Philosophy
The selfdriven Legal Framework begins with a rejection of the retrofit model. Most law firms approach AI as an enhancement layer — tools bolted onto existing pyramidal structures, billable-hour economics, and paper-based trust. The results are predictable: marginal efficiency gains, unchanged risk profiles, and no structural advantage.
AI-native legal practice means something fundamentally different. It means designing the practice around what AI agents do well — volume, consistency, speed, synthesis — and reserving human capacity for what only humans can provide: judgment, ethics, advocacy, and accountability. The organisational structure, the credential architecture, the pricing model, and the client relationship all follow from this inversion.
The distinction that matters: AI-enabled firms add technology to existing structures. AI-native practices are built so that every workflow, every credential, and every client interaction flows through intelligent systems — with humans conducting the orchestra, not managing the instruments.
Core Principles
Six principles govern every design decision in an AI-native legal practice. They are not aspirational — they are structural constraints that shape everything from agent delegation to billing.
1. Agents Execute, Humans Conduct
AI agents handle volume, consistency, and routine judgment. Human conductors exercise professional responsibility, strategic thinking, and ethical accountability. This division is absolute — it is not a spectrum.
2. Identity Before Everything
Every party, agent, credential, and action is anchored in cryptographic identity. Without a verified KERI AID and an ACDC credential establishing scope, no action proceeds. Trust is architectural, not procedural.
3. Outcomes Over Hours
When AI eliminates the labour content of routine legal work, billing for that labour becomes indefensible. The practice prices on scope, value, and outcomes — never on time spent by agents.
4. Radical Client Transparency
Clients have real-time access to matter dashboards, agent activity logs, verifiable provenance of all documents, and cost tracking. Opacity is not a feature of legal service — it is a legacy liability.
5. Access as a Design Goal
AI capacity reduces the marginal cost of legal work toward zero. The practice actively directs this capacity toward underserved communities — access to justice is not a pro bono gesture, it is a structural outcome.
6. Accountability is Cryptographic
Every consequential action — a filing, a delegation, a signature, a revision — is recorded in an append-only key event log. The audit trail is not a document; it is a mathematical proof.
AI-Enabled vs AI-Native
The difference between AI-enabled and AI-native legal practice is not degree — it is kind. The following comparison maps the structural divergences across the dimensions that matter for practitioners and clients.
| Dimension | AI-Enabled (Traditional) | AI-Native (selfdriven) |
|---|---|---|
| Organisational Structure | Partner-associate pyramid with AI tools | 8 Areas of Focus, human conductors + agent teams |
| Identity & Trust | Email, passwords, wet signatures, paper trails | KERI AIDs, ACDC credentials, passkey signing |
| Pricing Model | Billable hours (augmented by AI efficiency) | Fixed-scope, outcome-indexed, value-based |
| AI Role | Research and drafting assistant | Delegated agent with scoped, cryptographic authority |
| Audit Trail | File notes, version history, email chains | Append-only KEL, witnessed, mathematically verifiable |
| Client Visibility | Status updates from lawyer, invoice at end | Real-time dashboard, agent logs, verifiable provenance |
| Access to Justice | Pro bono as optional add-on | Structural outcome of AI cost reduction |
Practice Layers
The AI-native legal practice operates across five distinct layers — each layer dependent on the one below it. Identity is the foundation; client outcomes are the surface.
Cryptographic Identity (KERI/ACDC)
Self-certifying AIDs for every practitioner, client, and agent. ACDC credentials establishing scope, authority, and role. The mathematical root of trust from which everything else derives.
Agent Infrastructure
AI agents operating under KERI-delegated AIDs — scoped to specific matter types, actions, and time windows. Agents cannot exceed the authority their conductor delegated. Every action logged in the KEL.
Organisational Framework (8 Areas)
The practice structured across 8 Areas of Focus — each led by a human conductor, each served by a team of purpose-built AI agents. Strategy, engagement, operations, compliance, and culture all mapped to this structure.
Matter Workflows
AI-orchestrated pipelines for intake, research, drafting, review, filing, and delivery. Each workflow anchored to a matter ACDC credential and logged against the client's AID. Conductors intervene at defined decision gates.
Client Outcomes
Verifiable documents, transparent pricing, real-time matter visibility, and cryptographically provable chain of custody — delivered at a cost structure that serves clients at every economic level.
The Conductor Model
The Human Conductor is the central organisational concept of selfdriven.legal. Each of the 8 Areas of Focus is led by a conductor — a senior practitioner or specialist who orchestrates the AI agents in their domain.
What Conductors Do
Conductors dedicate approximately 70% of their professional energy to their primary Area of Focus, and 30% across supporting areas. They exercise the judgment, professional responsibility, and strategic thinking that AI agents cannot provide. They sign consequential outputs via WebAuthn passkey — creating a legally meaningful, cryptographically verifiable KERI interaction event.
What Agents Do
Agents handle volume, consistency, monitoring, synthesis, and routine execution. They operate under KERI-delegated AIDs — their authority is explicitly scoped to the matter type and action set their conductor authorised. An agent cannot exceed its delegation.
Conductor Registry
All conductors hold KERI AIDs with ACDC role credentials issued by the selfdriven.legal entity. Their appointment is recorded in the entity's KEL — making conductor changes verifiable and auditable without any central administrator.
Applying the Framework
The selfdriven Legal Framework is jurisdiction-agnostic and practice-area-neutral. It provides the organisational architecture; practitioners adapt the regulatory mapping to their context. The following guidance applies the framework across common legal verticals.
Commercial Law
Contract drafting agents, due diligence automation, regulatory compliance monitoring. Conductor focus on deal strategy, risk assessment, and counterparty negotiation. Outcome pricing based on deal size and complexity.
Litigation
Discovery review agents, precedent research agents, submissions drafting. Conductors lead on courtroom strategy, witness management, and tactical judgment. ACDC-anchored filing records create unambiguous procedural chain of custody.
Access to Justice (Community Legal)
AI agents handle the high-volume intake, issue identification, and initial advice generation that makes community legal centres economically impossible under the billable-hour model. Conductors supervise and sign off — agents serve the 92% of civil matters that currently receive inadequate help.
Regulatory & Compliance
Monitoring agents track regulatory changes across multiple jurisdictions in real time. Compliance review agents audit documents against current rules. Conductors translate regulatory intelligence into strategic advice and manage regulator relationships.