Legal professional skepticism of the future value of change investment leads to underinvestment.

A first-pass look at our future.

In earlier Legal Evolution posts, I’ve shared reflections upon my career journey (080), professional evolution (143), and current area of focus (159).  This article describes an investment hypothesis for the upcoming decade focused on building the future of the practice of law [hereafter, “future practice”]. The conclusions are inaccurate, which obviously requires some explanation.

The goal of this essay is to frame the problem, describe the starting direction, and share updates as we get smarter.  Indeed, what we are trying to solve is what policy analysts call a “wicked problem”— a problem so complex that it is highly resistant to resolution.  Because “[t]ackling wicked problems is an evolving art,” Australian Public Service Commission, “Tackling wicked problems: A public policy perspective,” June 12, 2018, we must give away the playbook so that the entire ecosystem can evolve to support our adaptation.

[Editor’s note: This is a dense essay that’s worth your time.  If Susskind gets credit for the concepts, Barnwell is the practice master who can draw the blueprints. wdh]

Society needs better legal infrastructure

Legal systems matter because they are infrastructure for civil society. We all share these roads. Our current legal infrastructure was designed for the pre-industrialized 19th century. See Gillian Hadfield, Rules for a Flat World (2016) (history of the development of legal infrastructure); see also Post 207 (discussing Restatements of Law as a prime example of shared legal infrastructure). But we live, work, and play in a world with far more volume and types of interactions. Our legal infrastructure’s systemic lack of process rigor, information, and adaptability will produce consequences with increasing frequency and systemic outcome severity. See Chris Clearfield & András Tilcsik, Meltdown  (2018) (when complex, tightly coupled systems fail, they do so spectacularly).

Source: [click on to enlarge]
The implication is as obvious as it is hard:  We must evolve our legal systems and how we perform legal work within them to provide speed, scale, quality, and access. See Daniel W. Linna, “Evaluating Legal Services: The Need for a Quality Movement and Standard Measures of Quality and Value” in Research Handbook on Big Data Law (Ronald Vogl, editor, forthcoming 2021)(law needs standardization, empiricism, and data to evolve). The emerging dynamics that require these changes are being examined with more rigor and provide investment targets. See Nathan Bennett & G. James Lemoine, “What VUCA Really Means for You,” Harv Bus Rev, Jan-Feb 2014 (offers general approaches to volatility, uncertainty, complexity, and ambiguity, referred to as “VUCA”).

Legal Infrastructure’s change resistance was a feature and is a defect

Legal infrastructure is built upon institutions, systems, and culture that change slowly. This inertia insulates the participants from capricious change. It also limits the system’s ability to react to necessary beneficial change. Society’s accelerating evolution makes systemic non-reactivity to change more of a liability than an asset. These systemic traits devolve from us as legal professionals.

The legal profession has long selected for, trained, and rewarded narrowly focused subject matter experts who draw from the past to build the future.  Cf. Post 043 (reporting on launching of IFLP as a vehicle for training T-shaped lawyers with process and design complements). This backward focus has been rational behavior that has launched many successful careers. But it also carries downside risk, particularly for the profession as a whole, such as an underinvestment in assets that could prepare us for a future that looks very different from the past.

In the famous “multi-armed bandit” resource allocation problem,  the agent seeking to maximize her chances of survival has to allocate her finite attention between acquiring new knowledge (exploration) versus deciding based on existing knowledge (exploitation). If the maintenance of civil society depends upon redesigning and upgrading legal infrastructure, what is the optimal mix of our attention between exploration versus exploitation?

As this essay’s lead graphic suggests, if we are paying any attention at all, it’s obvious that our profession’s skepticism toward change affects our change investment.  See, e.g., Post 160 (per Randy Kiser, lawyer attitudes toward “non-essential” learning limits profession’s adaptability); Larry Richard, “Herding Cats: The Lawyer Personality Revealed,” Report to Legal Management, Aug 2020 (attorneys average 90th percentile for skepticism); Ron Friedmann, “Our Perception of Change,” Prism Legal, Dec 2019 (perception of change is relative to observer’s career starting point).

