Course Format and Delivery
Includes pre-session use case discovery consulting (up to 2 hours), the 4-hour training session, the optional 2-hour hackathon, and all content customization to the team's actual use cases. $300 per additional participant beyond 10. On-site delivery: plus travel at actual cost. Larger groups, extended consulting, or multi-session engagements: contact us for a tailored quote.
Course Overview
Custom Claude Cowork Power User Training is a hands-on course for enterprise teams that have moved past basic AI usage and are ready to build real productivity leverage. The course is built around a four-hour hands-on session per group, anchored by the team's actual use cases identified through pre-session consulting, and optionally extended with a two-hour hackathon where participants apply skills to real work with an instructor present.
Who This Training Is For
Custom Claude Cowork Power User Training is designed for enterprise teams who have already moved past the AI 101 stage and now need to convert access into actual productivity. The typical engagement profile looks like this:
Team size
10 to 40 participants across a mix of roles, from individual contributors to senior leaders.
AI access already in place
Cowork is licensed, laptops are issued, and some team members are already exploring use cases on their own.
Uneven adoption
Power users sit alongside quiet non-users. Some are getting real leverage from AI; others are asking it basic questions, not building workflows.
High-stakes work
Financial modeling, deal execution, research, content production, client deliverables, briefings, and the documents that present that work to leadership.
If this picture matches the team in front of you, this training was designed for groups exactly like yours. SoftEd has been training enterprise teams in generative AI since early 2023, runs the semi-annual GenAI Day conference, and has delivered more than 30 AI courses to over 2,000 learners.
How the Course Works
Three stages: discovery, training, and practice.
Use case discovery consulting
A working session with the team's power users to map the use cases each group needs to be fluent in. Output: a prioritized list of two to three core use cases per group, plus the growth use cases the team wants to add over the next year. We also identify which client context files (deal memos, brand and style guides, model templates, sample decks, internal playbooks) should be loaded into Cowork projects for the training.
Four-hour hands-on power user session
A working session for groups of up to 20 to 25 participants, focused on the use cases identified in discovery. The curriculum is the standard structure; the specifics of what we build and demonstrate get customized based on the group's discovery output.
Two-hour hackathon (recommended add-on)
Free-form working time, with the instructor circulating as a coach. Participants apply what they just learned to their real, current work and present a best-find at the end.
The Four-Hour Power User Session
For mixed-role teams, we typically recommend two breakouts. Specialist and analyst participants focus on their core work: research, modeling, content production, and the documents that present that work to leadership. Operations and admin participants focus on calendar workflows, briefing prep, document drafting, meeting follow-up, and the recurring tasks that drain hours from their week. The hour-by-hour structure below applies to either group; the practice use cases swap in based on role.
Cowork as a power user's workspace
Reframing Cowork from chatbox to agentic workspace.
- The Cowork mental model: projects, skills, connectors, agents
- Native skills versus custom skills versus MCP connectors
- Setting up Cowork for the kind of high-stakes work the team does day to day
- Practical demo: turning a generic Cowork session into a configured workspace for the team's actual work
Customizing Cowork for your work
This is where most power user trainings stop short. We get hands-on with the configuration layer.
- Projects in Cowork: structuring per workstream, per client, per recurring project
- Custom instructions that bake in the team's voice, formatting standards, and analytical conventions
- Loading context: source documents, internal templates, brand guidelines, model conventions
- Skills: when to invoke them, how to chain them, how to know when to build a custom one
- Connectors and MCPs: linking Cowork to the systems the team already lives in
By the end of this block, every participant has at least one of their own Cowork projects fully configured and ready to use Monday morning.
Live use case build-out
Skill-clustered sub-groups work the use cases that came out of discovery, with the instructor circulating between them. This is the heart of the session: we teach, demonstrate, then participants build.
Example: specialist and analyst focus
- Research and screening: turning a prospect list, target list, or document set into structured briefing memos
- Analytical work support: validating assumptions, exception checking, sensitivity narratives, model review
- Document and deck generation: turning analytical output into a presentation-ready story for leadership
Example: operations and admin focus
- Executive calendar and meeting prep workflows
- Document drafting and editing with team voice consistency
- Meeting note synthesis and follow-up
Power user quality control and team practice
The skill set that separates an enthusiastic AI user from a true power user is knowing when to trust the output, when to verify, and how to capture what works so the team gets faster over time.
