Děkujeme za vaši dotaz! Jeden z našich pracovníků vás brzy kontaktuje.
Děkujeme za rezervaci! Jeden z našich pracovníků vás brzy kontaktuje.
Doba trvání 14 hodiny
Návrh Školení
- Day 1: Understanding Claude and Applying It to Daily Work
- Arrival and Defining Objectives
- Review of each participant's current status and available licenses
- Participants identify one use case to work on throughout the two days
- Overview of Claude's current capabilities and known limitations
- Exercise: Document your use case on a card, detailing the goal, input materials, and desired outcome.
- How Language Models Function
- Understanding tokens and the context window—how much text a model processes at once
- Explaining why identical questions can yield different answers
- Mitigating model hallucinations and containing invented details
- Comparing Haiku, Sonnet, and Opus models based on quality, speed, and usage limits
- Determining when Claude, Copilot, or ChatGPT is the more appropriate choice
- Exercise: Complete the same task using two different models and compare results in terms of quality and speed.
- Effective Instructions and Context
- Essential components of a strong prompt: role, task, context, examples, format, and constraints
- Enabling Claude to ask clarifying questions before executing a task
- Summarizing and comparing lengthy documents and requesting specific supporting passages
- Incorporating review steps directly into the instruction set
- Exercise: Refine a weak prompt from your daily work over three iterations and compare the versions side by side.
- Projects as Shared Memory
- Logical separation of project instructions and knowledge files
- Centralized storage for house style, product knowledge, and templates
- Separating team and private areas to keep sensitive content distinct
- Verifying, correcting, and disabling memory when necessary
- Exercise: Create a project for your use case, including specific instructions and three knowledge files.
- Artifacts, Files, and Analyses
- Generating reports, tables, charts, and presentation drafts as artifacts
- Querying and comparing Excel, PDF, and Word files
- Using analysis tools to calculate figures and performing spot-checks
- Exporting, publishing, and sharing artifacts within the team
- Exercise: Create an analysis featuring a chart and a short comment based on an anonymized Excel list.
- Outcome: Each participant will have a configured project for their use case, two improved templates, and an initial artifact built from real materials.
- Arrival and Defining Objectives
- Day 2: Extending Claude with Skills, Connectors, Cowork, and Routines
- Ecosystem Overview
- Platform locations: Chat, Desktop App, Claude in Chrome, Cowork, and Claude Code
- Definitions of skills, connectors, MCP, and routines
- Simplified explanation of agents: workflows combining AI and tools
- Live demonstration from concept to completed skill
- Exercise: Map your use case to the appropriate layer and justify your choice.
- Creating the First Custom Skill
- Structuring skills with descriptions, application scenarios, workflows, examples, and templates
- Collaboratively writing and refining skills with Claude
- Testing skills with three real-world cases
- Sharing and maintaining skills within the team
- Exercise: Each participant builds a skill for their specific use case, such as drafting replies in house style or generating review reports.
- Connectors and Integrated Services
- Connecting Microsoft 365 or Google Workspace, file storage, and calendars
- Maintaining minimal and deliberate permission scopes
- Identifying hidden instructions in emails and documents (prompt injection)
- Exercise: Connect a service and execute a task across mailbox and calendar, such as preparing for an upcoming appointment.
- Claude Cowork and Routines
- Delegating tasks with access to authorized folders instead of conversational interaction
- Monitoring plans, tracking progress, and reviewing interim results
- Scheduling tasks, such as morning overviews or weekly reports
- Implementing approval points before sending or deleting content
- Exercise: Set up a routine for a daily overview and evaluate the first execution together.
- Team Guidelines and the Next 30 Days
- Defining what data can and cannot be uploaded
- Maintaining shared skills and projects while assigning responsibilities
- Distributing the model matrix and reference sheet as a handout
- Exercise: Draft a personal 30-day plan including three concrete steps.
- Outcome: Participants will complete the course with a working skill, a connected connector, an established routine, and a 30-day plan.
- Ecosystem Overview
Požadavky
- A laptop with an up-to-date browser, a Claude account (Pro, Team, or Enterprise), the Claude desktop app (recommended), and a practical use case from daily operations.
- No programming knowledge is required.
Audience
- Professionals in administration, finance, purchasing, HR, project management, and process management who are utilizing or introducing Claude in their organization and seek to advance beyond basic chat interactions.
Reference (3)
Komplexní pohled na pozadí; jak správně formulovat prompt.
Olaf Kannt - Helios Klinikum Krefeld
Kurz - Generative AI and Prompt Engineering in Healthcare
Přeloženo strojem
tipy a doporučené prompty, které si z tohoto školení odneseme
Lee Mei Lin - ST Engineering IHQ Pte Ltd
Kurz - InVideo AI: Creating Engaging Short-Form Videos
Přeloženo strojem
Využívání správných a efektivních promptů
Marses Pacaldo
Kurz - Generative AI and Prompt Engineering for Corporate Professionals
Přeloženo strojem
