Prompt Engineering & AI Skills Training

The Gap Between Deploying AI and Using It Well

Organisations can deploy capable AI tools and still see limited adoption. The barrier is rarely technical — it is the gap between having access to an AI system and knowing how to direct it effectively. Staff who have not been taught to work with AI produce mediocre results, grow frustrated, and revert to previous workflows. The AI investment goes underutilised.

Prompt engineering training closes this gap. It is the practical skill of communicating with AI systems precisely enough to get reliable, useful output — and critically evaluating that output before acting on it.

What the Training Covers

Fundamentals of Effective Prompting. How AI language models interpret instructions, why prompt phrasing affects output quality, and the structural techniques — context setting, role assignment, output specification, chain-of-thought prompting — that consistently improve results. Participants practise with the AI tools their organisation has deployed, not generic examples.

Task-Specific Prompt Patterns. We develop prompt libraries tailored to the participant's role: policy analysts learn patterns for synthesising regulatory text; lawyers learn patterns for document review and legal research; financial analysts learn patterns for data interpretation and report drafting. These libraries become a shared organisational resource.

Critical Evaluation of AI Output. The most important skill we teach is not prompting — it is the ability to evaluate what the AI produces. Participants learn to identify factual errors, logical inconsistencies, inappropriate confidence, and cases where the AI has misunderstood the task. Professional judgement, applied to AI output, is the safeguard that makes AI safe to use in consequential work.

Managing AI in Professional Workflows. How to integrate AI tools into existing work processes without creating new risks — including appropriate disclosure when AI has contributed to a work product, version control of prompts used for recurring tasks, and escalation procedures when AI output is uncertain.

Information Security and Data Handling. What information may and may not be entered into AI systems, particularly cloud-hosted tools. For government clients, this maps directly to the classification of information under the Policy on Government Security and the Access to Information Act. For legal clients, this includes solicitor-client privilege implications of using AI tools with third-party providers.

The Legal Sector Context

The legal profession faces particular obligations around the use of AI — competence obligations, accuracy requirements, and privilege considerations. Our training for legal professionals addresses these directly. Participants learn not only how to use AI tools but how to document their use appropriately, verify AI-assisted legal research, and comply with Law Society guidance on AI in legal practice.

The Government Context

Federal departments using AI tools must navigate the Directive on Automated Decision-Making, the Policy on Government Security, and obligations under the Access to Information Act and Privacy Act. We incorporate these requirements into training content so that staff understand not just how to use AI but the legal and policy boundaries within which they must use it.

For departments with bilingual service obligations, we address the particular challenges of using AI tools for content that will be used in both official languages — including quality verification requirements for AI-assisted translation or drafting.

Delivery Formats

Training is available in full-day workshops, half-day intensive sessions, and multi-week cohort programmes with practice assignments between sessions. We recommend the cohort format for organisations seeking sustained behaviour change rather than a single training event. All formats include hands-on practice with the AI tools participants will use in their roles.

Delivery Process

Step 1: Discovery and Needs Analysis (Week 1). We meet with programme sponsors and a sample of intended participants to understand the AI tools in use, the specific tasks participants perform, the current level of AI familiarity, and the outcomes the organisation needs. This informs every subsequent design decision.

Step 2: Programme Design and Content Development (Weeks 2–3). We develop role-specific training content, prompt libraries calibrated to participant workflows, and practical exercises using real (or representative) documents and tasks from the organisation. For government clients, this stage includes mapping training content to applicable directives and policy frameworks.

Step 3: Pilot Delivery and Calibration (Week 4). We run the programme with a pilot cohort of 8–15 participants. Structured feedback after each session is used to refine pacing, examples, and difficulty calibration before broader rollout.

Step 4: Full Rollout and Reinforcement (Weeks 5–12, depending on cohort size). We deliver the programme across all cohorts on the agreed schedule. Between-session assignments embed practice in real work. We monitor progress and adjust content where adoption lags.

Typical Engagement

Duration: 6–12 weeks from scoping to final cohort completion. A single department of 50 staff typically runs across three cohorts over eight weeks.

Phases:

• Weeks 1–3: Discovery, design, content development
• Week 4: Pilot cohort delivery and calibration
• Weeks 5–12: Remaining cohort delivery and between-session support
• Post-programme: Six-month refresh cycle scoping

Client responsibilities: Designate a programme coordinator; secure calendar time for participant cohorts; provide access to the AI tools participants will use during training; assign subject-matter reviewers for prompt library validation.

Remolda responsibilities: All content design, facilitation, prompt library development, participant progress tracking, and post-programme refresh design.

Technology & Integrations

Our prompt engineering training is designed around the specific AI tools your organisation has deployed, not generic examples. We work with all major platforms commonly deployed in Canadian public and regulated-sector environments, including Microsoft Copilot (Microsoft 365 integration), Google Workspace Gemini, Anthropic Claude, OpenAI ChatGPT Enterprise, and sector-specific AI tools such as Relativity AI for legal document review and various government-approved generative AI platforms. We also train on retrieval-augmented generation (RAG) interfaces where organisations have deployed internal knowledge bases with AI query layers. Prompt libraries we develop are formatted for direct use in the participant's actual environment.

Canadian Regulatory Context

Prompt engineering training in Canadian government, legal, and financial contexts cannot be separated from the regulatory environment in which AI tools are used. For federal departments, the Directive on Automated Decision-Making (Treasury Board Secretariat) requires that staff understand when AI-assisted content constitutes an automated decision that triggers impact assessment obligations. The Policy on Government Security governs what information may be entered into cloud-hosted AI systems — a fundamental constraint participants must understand before using any generative AI tool. For legal professionals, Law Societies across Canada, including the Law Society of Ontario and the Barreau du Québec, have published guidance on the competent use of AI in legal practice that includes obligations around verification of AI-assisted research and advice. For federally regulated financial institutions, OSFI's supervisory expectations for model risk (Guideline E-23) apply to AI-assisted analytical work. We incorporate these requirements directly into training content so that participants understand not only how to use AI effectively, but the specific boundaries within which they must operate.

---

Further reading: Prompt Engineering for Business: A Practical Guide | AI Training for Corporate Teams