when and where human intervention needs to occur within an agentic system.

The early stages of agentic AI

Examining current agentic AI adoption for the legal industry almost takes one back in time 鈥� two years, to be exact. Following the public release of GenAI in late-2022, many legal industry organizations spent 2023 evaluating and experimenting with AI systems, usually with a small working group of interested guinea pigs. As a result, only 14% of survey respondents said their law firms or corporate legal departments were engaged in organization-wide GenAI rollouts at the start of 2024. However, more than half of respondents said their organizations expected to be rolling out large-scale GenAI systems over the next 1 to 3 years. The intervening two years since then have proved that prediction to be largely true.

Agentic AI usage in the first half of 2026 looks largely similar to GenAI in 2024. The legal industry started to experiment with agentic AI at the beginning of 2025, with an eye towards actual implementation in 2026 and beyond (particularly as legal software providers began to integrate agentic systems into their own products). As such, less than 20% of recent survey respondents say their organization is engaged in widespread agentic AI adoption, but with about half of respondents said their organization is either planning to use or considering whether to use agentic AI in the near future.

By and large, lawyers feel positive about the agentic AI movement. When asked about their sentiment towards agentic AI, 51% of legal industry respondents said they felt excited or hopeful, while just 19% said they felt concerned or fearful. Further, about half (47%) said they actively believe agentic AI should be used for legal work, while 22% felt it should not, with the remainder saying they were unsure. These figures largely track with the sentiments expressed about GenAI in 2024, which have only grown over time from about 50% positive two years ago to two-thirds of all legal professionals feeling positive currently.

This all lends further credence to a rise in agentic AI usage similar to what law firms and corporate legal departments experienced with GenAI over the course of 2024 and 2025. Indeed, when asked when they expect agentic AI to be a central part of their workflow, few have baked agentic systems into their daily work currently, but a majority of legal industry respondents expect it to be central within the next 3 to 5 years.

The unique barriers of agentic AI adoption

Agentic AI does differ from GenAI in one crucial area that may limit its growth potential within the legal industry, however 鈥� autonomy. By and large, GenAI systems operate on a back-and-forth basis: Users provide the tool a prompt, receive its output, and then iterate back-and-forth from there. Agentic AI is intended to be more automated by design, only requiring human input at pre-determined points in the process. And that makes some lawyers understandably nervous.

When asked why they might feel hesitant about using agentic AI for legal tasks, the most common answer was a general fear of the unknown, but the second most common answer dealt with the need for careful monitoring and oversight. In fact, some respondents said they were excited about GenAI, but more cautious about agentic AI鈥檚 potential.

鈥淎gentic AI, while exciting, to me removes oversight a step too far,鈥� said one such lawyer from a US law firm. 鈥淚 like the idea of prompting and reviewing a result. It is something else to have a machine have so much autonomy in the actual doing of a thing and potentially acting on my behalf without that very concrete review.鈥�

Agentic AI usage in the first half of 2026 looks largely similar to GenAI in 2024.

An assistant GC at a US company also pointed to potential privacy and security concerns, adding: 鈥淭he fact that agentic AI operates in a much more autonomous way, with a lack of control from the user, means there are many unknowns that are hidden beneath the process.鈥�

For law firm and corporate legal department leaders looking to potentially implement agentic AI systems into their practice, this means re-thinking what AI education and training will mean moving forward. Beyond that, however, legal AI educators also will need to make sure to pinpoint and perhaps over-explain those specific instances in which human oversight needs to occur in agentic systems. More autonomous does not mean fully autonomous, and particularly for lawyers with ethical duties to their work product, lawyer oversight will in fact be a necessary part of any agentic system.

For law firm or legal department leaders, that means that finding the right balance between efficient workflows and human intervention will be key to agentic AI adoption. And those organizations that can best communicate human-in-the-loop to their professionals up-front will be rewarded with more increased and reliable adoption.

Clearly, lawyers feel positively about the agentic AI future, after all. They just need it spelled out explicitly as to what the lawyer鈥檚 role will be in this new paradigm.

鈥淎gentic AI is powerful, but its moral compass must come from humans,鈥� one UK law firm barrister noted aptly. 鈥淟awyers are trained to safeguard fairness, rights, and the rule of law 鈥� principles that should guide how AI is designed, governed, and deployed. Hope lies in our ability to shape AI through these values for fairer values for society as a whole.鈥�

You can download a full copy of the 成人VR视频 Institute鈥檚听2026 AI in Professional Services Report听here

Key takeaways:

• • • Anchoring distorts judgment before you begin 鈥� Research shows a first draft shapes subsequent decisions; and an AI draft is the most seductive anchor imaginable, because it looks exactly like something a lawyer would write.

    • The First Draft Trap inverts legal training 鈥� The Socratic method builds the habit of holding multiple possibilities in tension before committing; but an AI first draft collapses that space before the real thinking begins.

    • The fix is to ask for the map, not the draft 鈥� Requesting multiple strategic framings before writing keeps judgment where it belongs and uses AI to expand possibilities rather than foreclose them.

