How can we keep up with exponential increases in demand and complexity? Invert the pyramid.
Bill Henderson once advised me not to use the term “industrialization” to describe changes in the legal profession to attorneys. It offends us, and we disengage. But I titled this field note “industrial evolution” because we must embrace industrialization as a necessary and valuable part of our transformation that will elevate the value of our profession in a digital age. Cf. Post 231 (Henderson breaking his own advice for the same reason, comparing legal to the early days of the auto industry).
This post is part of a series that reflects my legal industry learning journey, building upon my career journey (080), professional evolution (143), focus on knowledge work (159), and future practice design theory (210). This installment examines the changes happening now that require us to evolve to serve a civilization experiencing exponential change powered by the fourth industrial revolution, and how we might get there faster, together. See Erik Brynjolfsson & Andrew McAfee, The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies (2016) (cognitive automation will produce creative destruction).
This post was drafted proximal to the College of Law Practice Management’s 2021 Futures Conference, which offered expert commentary on the information work industrialization megatrend and strongly influenced the thesis presented here: that we are experiencing accelerating change as a secular trend.
As discussed below, my tentative solution is, in part, to invert the law’s traditional pyramid structure. None of this is likely to make much sense, however, without first describing the set of challenges before us.
That is not the tide coming in. The seas are rising.
If you perceive your organization is being asked to do more with less, you’re probably right. Prioritization and cost control can deliver incremental improvement and better is good. See Casey Flaherty, “The Savings Trap – The Value of Value Storytelling (#2),” 3 Geeks and a Law Blog, Sept 19, 2021 (our goal is delivering value, not savings); cf Casey Flaherty, “The Limits of Incremental Improvements,” 3 Geeks and a Law Blog, Mar 10, 2020 (the first improvements are typically the highest return on investment).
Yet, if our organizations are going to keep growing at an accelerating rate, these approaches are not good enough. We will need more capacity and new capabilities that mere conservation of resources cannot deliver. Our evolutionary path needs improvements on how we do work today that range from whole-multiples to orders-of-magnitude. This is transformation.
Transformation is system-level change. Transformation can start with point solution investments. These help us explore, validate hypotheses, and create investment cases. But system effectiveness is limited by the least capable component upon which the system depends. If your computer has a very fast processor but inadequate memory and slow networking, the processor cannot contribute at its full capacity. Legal work has a similar challenge. We need many elements to evolve simultaneously to produce coherent complements. We must start walking in the same direction to transform.
We can ground our transformation imperative in the information that explains what many of us in the commercial sector are feeling now. For example, law departments are targeting 16% in cost savings and predict workloads to increase by 25% on average over the next three years. See EY Law & Harvard Law School Center on the Legal Profession “The General Counsel Imperative: How do you turn barriers into building blocks?”, Apr 7, 2021 (NB: Cost savings of large and small departments averaged for modeling purposes here); see also Post 216 (Chief Legal Officers increasingly reporting spend control).
As shown in Figure 1 below, treating these as compounding curves and projecting them out suggests law departments must manufacture almost 4 times our current leverage within the decade. I believe this understates the necessary leverage and offer reasons why below. See Post 216 (“In the next decade, both buy- and sell-side will need to do 5x to 10x more with less”).
These projections are inaccurate, imprecise, and very useful. They highlight the direction of travel and probably understate the demand growth for sectors that will be increasingly powered by machines that are getting more capable at an exponential rate. See HBR, “Growth and the Energy Transition with Vaclav Smil”, Azeem Azhar’s Exponential View, Season 4, Episode 9, Nov 27, 2019 (information industries can experience growth indexed on Moore’s Law and physical industries will experience growth at a few percent per year).
Saying “we have more work” does not help us adequately understand our challenges and opportunities. To explain how and why that work is changing we must start with a simplified model of what commercial legal work is and how it creates value today.
Our work will be harder and more interesting.
Legal work creates value by helping clients optimize decisions and actions that are impacted by legal and policy constraints or depend upon legal or policy infrastructure. Legal work can require the skills of a researcher, detective, game theorist, behavioral economist, philosopher, preacher, advice columnist, and poker player. But when we unromantically strip legal work down to its component parts, as in Model 1 of Post 159 (Model1), we are left with something like the following list:
- Understand the client’s objective.
