Defensive prompting
Attorneys + staffStructuring instructions so the model can’t wander: scoping, constraints, and the discipline of demanding verifiable output instead of plausible prose.
MCLE · California Bar–accredited
Practical AI, on the record.
Four working sessions that turn attorneys and staff into competent, skeptical operators of modern language models. Written and taught by a litigator — because the sanctions motion lands on your desk, not the vendor’s.
Privacy boundaries & privilege
Required for every seatWhat may never enter a third-party context window. How privilege interacts with vendor terms of service, and how to write firm policy that holds under pressure.
Citation integrity
AttorneysFact-checking model output against primary sources: cite-checking drills and a verification workflow that catches fabricated authority before it reaches a filing.
Hallucination detection
Attorneys + staffRecognizing confident nonsense: the failure patterns of large models, adversarial spot-tests, and the habits behind a zero-tolerance standard.
Accreditation
On the syllabus, and on the record.
Accredited by the State Bar of California for MCLE credit. Delivered online, for cohorts of any size — from a single practice group to the entire firm.
- Bar-accredited credit for attending attorneys
- Staff seats included — policy only works firm-wide
- Firms deploying Sovrant infrastructure run this first