In turn, this underinvestment exposes all of us–lawyers, clients, and broader society–to potentially large system-level risk.  Below is a graphic that models the career investment behaviors of the typical legal professional. Note how our exploration tends to flatten out shortly after becoming a minimum viable lawyer (MVL).

Legal professionals invest less in exploration investment behaviors as their careers progress.

Modeling future behavior on past behavior without revision becomes irrational following system inflection points. Reshaping career investment patterns is just one small part of the overall wicked problem of legal infrastructure upgrade and adaption.

We are at a system change inflection point

People have predicted radical shifts in the delivery of legal services for decades. Pioneers have been doing the heavy lifting to prepare the market. See, e.g.,  “How to Run a Legal Department Like a Business,” Business of Law Podcast, July 2, 2020 (Jeff Carr details systematically creating a business-adaptive legal department). In my work for Microsoft I see signals in the commercial legal market that predict system change.

One of the most significant is that more legal professionals with elite executive and business management skills are ascending to leadership roles at enterprise legal services buyers and driving value-focused approaches to their business. Commercial engagement models among enterprise legal services buyers and sellers were historically driven by two primary factors: relationships and outcomes. The emerging breed of leaders are adding a third factor of how the work is done to maximize value for the business. Because they have executive leadership capabilities, they’re able to drive behavioral change at-scale within their organizations.

We also see legal executives who are data fluent. We will see far fewer senior legal leaders proudly proclaim they went to law school because they do not like math. Everyone who reports into the C-Suite will bring data that supports their story for the value their organization brings to the enterprise. See Corporate Legal Operations Consortium, “From Global Pandemic to Inspiring Innovation,” CLOC Global Institute, November 19, 2020 (General counsel share professional executive approaches to transformational leadership during a pandemic).

Enterprises are creating demands and constraints that require us to change how we deliver legal services.  As they invest in growth businesses, they are driving more work through their business capabilities, which in turn generates increased legal demand.  Yet, as the business value chain tightens, that work must be done more quickly. Businesses are simultaneously holding all functions more accountable for their resources and creating constraints that make it harder to service increased demand with conventional approaches. If capacity must increase by 10x, our current approaches breaks, as the option of a 10x increase in hiring is simply off the table. We will continue to see these demand/constraint pressures amplified as other parts of businesses make smart investments that force adaptation by their internal and external partners. See Walter Frick, “The Real Reason Superstar Firms are Pulling Ahead,” Harv Bus Rev, Oct 5, 2017 (outperforming firms holistically build business systems powered by IT investments).

Technology that can transform legal work has been available for decades. Unfortunately, it has not taken hold because it required too much specialized expertise to put into production and it required the people who do legal work to change too much to use the systems.  Yet, in recent years, the adoption costs have plummeted:

  • Cloud services ease the burden of acquiring new technology;
  • No/low-code application development platforms reduce dependencies upon scarce development talent; and
  • Natural language machine learning models create software powered experiences that do not require legal professionals to change much about how they work.

Indeed, low-code is akin to adding moveable type to the printing press—you no longer need a scarce expert to produce something good enough that scales. See Annie Keating, “Low-Code Automation and the Future of Work,” Forbes, Apr 28, 2020 (low-code software solutions bridge the divide between business needs and solutions).

Early career legal professionals are acquiring skills that complement their substantive legal capabilities that will let them leapfrog their conventional peers. People like Cat Moon, Dan Linna, Gabriel Teninbaum, and Houman Shadab are imbuing legal professionals with design and technical skills that change how they can solve problems.

This influx of skills changes the scale of the problems we can address. It also radically reduces the effort required to produce a solution because you do not have to attract skills that don’t normally interact. The combination of some system thinking and process skills (perceive the problem), some design skills (envision a potential solution), and some technical skills (implement the potential solution) converges to produce a legal architect who can make an alpha release product with little support. In small steps, this convergence is already happening. See, e.g., “Design Thinking,” Business of Law Podcast, Sept 3, 2019 (Microsoft partnered with Bold Duck Studio to catalyze innovation with basic business design skills).

