Built for ABA Formal Opinion 512.

• Model Rule 1.1

  Competence

  Understand what the tool can and cannot do, and verify its output.

  • JILL works from your case file and the controlling authority, with citations validated before you see them.
  • Outputs are confidence-scored, so you can see what needs a closer look.

• Model Rule 1.6

  Confidentiality

  Protect client information when it goes into an AI tool.

  • US-based infrastructure, encryption in transit and at rest, and SOC 2 Type II and HIPAA compliant controls.
  • Role-based access, restricted matters and ethical walls enforced in the platform.
  • No training of third-party AI models on firm data.

• Model Rule 1.4

  Communication

  Keep clients informed, including about the use of AI where it matters.

  • A record of what AI produced on each matter, so the firm can explain its use to a client.
  • Per-document control over what each client sees in the portal.

• Model Rule 3.3

  Candor toward the tribunal

  Never put an unverified or fabricated authority in front of a court.

  • Citation validation against authoritative sources before output reaches your screen.
  • Court rules, judge profiles and local requirements built into drafting and deadlines.

• Model Rules 5.1 and 5.3

  Supervision

  Supervise AI output the way you supervise an associate or a paralegal.

  • DDEAS signs, hashes and chains every action into a tamper-evident audit log.
  • Review stays with the attorney: the platform drafts and analyzes, the lawyer decides.

• Model Rule 1.5

  Reasonable fees

  Bill for actual time and value, not for effort the AI saved.

  • Time, billing and trust accounting tied to the matter, so billing reflects the work actually done.
  • The firm decides how AI-assisted work is billed, consistent with its engagement terms.

Read the full breakdown in the ABA Opinion 512 article, or see the DDEAS framework and our security practices.