Framework
There is a clear, well-regarded answer to what it takes to work well with AI. The catch for a law firm, or for any enterprise, is that the framework was designed for an isolated professional, rather than a scaled organization.
True AI fluency is more than basic tech literacy; it is the ability to critically direct the tool, accurately evaluate its output, and safely integrate these systems into high-stakes work. The most useful account of what AI fluency entails comes from the "AI Fluency Framework", developed by Professor Rick Dakan of Ringling College and Professor Joseph Feller of University College Cork, and published through Anthropic's AI Fluency Initiative. The program lists four competencies (known as the four 'Ds'): delegation, description, discernment, and diligence.
In plain terms: delegation is deciding what work remains with a professional and what is assigned to the model; description is communicating clearly enough to elicit high-value output; discernment is judging whether the machine's resulting output is accurate and sound; and diligence is the institution assuming ultimate responsibility for how the work is executed and for the integrity of the final work product. It is the best short answer we have seen to a question most firms never quite frame: not "which tool," but "what it takes to deploy the technology effectively."
There is one structural reality to keep in mind about this methodology, one that changes everything for a firm. The framework prescribes fluency for an individual working with AI. That is the right unit for a single professional. It is the wrong unit for a law firm.
When the four Ds are left to the discretion of individual lawyers, the result is a fragmented reality found inside most firms today: individual attorneys improvising their own divisions of labor, drafting isolated prompts, setting their own subjective quality benchmarks, and enforcing ad hoc boundaries regarding data security and client confidentiality. Some are excellent at it. Most are not, and it is impossible to distinguish competence from risk from the outside. The result is inconsistent work product, severe exposure regarding confidentiality and privilege, and effort expended without commensurate efficiency gained.
The solution is not to train every attorney to individual proficiency and hope a consistent firm-wide standard emerges. Instead, the objective is to build the four Ds directly into how the firm operates - transforming fluency from an individual talent into an institutional methodology, ensuring the same standard is upheld regardless of which lawyer executes the task. This reframing is the core challenge a firm faces, and it is worth taking the four competencies in turn.
Delegation governs which aspects of a matter remain with senior counsel, which are allocated to junior associates or paralegals, and which are routed to the model, as well as the precise protocols for shifting tasks between human and machine. Left to an individual practitioner, this division of labor is a constantly shifting judgment call. For a firm, however, it has to become a formal, firm-wide division of labor, segmented by matter type, ensuring operational consistency across the firm's hierarchy, whether a file lands on the desk of a senior partner or an entry-level associate. Crucially, this baseline structure does more than protect consistency - it creates a measurable standard for the firm to systematically analyze and optimize over time. Without a formalized process, continuous improvement is impossible, leaving leadership entirely blind to how work is actually being executed.
Defining an institutional delegation framework is the most challenging of the four competencies to construct internally. The optimal structure depends entirely on a firm's specific practice portfolio, its risk tolerance, partner compensation models, and associate development pipelines. We have written separately on why delegation is the specific competency firms tend to outsource. For a successful implementation, however, the core requirement is clear: this operational boundary must be deliberately defined and formally codified, rather than left to individual practitioners, no matter their seniority and legal acumen.
Description is the skill of briefing the model clearly and precisely enough to generate reliable and high-quality work product. Left to individuals, this competency fragments into individual, inconsistent habits. Because no centralized baseline exists, the firm cannot systematically impart its standards. Instead, incoming lawyers are left to figure out the technology through unguided trial and error.
Built for the firm, description becomes shared property: structured prompts, system instructions, and templates grounded in the firm's style, precedent, and standards. This ensures the technology is directed by the firm's accumulated knowledge, rather than the variable approach of an individual user. Critically, this institutionalization is how a firm protects its distinctiveness. While a generic tool produces competent but anonymous work that looks like any competitor's, embedding the firm's specific standards ensures the very first draft reflects the firm's signature voice, eliminating the need for heavy rewriting.
Discernment is a practitioner's ability to judge whether the model's output is legally sound, but also aligned with the firm's standards, precedent, and best practices. Yet it is this skill that firms most frequently take for granted. While AI significantly speeds up routine, boilerplate tasks, it often falters when managing work that requires strategic judgment and nuanced legal analysis. In controlled studies, professionals using AI for these judgment-heavy tasks were actually less likely to find the right answer than those working entirely without digital assistance. In the legal profession, this performance gap creates a particular risk: confident prose that hides overlooked issues.
For a firm, discernment cannot rely on an individual attorney's intuition to spot errors. It must be an explicit institutional standard that defines the exact quality thresholds for each matter type, establishes clear review protocols, and identifies prohibited use cases - ensuring quality control does not depend on a specific practitioner's chance observation. In practice, this requires benchmarking the model against the firm's historical work product before deployment, grounding the standard in empirical evidence rather than subjective impressions.
Diligence requires the firm and its practitioners to accept accountability for how technology is deployed and for the outputs it generates. Every deliverable issued under the firm's banner remains the sole responsibility of the organization; ethical obligations, factual accuracy, and legal liabilities cannot be outsourced to a machine. At the individual level, this is a matter of personal professional ethics. For the institution, however, it demands formalized risk governance: a comprehensive acceptable-use policy, strict protocols for safeguarding confidentiality and privilege, an explicit posture on client disclosure, alignment with professional conduct rules, strict compliance with malpractice insurance underwriting standards, and a designated owner responsible for updating these frameworks as technology and regulations evolve.
Because ad hoc usage lacks institutional oversight, it is where diligence most reliably fails, leaving the firm exposed to errors that carry a long malpractice tail. This burden of accountability belongs entirely to the institution, not the individual.
When examined together, the four Ds reveal a required operational sequence. A firm cannot accurately describe a workflow until it has determined what to delegate, nor can it establish a validation bar until the precise demands of the work are understood. And ultimately, diligence must safeguard the entire process.
Effective deployment requires a precise order of operations: codify how the firm actually practices, then fit (delegation), brief (description), validate (discernment), and govern (diligence) the tool to match that reality. The four Ds offer a clear blueprint for why practice must always come before platform, and why building fluency into the institution, rather than scattering it across individuals, is the ultimate goal of a successful deployment.
← Back to insights See the steps, in order →Personal fluency comes and goes with whoever is at the keyboard. We build the standard into how the firm practices, so it holds across the firm. A discovery call is a good place to start.
Schedule a discovery call →
Codified Counsel