In 2017 we ran an experiment in our productivity hacking community to empower anyone in the Microsoft Legal Department who wanted to create a bot to make a bot. See “Orrick and Microsoft Legal Productivity Hackers discuss innovation and Microsoft Teams,” Business of Law Podcast, May 23, 2019. We produced several simple bots that did basic things, but the real value was acquiring know-how and activating our people. One of those intrepid bot builders was not satisfied with his out-of-office bot and partnered with an engineering team to produce an enterprise-scale bot that supports the commercial legal team. See “Building an Enterprise-Ready Bot,” Business of Law Podcast, July 11, 2019.

Indeed, in 2020, people can build useful bots very quickly.  In part, this is because the required technical effort is now a fraction of what it was just a few years.  Another part is the combination of low/no-code and off-the-shelf machine learning services, which results in tools that are truly useful to the sophisticated tasks of in-house lawyers.

Indeed, these building blocks are available now. What is missing, however, is the cultural and technical capabilities to deploy them—which is why we need to invest now.  We can build or buy technology, but culture is always a nonnegotiable build.  At Microsoft, our know-how on this is hard-earned.

The future practice is in production at enterprises

Microsoft’s Open Source and Standards practice demonstrates a pattern for delivering legal services at massive scale. See “How to Scale a Practice,” Business of Law Podcast, July 20, 2020.

The open source practice has scaled to service 1,000 times the starting request volume over the last decade, yet the legal team has not grown. Scaling support for orders of magnitude of growth requires rebuilding the processes from the ground up, building optimized machine support, and a ruthless focus on conservation of human attention.

The open source practice built an extensible model for identifying the type of issue they were facing (e.g., a given open source family type applied to a given business motion) and what to do about it (e.g., ask these questions of the business, and if within specification, deliver this guidance). They built a systematic process for identifying the gaps in their data and processes and created mechanisms to constantly convert their known-unknowns to known-knowns. They also partnered very closely with their business clients to develop a shared taxonomy and set of procedures that work across internal organizational boundaries.

The open source practice’s common taxonomy process and framework was the basis for creating tooling that scales and accelerates the business and legal processes. It pushes as much work as possible into rule-driven workflows that meet the end-user customers where they are. The first two generations of open source tools pulled engineers out of their native tools and created friction. The third generation was built by the business with constant input from the legal team and was native to the engineer tool space, but still had substantial manual elements. The fourth generation turned the tool into a background process analogous to other standard elements of the build and testing process for most scenarios. It runs with minimal human intervention unless there are exceptions.

The success of the open source practice required business investment to solve the system-level problem.  It took years to create buy-in that (a) the work had value, and (b) needed to be a business native solution that includes continuing architecture and expertise input role for legal. Yet, the result was a 1000x gain in productivity.

Today, the open source team spends most of their time looking for unknown unknowns and working on known unknowns. They are a small team that supports the work of tens of thousands of engineers. This leverage is built upon a disciplined approach to avoiding duplication of effort. They structure the way they ask for work and the delivery form of work product from outside counsel to ensure optimal re-use with minimal work translation effort by their team. They spend more of their time policing the frontier because they did the foundational work of installing traffic lights in town. They are simultaneously pioneers and town planners. See Simon Wardley, “Pioneers, Settlers and Town Planners,” Pieces or Bits, June 4, 2012 (role archetypes for scaled value creations); Simon Wardley, “On Pioneers, Settlers, Town Planners and Theft,” Pieces or Bits, Mar 13, 2015 (how to foster the archetypes).

Some believe this scaling model cannot be applied to other types of legal work. They presume the nature of the open source pattern’s legal work creates implicit work standardization, and the interaction with engineering clients increases tooling automation potential. But the pattern can be applied to so many kinds of practices if the legal professionals can get past two challenges.