- Collect and analyze the starting facts and known constraints to simulate options and outcomes.
- Formulate advice that takes the client’s objectives and context into account and probabilistically produces the highest value outcomes.
- Drive specific activities within the selected approach that require the legal professional’s franchise.
Framing client objectives is getting harder because we operate in complex, multi-stakeholder environments. Client value propositions are becoming more holistic and less linear. How an actor generates, allocates, distributes, and reinvests surplus-value in free-market scenarios is changing because an increasing set of stakeholders can hold them accountable for their impacts on the larger web of society. Historically “stakeholders” focused on customers and a few policing authorities. Today, however, it often includes employees, partners, suppliers, NGOs, political action committees, influencers, and more.
For example, diversity, equity, and inclusion (DEI) and environmental, social, and governance (ESG) are categorical examples that reflect megatrends and highlight how this influence operates in conventional and emerging ways. Likewise, the sovereign can limit entities’ freedom-to-operate. Employees with scarce talent can choose to work elsewhere. Customers with transferable needs can work with competitors better aligned with their values.
Optimizing for outcomes was easier when businesses could ignore certain impacts of their choices as externalities when designing their objectives. Objectives are now defined by and balanced across more dimensions. Legal professionals are increasingly asked to look beyond conventional “legal risk” to help optimize “enterprise risk.” See, e.g., Leaders of Legal series, In-House Focus (General Counsel and Chief Legal Officer manage to business outcomes).
Managing the information and constraints that are used to build the simulations that underpin legal analysis is getting harder because information volume is growing at an exponential rate and constraints are growing fast and incoherently. The amount of information in the world is increasing at more than 19% compound annual growth rate. See Arne Holst “Total data volume worldwide 2010-2025,” Statista, June 7, 2021. US Federal Statutes grew by 51% from 1994 to 2018. See Press Release, “Chicago-Kent Professor Measures Quantifiable Growth of Modern Law,” July 6, 2021. See also Post 216 (increasing regulatory burden will drive work). Of course, jurisdictional requirements can directly conflict. See Greg Baker, et al., “Best Practices for Responding to Subpoenas That Conflict With Foreign Data Privacy Laws,” JD Supra, Mar 27, 2020 (conflict between United States discovery and EU data protection laws).
We are trying to process more information through a more complex analytical framework. And many of the most compelling opportunities for clients to create value are in increasingly regulated spaces that combine industry and subject matter regulatory constraints. We are moving towards more complex work because that is what is required to access the next frontier of value.
Integrating with client decision-making processes and scaling to demand is getting harder as the business substrate converts to digital and the legal substrate remains analog. Modern organizations are increasingly data-driven. This requires influencing business decisions with data and hypotheses that help ask the right questions. These businesses can generate an astounding amount of work volume built upon machine-based leverage as they transform. See Post 210 (open source practices result in 1000x gain in productivity).
Because our clients are pursuing value creation in increasingly complex business spaces accelerated by machines, the net impact on legal includes:
- Increasing work volume (number of units)
- Increasing complexity (unit cost)
- increasing velocity (unit clearing time).
Demand will accelerate because many clients are chasing an exponential growth curve. And I suspect we are underestimating the true cost of demand growth because we are not factoring in increasing complexity and the requirements for rapid clearing times. Increasing complexity typically requires more effort to do the same work because the solutions are harder. And faster unit clearing times typically require more peak capacity to allow parallel effort.
Translated to the physical realm we are trying to operate a modern city with 19th-century transportation infrastructure. This will not work.
I expect 10x leverage scenarios by the middle of the decade driven by demand with increasing complexity and velocity characteristics. As noted earlier, my prediction is inaccurate, imprecise, and still useful. It presents an outlier opinion that is possible because machine intelligence is getting smarter much faster. See Julien Simon, “Large Language Models: A New Moore’s Law?”, Hugging Face, Oct 26, 2021 (machine learning models are growing very fast, and well-deployed smaller models are probably good enough).
Our experiences suggest we do not need machines this powerful to do useful things that help legal professionals be more effective. We can build the leverage we need by thoughtfully applying and combining the commodity tools we have now. So why are we stuck?
