Designed for
Technical and operational leaders, managers, analysts, consultants, and practitioners responsible for selecting, designing, supervising, or improving AI-enabled work.
← Back to the Longhand AI Roadmap
Program syllabus and certification policies
Longhand AI Competency Program
This learner-facing program guide explains who the curriculum serves, what learners are expected to produce, how proficiency will be evaluated, and which systems must be operational before Longhand can issue a competency certificate.
In development
Longhand is not currently issuing the competency certificate.
The instruction and formative practice are available. Credential issuance remains blocked until authentic assessments, qualified human review, feedback and appeals, accessible delivery, evidence retention, pilot validation, and governance approval operate end to end.
Audience and prerequisites
Prerequisites
- Basic experience using an AI chat or assistant.
- Access to a recurring workflow or an authorized synthetic case.
- Ability to save practice artifacts outside the lesson.
- Commitment to follow data, security, privacy, and professional-review rules.
Participation boundary: Instruction does not authorize access, deployment, external action, professional reliance, or the use of confidential or regulated data in an unapproved system.
Program outcomes
By the end of the complete program and planned assessment sequence, successful learners will be able to produce evidence of the following six outcomes.
Diagnose and prioritize valuable AI opportunities
Analyze recurring work, define a measurable result, compare alternatives with evidence, and defend what should or should not be built.
Design source-grounded assistants and workflows
Select the right AI surface and create grounded, reusable workflows using appropriate instructions, context, sources, tools, and verification.
Build appropriately bounded automations and agents
Specify and supervise agentic work with authorized access, explicit outcomes, ordered execution, failure handling, stop rules, and reviewable evidence.
Verify quality, reliability, and business results
Operate goal-directed loops, test repeated runs, evaluate evidence, and improve the system against a measurable result.
Apply safety, privacy, ethical, and professional controls
Protect data, preserve accountable human judgement, identify high-stakes boundaries, and stop or escalate when authority or evidence is insufficient.
Document, transfer, and continuously improve AI-enabled work
Turn effective practices into inspectable, reusable skills that another authorized person can run, evaluate, and improve.
Learning sequence and workload
Begin with the overview, complete the four core preparation courses, then use the Gear Audit to locate supporting instruction. The four core courses total 5 hours and 5 minutes. Total program workload will be published after pilot timing data covers the additional gear courses and assessments.
1. Orientation
Learn the gears, human judgement, Agent Skills, TDAR, DMP, and deterministic boundaries.
Open the overview →2. Core preparation
Complete TDAR, DMP, RICE, and Responsible AI Practice and save the independent artifacts.
Open the four courses →3. Gap-based instruction
Use the formative Gear Audit to identify supporting Chat, Assistant, Agent, and Loop lessons.
Open the Gear Audit →4. Performance assessment
Submit authentic milestones and a capstone for qualified human evaluation.
Assessment and evidence plan
Knowledge checks and browser progress are formative. Certification will require authentic work retained with the task, rubric, assessor, policy version, feedback, and final disposition.
| Planned assessment | Primary evidence | Status |
|---|---|---|
| Milestone 1 — Opportunity Diagnosis | TDAR result brief, DMP decision record, RICE comparison, sources, assumptions, and reflection. | Planned |
| Milestone 2 — Grounded Assistant | Assistant specification, source-grounded artifact, verification record, and human-review evidence. | Planned |
| Milestone 3 — Bounded Workflow Automation | Workflow brief, access/action boundaries, failure handling, tests, and operating evidence. | Planned |
| Final AI Work Capstone | End-to-end workflow, safety and governance record, measured result, reliability evidence, transfer package, and improvement recommendation. | Planned |
Scoring model
Published analytic rubrics will use four performance levels: insufficient evidence, developing, proficient, and advanced. Safety, integrity, and professional-boundary criteria may be non-compensable.
Qualified human evaluation
A credential decision may not be made from navigation, self-rating, an unreviewed automated score, or the model evaluating its own professional correctness.
Learner policies
- Responsible AI and integrity: Learners must disclose material AI assistance, preserve sources and versions, distinguish observation from inference, and may not fabricate evidence or reviewer approval.
- Data and privacy: Use only authorized data and systems. Minimize sensitive data, preserve required provenance, and use synthetic or de-identified cases when authorization is unclear.
- High-stakes work: Legal, financial, medical, employment, education, security, privacy, and similarly consequential work requires qualified professional review appropriate to the context.
- Feedback and resubmission: Planned assessments will provide criterion-specific feedback and a defined resubmission path without erasing the original evidence or decision.
- Appeals: A learner may appeal procedural error, inconsistent scoring, inaccessible conditions, or a material factual mistake. A qualified evaluator other than the original decision-maker must review the appeal.
- Accessibility: Longhand intends to provide accessible required content, equivalent alternatives, keyboard operation, and a documented accommodation process before certification opens.
Credential issuance gate
The governing authority may approve certificate issuance only when all of the following are true.
- Every claimed competency is represented by authentic assessment evidence and published rubric criteria.
- Qualified assessors meet training and calibration requirements, and scoring reliability is acceptable.
- Evidence retention, feedback, resubmission, appeals, corrections, expiration, and revocation are auditable.
- Required learner journeys pass accessibility review and support commitments are staffed.
- Pilot evidence supports valid interpretation, fairness, workload, transfer, and the public credential wording.
- A named governance authority approves issuance in writing. Until then, Longhand remains training-only.
Questions, support, and accommodations
Contact the coaching team with curriculum questions, support needs, accessibility barriers, or accommodation requests. Do not email confidential learner evidence or regulated information without an approved secure process.
Tech Leaders coaching team
Email coaches@technical-leaders.comProgram and learner support