  1. Status and identity.  Some legal professionals view these implications of process and mechanization as a threat to their status and identity. Many legal professionals take pride in being craftspeople with the exquisite capability to deal with nuance, context, and ambiguity. And they should. But being able to describe the decision tree and operations that produce an outcome does not diminish the impact of the outcome produced. Some of those operations are complex and can only be performed by skilled experts. Some of those operations are highly context and judgment dependent and can only be performed by experienced experts. It is often the most adept who can describe complicated things in simple terms, and sort matters as novel or routine.
  2. Critical thinking swamps system thinking. Some legal professionals struggle to see the patterns in their work because they developed strong critical thinking skills to the exclusion of system thinking skills. They are so adept at distinguishing and advocating for why things that seem the same other are different that they are challenged to observe and describe patterns based upon similarity. See Post 080 (discussing tension); see also Herman Kahn, “The Expert and Educated Incapacity,” Hudson Institute, Jun 1, 1979 (expertise can blind us to solutions outside our own discipline). Design thinking exercises that bring together groups of experts facilitated by allied professionals can be invaluable. Exposing legal professionals to these ways of thinking and skills earlier in career changes their potential to create an impact because it accelerates their progress to change-makers. And because most legal work is a form of knowledge work, see Post 159, we can import and build upon general knowledge work concepts to build an investment strategy that can balance cost, risk, and talent over the long term.

The graphic below attempts to summarize the enormous skills and mind shift needed for us to transition to big-picture systems-level thinking and problem-solving.

A learning and development process that helps people embrace systems thinking.

Because of a lack of investment in foundational training, we often get stuck at the “collaborate effectively” phase. Obviously, this is a transition destined to last a full generation or more.

System dynamics shape investments that create the future practice

The environments in which legal services will be delivered are increasingly volatile, uncertain, complex, and ambiguous (VUCA).  Building upon the Bennett and Lemoine VUCA approach, we can create an investment framework that helps our organizations and institutions systematically adapt

We can define a problem space that describes the characteristics of legal issues in relation to how much we know about an issue (information) and what we know to do with an issue (approach). These defined spaces for volatility, uncertainty, complexity, and ambiguity produce a prototype virtuous cycle solving sequence that can be used to develop and sequence a strategy for systematic adaptation. Such a prototype might look like this:

  1. Take what you know how do to, identify and extract the processes, refine them, and instrument them to start producing data from the work.
  2. Feed the data from your upgraded processes into an intelligence mechanism that classifies, derives insights, and shares information broadly with your organization and partners.
  3. Build processes that influence the behaviors of your people and collect data on the outcomes, and feed what is learned into process redesign and intelligence mechanisms.
  4. Use the efficiencies gained from optimizing activity for impact to redeploy resources as reinvestment, and offer your investment to others who have resources deployed to underperforming assets.

The graphic below is a visual depiction of this suggested prototype:

An investment cycle framework for addressing volatile, uncertain, ambiguous, and complex issues.

The path to structure and process invites thinking about scale with the premise that expert human attention is the primary constraint on scaled legal work. [emphasis added by editor. Go ahead and read it a second time. It’s that important. wdh]

The future practice scales

Growing demands for legal services can be satisfied by using approaches that scale, which in turn manages the problem of growing complexity by applying structure. This typically means providing services that deliver adequate quality, faster, using techniques that consume fewer scarce resources, such as the time of experts.

A conventional approach to scaling a practice by applying structure might look like this:

  1. Experts teach novices enough about what to look for in the work and what to do when found to make them generally work-capable within mostly known scenarios with continuing expert oversight. This might create 10x leverage on the expert.
  2. Expert and capable workers partner with process experts to produce instructions that allow humans with very limited understanding of the subject matter to operate within fully known scenarios with continuing expert and capable oversight. This might create 100x leverage on the expert.
  3. Process experts partner with the expert workers, capable workers, and tools experts to build machine powered capabilities based upon the information observed through the instructable workers with continuing expert, capable, and process oversight. This might have no upper bound on the expert leverage.

We need to get from expert-delivered to machine-delivered experiences faster. The conventional approach is challenging because it has high coordination costs and often requires selling the experts on the value of investments that may not be aligned with their personal interests (subject matter interest and prestige) or economic incentives. See, e.g., Post 051 (Jae Um noting that legal innovation is hard because the ecosystem is complicated and actors are obscured); Post 010 (discussing how professional service firms, even in the managed services space, must balance interests of clients and talent to achieve both profit and long-term stability).