Our future requires industrial scale
I believe our industry’s evolution moves slowly because our profession is controlled by people who are comfortable with disorder. They are comfortable because their primary training and adaptation occurred at a time that did not require scale. Cf Post 233 (arguing that Planck’s principle applies to law, thus limiting the pace of change). Yet, we are now a trillion-dollar industry with no SKUs, few standards, no quality measures, connected by phone/email/parcel delivery/fax. See Daniel W Linna, “Evaluating Legal Services: The Need for a Quality Movement and Standard Measures of Quality and Value”, March 11, 2020 (draft book chapter).
In his book, Designing Dynamic Organizations (2002), Jay Galbraith, one of the world’s most renowned organizational design consultants, summarizes how different types of organizations create value. In my view, the following passage offers the evolutionary path we will likely follow to industrialize and scale.
In order to identify the capabilities most important for your organization, you first need to determine what type of value proposition your strategy offers…
In general, organizations try to differentiate themselves through a product, operations, or customer focus…
Product: A product-focused company not only creates the best product in the industry, but creates products and services that buyers may not even know they need. Product companies focus on innovation and new product development. Their advantage is not just the products themselves, but their product development processes that allow them to get new products out to the market faster than competitors…
Operations: Operationally excellent companies deliver a combination of quality, prices, and ease of interaction that others can’t match. They may not be first in the marketplace with new products and their products or services may lack some of the features of competitors. They do promise value—whether measured in cost, convenience, quality, or consistency of experience—which is seen by buyers as more important than other product or service attributes…
Customer: Customer-centric companies build long-term customer relationships by tailoring the products and services they offer. They provide total solutions to customer needs rather than stand-alone services or products. The professional services industry (e.g., consulting, law, accounting) provides classic examples of customer-centric organizations … . The more the firm understands its customers, the better solutions it can provide. Further, customers don’t typically buy these services based on the lowest price or latest methodology, but on trust and quality of the relationship…
Each strategic focus implies a different culture and type of person, and different key processes and measures to execute it successfully.
Jay Galbraith, Diane Downey, & Amy Kates, Designing Dynamic Organizations (2002) at 26-28. To create scalable practices, we should move more experiences from services, to operations, to products. This approach can create far more value from limited subject matter expertise for a broader range of clients and customers.
This evolution requires new investments—investments that introduce structure and sameness to similar work types. The legal profession is mature. Why haven’t we evolved to more sameness already?
My own conclusion is that is time to invert the pyramid. See Figure 2 below.
Legal services must evolve beyond luxury goods
Legal professionals historically command a premium for offering our customers a luxury experience that accommodates high entropy work. See Kanye West, Steve Stoute, & Ben Horowitz, “Technology, Culture, and Consumer Adoption: Learning to Read the Cultural Landscape,” YouTube, June 23, 2014 (Kanye West observes that time is the only luxury).
Entropy has different formal definitions depending upon the scientific domain (thermodynamics, physics, information theory). The concept generally reduces to disorder that limits effectiveness. Consumers of legal services, who can gain access, pay more because we legal professionals rarely explain the tradeoffs that would reduce our shared costs and let the work move “downmarket” if the customer would embrace more structured approaches to their interactions.
The current state of disorder is experienced as a feature, not a defect, by the controlling actors (the people who choose the internal and external routes for legal services work). Buyers who can pay for the legal services work they need feel cared for. Further, they have not historically asked for options or alternatives to what is being offered—what’s better than a skilled, smart expert working diligently to get you to a solution? Likewise, the Sellers get their market-clearing rate on their time, regardless of the utility of their services. Sellers have not historically offered bargaining options because it has not served their self-interests.
The need for more leverage is changing preferences. Yet, as our preferences change, we still feel stuck. Why?
Exquisite service lets us push the work of sorting out the chaos to our counterparty so we can focus on other things. Buyers of conventional legal services offered by law firms enjoy local utility maximization because our idiosyncratic preferences are accommodated—after all, we’re buying a de facto luxury good. The effort to understand and structure our work with undocumented features that create value for us (i.e., our preferences) is pushed to the seller and turned into a durable asset in the form of “know-how.” See IBM Cloud Education, “What is Knowledge Management,” Nov 12, 2020 (tacit and implicit knowledge are valuable but may be hard to store or share).