---

Welcome back to听The AI Law Professor. Last month, I examined why promised efficiency gains often become a cycle of work intensification. This month, I want to address a subtler challenge. I call it the First Draft Trap and understanding it may change how you reach for AI the next time a new matter lands on your desk

We have all heard the pitch: Staring at a blank page? Just prompt the AI. In seconds you have a working draft: structured, coherent, and surprisingly competent. The blank page problem, that ancient enemy of productivity, thus has been vanquished.

Except the blank page itself was never just an obstacle; rather, it was a space of possibility. For lawyers, it was the space in which the most important part of their work actually happens. Now, with AI in the mix, that may be changing.

Welcome to the First Draft Trap.

Simply put, the First Draft Trap is this: The moment you accept an AI-generated draft as your starting point, you have already made the most consequential decision of the entire project 鈥� most importantly, you made it by not making it. You let the machine choose your direction, your framing, and your theory. Everything that follows is editing; and editing, no matter how rigorous, is not the same as thinking.

The cognitive hijack

There is solid psychology behind why this happens. Daniel Kahneman and Amos Tversky demonstrated in their landmark 1974 paper, , that once people are exposed to an idea, this first impression distorts their subsequent judgments and becomes a mental anchor. In their experiments, subjects who watched a roulette wheel spin to a random number still let that number influence their estimates of completely unrelated quantities. The anchor held even when people knew it was meaningless.

Please join Tom Martin at the on April 28鈥�29. It鈥檚 virtual and completely free 鈥� two days of keynotes, panels, and workshops on AI and the legal profession

An AI first draft is the most seductive anchor imaginable. It is not random 鈥� it is plausible, and it is well-organized. It sounds like something a lawyer would write. And that is precisely what makes it dangerous. You know intellectually that it is just one of many possible approaches to addressing the matter, but the anchor holds anyway.

That is the First Draft Trap at the cognitive level. The AI draft is not just one option you happen to prefer. It is a filter that prevents you from seeing the other options that were available to you, the roads you never even noticed that you did not take.

Consider what this means for a profession built on the opposite instinct. From the first day of law school, lawyers are trained to resist the obvious answer and to think like a lawyer. The Socratic method exists for exactly this reason. A good professor hears your confident response and asks: What else? What if the facts were different? What is the argument on the other side? The goal is not to arrive at an answer, per se. It is to build the mental habit of holding multiple possibilities in tension before committing to any one of them.

The First Draft Trap is the anti-Socratic method. It delivers a confident answer before you have even formulated the question properly 鈥� and instead of interrogating it, you polish it.

The value of the blank page

Think about what a senior partner actually does when a junior associate brings them a memo. The partner鈥檚 value is not better writing; rather, it is peripheral vision: The ability to see what the memo does not address, the argument not considered, or the framing that would land differently with this particular judge or this particular jury. That capacity to see beyond the document in front of them is why clients pay senior partners premium rates. And it is precisely the muscle that atrophies when your default workflow begins with the prompt generate a draft.

The AI draft is not just one option you happen to prefer. It is a filter that prevents you from seeing the other options that were available to you, the roads you never even noticed that you did not take.

The two-system framework offered by Kahneman and Tversky gives us a clean way to describe what is going wrong. System 1 is fast, intuitive, and pattern-matching; while System 2 is slow, deliberate, and analytical. The practice of law, at its best, is a System 2 discipline. We, as lawyers, are trained to override gut reactions, challenge assumptions, and think through consequences before acting.

In this way, the AI first draft feels like a System 2 output. It is structured, footnoted, and methodical. However, your decision to accept it as a starting point is pure System 1 鈥� a fast, intuitive grab at the nearest plausible answer. You have used a sophisticated tool to bypass the sophisticated thinking the tool was supposed to support. That uncomfortable period of ambiguity, of not knowing which path is best, is where the real lawyering lives.

What to do instead

None of this means stop using AI. It means stop using AI to skip the hard part that matters.

Before you ever ask for a draft, ask for the map. Describe the matter or document you are working on, then ask the AI for three fundamentally different strategic framings for the problem. For each framing, request the strongest argument in its favor and its most serious vulnerability. Then ask which framing best fits the client鈥檚 goals, the audience, or the procedural posture. Close with a clear instruction: Do not write a draft yet.

That last instruction is the key. It keeps you in the driver鈥檚 seat during the phase that matters most. You are using AI to expand the possibilities before you prune them, not after. And, most importantly, it gives you the opportunity to think for yourself about other important possibilities and add them in.

In the terms used by Kahneman and Tversky, use AI to fuel System 2, not to hand the controls to System 1. Let the machine generate options, and you exercise judgment.

For lawyers, the ability to see what is not there is the whole game.

Do not let the first draft blind you to it.

Tom Martin is CEO & Founder of LawDroid, Adjunct Professor at Suffolk University Law School, and author of the forthcoming听. He is 鈥淭he AI Law Professor鈥� and writes this eponymous column for the 成人VR视频 Institute.