Prompt Engineering and Context Engineering
Instruction design for reliable SI output: structured prompts, reasoning models, context design, reusable templates and systematic quality testing.
- Duration
- 3 daysor 6 half-day virtual sessions
- Level
- Practitioner
- Formats
- Public workshop, Private corporate, On-site, Live virtual, Blended
- Public dates
- To be announced
01 — Overview
The challenge
Output quality varies widely between employees using the same tools. Teams copy prompts that worked once without understanding why, and results degrade when tasks, documents or models change. Reliable results depend on deliberate context design and testing — skills few employees have been taught.
About this workshop
Moving beyond tips and tricks, this workshop treats instructions as designed and tested work products. Participants learn how current models use context, how to give them the right information in the right structure, and how to test prompts so that results stay reliable across people, documents and model updates.
02 — Who it is for
- All professionals
- Functional specialists
- Technical & digital teams
- Graduates & early career
- 03.1Professionals who use SI assistants regularly and want consistent, high-quality results.
- 03.2SI champions and power users who create templates for colleagues.
- 03.3Analysts, consultants, writers and specialists who rely on precise output.
- 03.4Technical staff preparing to build SI-enabled applications or agents.
Recommended prerequisites
- Regular use of at least one SI assistant for three months or more, or completion of SI Foundations for Business Professionals.
- No programming required. Optional exercises use structured formats such as JSON and Markdown.
Level: Practitioner
For regular SI users who want reliable, role-specific results.
Competency families
Technical
03 — Learning objectives
By the end of the workshop, participants will be able to:
- 01Structure instructions with role, goal, context, constraints, examples and an output specification.
- 02Apply context engineering: select, structure and order the information a model needs — and leave out what it does not.
- 03Use reasoning models and step-by-step decomposition where they help, and recognize when they only add cost or delay.
- 04Produce structured output (tables, JSON, templates) that people or systems can reuse.
- 05Test and compare prompts systematically against defined quality criteria.
04 — Curriculum
3 days, 6 modules. In live virtual delivery each day runs as two half-day sessions.
Day01
Anatomy of an effective instruction
- Specification thinking: goal, audience, constraints and success criteria
- Role and perspective without gimmicks
- Output specifications and formats
Examples and decomposition
- Few-shot examples and counter-examples
- Breaking complex tasks into chained steps
- Hands-on: rebuilding a weak prompt into a reliable one
Day02
Context engineering
- How models use the context window: position and noise effects
- Selecting and structuring source material
- System instructions, project knowledge and memory features
Grounding and reasoning
- Requiring citations and quoting sources
- Reasoning models: when deliberate reasoning helps
- Reducing hallucination through grounding and verification steps
Day03
Structured output and reuse
- Templates, tables and JSON for downstream use
- Prompt templates with variables
- Designing instructions for colleagues, not just yourself
Testing and maintaining prompts
- Evaluation sets and scoring rubrics
- Comparing prompts and models objectively
- Versioning, documenting and sharing a prompt library
05 — Practice & tools
Hands-on exercises
- 01Prompt rebuild: diagnose and improve five underperforming prompts from real workplaces.
- 02Context challenge: answer questions from a large document set with correct citations.
- 03Structured extraction: turn unstructured contracts or reports into a consistent table.
- 04A/B evaluation: score two prompt versions against a test set and justify the result.
- 05Library build: document three team-ready prompt templates with usage notes.
Practical skills developed
- 01Writing specifications for complex, multi-part tasks.
- 02Using examples and counter-examples to shape output.
- 03Grounding output in provided documents and requiring citations.
- 04Building evaluation sets and scoring rubrics for prompts.
- 05Maintaining a versioned, documented prompt library for a team.
Tools, technologies and workflows
- Assistants and reasoning models
- e.g. ChatGPT, Claude and Gemini, including their reasoning modes
- Context features
- projects, knowledge files, system instructions and custom assistants
- Structured formats
- Markdown, tables and JSON schemas
- Testing
- spreadsheet-based evaluation sets and, optionally, model playgrounds
Product names are examples, not endorsements. Private workshops use your organization’s approved tools; our teaching is vendor-neutral.
Technology maturity: Established
Prompt design is an established practice. Context engineering for long-context and retrieval-based systems is maturing quickly; the workshop shows current methods and is explicit about where practice is still evolving.
Maturity map06 — Outcomes
Participants produce consistent, high-quality output across varied tasks.
Expected outcomes
- Prompts are designed, tested and documented rather than improvised.
- Teams gain a shared, maintained prompt library.
- Participants are prepared to contribute to agent and automation projects.
Participants leave with
- 01A tested prompt library of at least five templates, with documentation.
- 02An evaluation set and scoring rubric for one business task.
- 03A context-design checklist for document-based tasks.
07 — Dates & delivery
Public sessions
No public dates are published yet.
Register your interest below and we will contact you as soon as a session is scheduled. Dates, venue and fees are confirmed before you commit to anything.
Available formats
- Public instructor-led workshopsIndividual professionals building their own skills.Available
- Private corporate workshopsTeams and departments learning together.Available
- On-site corporate trainingLarger groups and several cohorts in sequence.Available
- Live virtual instructor-led trainingDistributed and international teams.Available
- Blended learning journeysSkills that must change daily working habits.By arrangement
For your organization
This workshop can be delivered privately for your teams — on-site, at a venue or live virtual — with cases and exercises adapted to your context.
Request a proposal08 — Register interest
Register interest
Tell us that you would like to join this workshop. Registering interest is not a booking: we contact you when dates are confirmed, and you decide then.