Remarkably, repeat interactions don’t change the dynamic. The Seller of the service enjoys a stickier relationship with rent-seeking opportunities because the undocumented specification of the work creates switching costs for the Buyer. Buyers must drive the creation of better-defined interfaces to get where we need to go, which is bound to feel like an unnatural act. See Post 219 (highlights the buy-sell system dynamic required for sustainable innovation); see also Post 040 (In-house and outside counsel are caught in the same culture loop).
So, what attracts us to the current state?
The action is the juice, but we’re getting squeezed
I suspect legal professionals thrive in higher entropy environments because of the traits our profession selects for, trains, recognizes, rewards, and promotes.
The strengths of conventionally competent legal professionals may be less likely to produce architected systems-approach to legal work. If we plot a sample of traits on a strengths-weaknesses ledger the pattern emerges.
Trait | As a strength | As a weakness |
Critical Thinking | Able to distinguish between things that are similar. | Struggles to reduce similar things to patterns. |
Client Service Focused | High attention to client needs. | Functions beyond boundaries of expertise. |
Bias for Action | Responds to request with urgency. | Does not prioritize the important over the urgent. |
Comfort with Ambiguity | Applies judgment to make decisions with limited information. | Does not press for more rigorous and structured inputs for decision making. |
Figure 3. A strengths-weaknesses ledger view for common traits of legal professionals.
Our industry’s culture has not historically attracted or accommodated the competing traits of systems thinkers, and a known exception supports the generalization.
Intellectual property prosecution often draws talent from the STEM disciplines. These professionals often bring a higher degree of systems thinking traits with them. See, e.g., “Isabella Fu talks about decision making and leadership,” Business of Law Podcast, Dec 26, 2018 (An elite critical and systems thinker attorney with a physics undergraduate degree). People who focus on other practices may presume that the systemization of intellectual property is unique to the underlying subject and cannot be imported to other legal work. I disagree. Patent prosecution is a blend of fact gathering, research, pattern matching, drafting, and advocacy wrapped in a repeatable process.
Many legal professionals struggle to see their work as something that can be patterned and systematized because “things are different every time.” Yet, I am hearing less of this fatalism in my work, and I have a theory.
My colleagues’ work volume is forcing us to think differently. We support a business (Microsoft) that is intentional in its exponential journey. It is forging paths for ecosystem value that generate demand for legal services breaking through our learned patterns. And my less tenured colleagues seem to think differently about their work. It is not because they are superior technologists. See Darth Vaughn & Casey Flaherty, “Tech comes naturally to ‘digital native’ millennials? That’s a myth,” ABA Journal, Oct 13, 2016 (technology proficiency is a learned skill, not a talent). Rather, they want to have more balanced lives and, faced with a relentless onslaught of work demands over a longer time horizon, they are willing to try different things to figure that out.
In the short term, this often means working harder and taking more risks on investments. This is evolution. We cannot afford to waste their energy and intrepidity. We are thinking about how to build the experiences we need to support these professionals who serve an exponential business within our primary constraints.
Solving for our constraints = the foundation of our future
Within our emerging knowledge management experiences, we use design principles that serve our people and objectives. See Post 159 (Model 2 shows a generalized knowledge work process model). These design principles flow from our greatest scarcity: the time of subject matter experts.
Thus, it follows, to conserve and grow value:
- Professionals should not have to answer the same question twice.
- Professionals should not be routers for known work.
- Professionals should remain in tools they know with minimal training or switching.
There are also core fragments of information work that can be systematized and machine-assisted to produce industrial-scale leverage from our human work. Figure 4 below organizes a sampling of those fragments into a loose maturity progression that focuses on building-block capabilities (foundation); efficient aggregation points (platform and scale); and extensions that can help us extend our art (new categories).
We can apply the Figure 4 model to a simple scenario operating at the Foundation and Platform layers.
Imagine you are a law department that supports a business function that engages with many external parties that provide products your business might want to use. Think of this as a very low-risk procurement scenario. The business is not taking a hard dependency or paying serious money. To better evaluate the value proposition before moving to negotiated terms, they want to see what is available in the standard offerings. Further, they want to do this quickly for many potential providers. The fastest path is to determine whether the providers’ standard terms are acceptable for this scope. Your department has identified the specific terms that require more scrutiny. If those terms are absent, the business can proceed.
You have been asked to design a system that enables this scenario, including designing for minimal waste (appropriate resources, minimized duplication), high capacity (parallelism), and high adaptability (new constraints). You might do the following:
- Require all business “asks” be submitted into a structured intake that pushes information capture costs to the business owner, and requires all information to be provided before analysis commences.