- Iteration patterns: structured build, test, refine cycles
- Quality control: where AI commonly fails on high-stakes professional work, what to verify, when human review is non-negotiable
- Prompt and project patterns we have seen work for enterprise teams across industries
- Sharing skills across the team: how to capture what one person figured out so the rest of the group benefits
- Lightweight documentation: what to keep, what to discard
Each participant leaves with their own configured project, a small library of working prompt patterns specific to their role, and a clear next step list.
Two-Hour Hackathon (Recommended Add-On)
A hackathon, in this context, is structured practice. Participants stay together in the same live session, with the instructor present as a coach. The hackathon turns the morning's training into muscle memory before everyone goes back to their inboxes.
Identify real work
Each participant identifies one piece of real, current work they want to take a run at in Cowork.
Kick-off and circulate
The instructor sets a kick-off frame and circulates between participants and sub-groups (physically in-room or across virtual breakout rooms).
Real-time unblocking
Participants get unstuck in real time, see how others are solving similar problems, and trade techniques across the group.
Share-out
Last 20 minutes: four to six participants demonstrate the best work or biggest unlock from the session.
Our Training Team
The following subject matter experts staff custom Claude Cowork engagements. For each engagement, SoftEd assigns a primary lead based on the team's role mix and the use cases identified during discovery.
Zaki Medina
Lead AI Instructor, Generative AI and Enterprise StrategyZaki brings 17+ years of enterprise AI experience to SoftEd training engagements, with AI strategies he has shaped generating over $1 billion in business value across Fortune 100 companies. A former Microsoft AI Business School Instructor, he specializes in generative AI, AI-driven automation, and intelligent systems. Microsoft-certified: Azure AI Engineer, Azure Data Scientist, and Azure Data Engineer Associate.
George Churchwell
Lead Facilitator, Generative AIGeorge brings 35+ years in customer experience and enterprise learning to SoftEd, with deep specialty in generative AI, custom GPT design, and prompt engineering. He has built AI applications targeting $1 billion markets and AI-assisted eLearning for Cisco, and has shaped certification programs at Apple, Cisco, AT&T, Verizon, British Telecom, and VMware. Co-founder of two companies sold to private equity. Two patents in mobile learning.
Lara Hill
VP Marketing and AI Literacy LeadLara brings 15 years of marketing strategy leadership to SoftEd, with a focus on AI literacy, ethical AI practice, and the applied workflows that turn an AI license into actual productivity. She designs and uses the same Cowork power user workflows clients train on: brand-voice-aware content generation, deck and document production from source data, and project setups that preserve organizational tone.
Prerequisites and Delivery Requirements
What we need from the team
- Confirmation of the Claude tier in use. The standard assumption is the Claude Max plan that includes Cowork.
- A pre-session use case discovery working session, scheduled before the training day.
- Optional but high-value: a small set of representative files (a sample memo, a deck template, a brand or style guide, an internal playbook, a model template) shared in advance so the training environment reflects how the team actually works.
On-site delivery
- A conference room where the full group can work together.
- Participants bring their own laptops (preferred setup) so they leave the session with their own configured Cowork projects.
Virtual delivery
- A stable video conferencing platform (Zoom, Teams, Google Meet) with breakout room support for sub-group work.
- Participants on their own machines with Cowork access ready to go before the session begins.
- A working session audio setup so the instructor and participants can hear each other clearly throughout.
Group size recommendations
- Maximum 20 to 25 participants per session for true power user depth. We can run up to 30, but depth suffers.
- For larger total headcounts, we recommend splitting into multiple sessions, ideally by role cluster (operations and admins together, specialists and executives together).
About SoftEd
SoftEd is a B2B training company with 36 years of experience serving enterprise teams across financial services, insurance, retail, technology, and other regulated industries. Founded in 1990 and now operating as an independent North American company, SoftEd has trained teams at organizations including DBS Bank, National Australia Bank, Walmart, and BP. We have been delivering AI training since February 2023, with more than 30 active AI courses and over 2,000 learners trained.
Our role-specific programs span learning and development, project management, business analysis, product management, executive teams, agile methodology, and high-end agentic AI work. Every custom Claude Cowork engagement is built around the actual work the client team does every day.