Organizations that want to accelerate legal innovation with a radical approach will hire and/or retain subject matter experts who have some process and tool capabilities that tighten the innovation loop. The graphic below provides a preliminary sketch of how such an approach could be conceptualized and built.

Scaling legal work requires complementary skill sets.

The above model might work in theory, but what about in practice?  Another part of our wicked problem is the alignment of both intrinsic and extrinsic incentives with innovation behaviors.  Getting this design element right is no less tricky than a complex workflow or a sophisticated piece of technology. Cf. Post 203 (noting how one-to-many solutions require the creation of new business models, as the technology already exists).  William Henderson, “The Legal Profession’s Last Mile Problem,” May 26, 2017 (same).

That said, in pursuit of scalable legal work, we should not try to turn legal professionals into software engineers. Success is producing legal professionals capable with basic process and design thinking approaches and the emerging set of tools that can produce software solutions without writing code. We are capable of learning new tools when we have proper incentives and support. Indeed, there was a time most attorneys relied upon secretaries to type for them.

Organizations that make these investments create an option to disrupt existing businesses, including their own. See Clayton M. Christensen, Michael E. Raynor, & Rory McDonald, “What Is Disruptive Innovation?,” Harv Bus Rev, Dec 2015 (addressing the underserved bottom of a market creates growth trajectory that can reach upmarket and dislodge entrenched providers). But this “disruption” does not have to be destructive for the value or experience of experts. Rather, it can give them more of what they want.

The future practice explores

Most attorneys probably went to law school seeking an aspirational career outcome. They wanted to be intellectually stimulated and produce work that creates a beneficial impact. They did not seek a practice of repetitive tasks, acting as human information routers and middleware, and otherwise spending their time on the realm of known-knowns that do not demand their analytical skills.

Experts spending so much time with the known is a perverse outcome of needing more information to adapt to increasing uncertainty. The tedious gathering of that information by lawyers is justified (or rationalized) based on reducing or managing risk, albeit framing it this way does not make the work more enjoyable for the lawyers.

We spend more of our time on this mundane work because we do not wrap our work in structure, systems, and tools that shape work and feed back refined information. Instead, we become the system that constantly translates inbound and outbound work, even when that work is well known for type and context. Attorneys will try to push the work they do not want to do to other humans with different skills. Delegation is critical. But the delegation must happen in a way that supports the future practice.

Work and information must land in a machine-supported state that creates shared knowledge. When we transact information solely with humans, we are subject to the limits of our collective human cognition. This includes our limited abilities to reason across massive amounts of information, the speed with which we can service requests, and our limited capacity to switch across information formats and experiences. We see the normal human response to these burdens when people do not share information and do not consume information.

The graphic below depicts the system friction inherent in our information flows.  Along the left and right sides are four hypotheses that illustrate some possible reasons why this is a very wicked problem to solve.

A surface plot highlighting communication switching costs.

We can frame the investment pattern for systematic adaptation that builds the future practice in a way that (a) elevates the work legal professionals crave and (b) creates less identity threat and incentive misalignment.  Indeed, we want legal subject-matter experts exploring the frontier of their practices looking for unknown-unknowns and solving for known-unknowns. See Post 071 (Bill Henderson describing investment in substantive law (Type 0) and service delivery (Type 1) and how rarely new opportunities are pursued);  “Clive Gringras discusses an International Legal Practice Built for the Future,” Business of Law Podcast, Apr 2, 2020 (how to learn clients’ business and preferences to identify work and shape innovation investment).

The investment pattern asks experts to do something that is both unnatural and difficult for them:  work within a process that reduces the concepts that identify issues and the principles that govern how they address issues into increasingly concrete and granular descriptions that can be handed off to actors with decreasing context and judgment capacity.  In its highest form, work is delegated into the machine realm. This investment pattern and evolution gives experts more time to roam the frontier pursuing the most interesting work and brings them data-powered insights that help them cover more ground faster. This systematic approach also unlocks a higher tier of the future practice that takes it from data-informed to data-driven.

Below is a graphic that illustrates how this type of systematic adaptation might work in practice.

Creating scale on legal work involves systematic adaptation that focuses on conserving the attention of the most skilled humans.