- Build a knowledge base of previously reviewed scenarios (provider, terms, usage scope) and their dispositions (approved with general guidance, approved with specific guidance, and denied).
- Build a machine learning-based scoring model that associates term types that influence dispositions with natural language analogs so your machines can automatically recognize the presence or absence of terms of interest.
- Build a workflow that routes the complete information of the request, including insights from the machine learning analysis to assigned law department reviewers.
- Build a standardized interface for routing review to outside counsel that fully describes the work to be done and the response format to make it very easy for humans to process the response and to retrain your machine model.
- Build a capture process into the workflow that collects the resulting guidance produced for the clients and stores that in the knowledge base.
Figure 5 below is a process map that reflects the above six steps.
This is far from a futuristic thought experiment. The system specification above can be created today with no/low-code building blocks and commodity services. It produces a closed-loop learning system that requires less human input, goes faster, and supports more parallelism over time. This one is purpose-built for a narrow scenario, but you can imagine a generalized form that can adapt to a broader set of constraints.
For example, imagine your law department supports a business that will be subject to increased regulatory constraints. You have been asked to design a system for the same efficiencies as the narrower scenario, but you don’t know all the signals, rules, and workflows. Instead, you know there are regulations that will be published, they will have to be observed and analyzed, and the business will have to be advised. As shown in Figure 6, you produce a naïve lifecycle and focus on the phases you can address with some reasonably bounded systems approaches.
Next, as shown in Figure 7 below, you decompose the focus areas of the lifecycle into a general set of capabilities, breaking them into pieces that can be composed and reused.
You start building the pieces that address your needs. The parts that find regulation changes, route them to the right processor, and produce rules that can be migrated into the business substrate to lower the cost of efficient compliance. Eventually, you start to realize that regulations, policies, contractual commitments, and the many forms of legal constraints that apply to business have a generalized form too. Thus, they can be created as a composable framework that can be adapted to many scenarios.
Finally, you realize that building this “machine” makes the most sense if you are not the only one doing it—specifically, that having different ways of doing similar things both inside and across organizations means that our work costs more but does not create more value for the premium we pay. Indeed, our lack of standardization is an invisible but heavy tax on the time of experts.
Figure 8 below is a conceptual framework of how we might industrialize a compliance system that delivers better, faster, and auditable guidance—something that is likely experienced as a product—with much less ongoing expert time.
We must build the future to be a part of it
It is in the long-term best interests of everyone in our legal profession to actively participate in and support our industrial evolution. The machines are going to keep getting smarter. We are very far from artificial general intelligence, but machine learning models will be able to reason over natural language and decode the words we commonly use into signals that become information and instructions that benefit humans. See Post 248 (Rob Saccone noting that delegating work to technology teammates is happening). Service providers can build products and services that do this and address other customer needs along the way.
Every company that creates value from information will need a data strategy. That strategy will need an approach that produces high-quality data from human-computer interactions because humans reflect the greatest intelligence in the system. When engaging systems with exponential capabilities, it is unwise to delay meaningful participation until after everything is fully sorted, as it is very easy to be left behind. See Post 63 (Jae Um discussing truly disruptive innovation models). Likewise, services-focused organizations should be understanding customers’ (not just clients’) jobs-to-be-done and developing data and product strategies as a co-development motion. See Clayton M. Christensen, et al., T, “Know Your Customers’ ‘Jobs to Be Done’,” Harv Bus Rev (Sept 2016).
We must all share if we are going to walk in the same direction. Finding ways to contribute to organizations like Standards Advancement of the Legal Industry (SALI) Alliance is important. The critical infrastructure for our future includes the entities that allow us to come together and coordinate our effort. So we must contribute to them on terms that make it easy for people to build upon our best practices.
Although multinational corporations are awash in a rising sea of legal complexity, the industrial know-how we are developing has the inherent capacity to grow the pie for our broader society. See Post 098 (discussing how human-centered ODR systems are improving access to justice in British Columbia). I fear our profession is not creating enough surplus value through our contributions to earn our franchise. And if we fail to evolve and adapt in ways that build upon industrialization principles that serve our humanity, our collective future is bleak. We can do better.