The future practice experiments

The future practice will support organizations that systematically adapt to pursue ambiguous opportunities using experimentation.  Indeed, these experiments are, in essence, how the organization (i.e., our client) gathers the data necessary to develop a high-value/low-risk strategy. Below is a graphic that illustrates the competitive advantage inherent in this approach.

Legal work can be scaled with the combination of knowledge management and systematic experimentation.

It is unrealistic to conclude that experimentation will not diffuse into the legal function. To prepare the way,  we need to adopt and support similar frameworks. A knowledge management (“KM”) process and system that builds upon machine intelligence are required to make this performant (computer jargon for “good enough”) and scalable.  In turn, by solving individuals’ problems, these systems will earn the access necessary to create collective knowledge. See, e.g., “Ballard Spahr’s COVID-19 Legal Response with Digital Transformation,” Business of Law Podcast, June 22, 2020 (practical approach to creating a scaled information resource).

In the years to come, we will see KM blossom as machine learning reduces adoption costs by meeting people where they are—in the communication layer. What does this mean? The ground truth of what is happening in an organization is probably best defined in the communications that flow among the organization’s people. Systems of record and formalized documents often reflect dated information because they require a person to break out of the inner loop and tools that power their work. Working in these secondary systems is expensive because they require an additional layer of skills to use them and context switching to access them. People minimize their tax by batching their engagement, which in turn creates delay and inaccuracy.

To operate within the communication layer, we must invest to build domain-specific, natural language, machine learning models. Legal professionals use a combination of practice, customer, and organization-specific jargon in their work. Unsupervised learning (a specific type of machine learning) offers promise for the future, but without help, current off-the-shelf models may not work well in these specialized environments. To bridge the gap, we will use commodity tools to train models that can operate in subject-matter experts’ native tools and perceive intent and signal accurately enough to act with confidence. See “Tagulous Demo for CLOC Las Vegas Institute 2019,” Business of Law Podcast, May 14, 2019 (prototype approach to capturing actionable signal from email using machine learning that supports KM built on commodity cloud services). Earning the privilege to act within the expert’s primary toolset is necessary to transition from data-informed to data-driven.

The future practice is elegantly data-driven when supporting systems of intelligence inform an actor with the least amount of information necessary to make a better decision before they are committed to a course of action.

My commute provides an example of what this looks like in real life. Shortly before I choose between two routes, I get specific information on a road sign that updates with current commute times of my options. This is exactly what I need to make a choice with better outcomes. As our approaches advance and mature, we will produce more experiences that inform while also taking extreme care to preserve scarce human attention.

The future practice is fully connected

To operate at the scale that unlocks the greatest potential of our legal infrastructure, the future practice need standardization across organizations that enables more machine supported patterns that help us connect our work.

The image below models a standard interaction pattern for legal work. Imagine a basic request for some form of contract review. It begins with a business customer, migrates through internal counsel, emigrates to outside counsel, and ultimately round trips back to the customer. This kind of interaction is typically transacted through a series of point-to-point email communications that may include attachments.

Scaling legal work with augmenting machine support.

We will build standard definitions for frequent, known-known work requests that operate inside and across our organizational boundaries. This will support the adaptation and KM investments by making it easier for machines to assist work as it moves through enterprise information systems. When we build open standards for work definitions, we can (1) transact knowledge across organizational boundaries with partners more efficiently, (2) measure work quality more easily, and (3) promote re-use by avoiding duplication of efforts. See Standards Advancement for the Legal Industry (SALI) Alliance (focused on developing open industry standards).

We also need a set of legal industry-focused application programming interfaces (“Legal APIs”) that allow us to start building machines in all parts of the ecosystem that can efficiently throw and catch work with different combinations of customers and partners. This is a critical investment that will bring speed, efficiency, and better decision-making, at-scale. And it will help our profession evolve into an industry that continues to provide the critical infrastructure society needs.

The future practice pattern will create more value through access

My perspectives are informed by commercial concerns, but the patterns can be applied to many types of legal services for many types of customers. Enterprise and consumer. Commercial, public sector, and public good. We will create more value if we can create more commonality across our respective needs. I believe we will get there. It is a question of path and pace. But the journey has already started.