Build AI-fluent teams.
An exclusive AI cohort program for Talent Scout clients. Deep AI skill training to upskill and enable your workforce.
The AI-ready agency
Make your people better.
Working with so many agencies across North America, we have frequent conversations with marketing leaders about the impact of AI. It isn't here to replace your team. AI is here to make every single person on it better. Faster. Sharper. More valuable. The companies pulling away right now aren't the ones cutting headcount. They're the ones whose people know how to use AI like a power tool.
The tool
This cohort isn't built around a stack of disconnected tools. It is built on one foundation: Claude Cowork. The cohort teaches your team how to sit in it and operate. Come back fluent in the operating environment the future is running on.
The program
Four weeks.
One intensive cohort.
Your team builds real AI agents from scratch, masters automation workflows that scale across your org, and learns the tools that turn AI ideas into repeatable systems. This isn't theory. It's hands-on. By week four, your people are shipping work that used to take weeks in days.
The curriculum
Learn to build AI automations and workflows.
This is not another AI overview course. By week four, your people are shipping work in days that used to take weeks. By the time they finish, every participant owns a complete, working Marketing Operator stack they drive.
Pre-read & Prep. Watch the trainer's video to prepare for this week's session.
2-hour live session and interactive workshop covering the week's core module.
Apply & Build. Apply the lessons learned on your own. Prepare any questions and feedback.
2-hour live session and interactive workshop covering the week's core module.
The deliverables
Deployed artifacts to bring back to work.
Not notes. Not slides. Not theory. Real, working, deployable systems built inside Cowork.
3 to 5 specialized Claude Cowork workspaces
5 to 10 production-grade system instruction templates
1 deployed marketing agent running a real recurring task
2 active automations on schedule
2 knowledge bases verified on-brand
A personal Obsidian vault for compounding knowledge
A live HTML artifact deployed to a public URL
A complete governance playbook
Your instructor
Led by Felipe Salgado
AI strategist, advisor to UNICEF on global AI strategy, and founder of the first AI school in Ecuador. He teaches from production stacks he built himself.
After the cohort
Private Slack Community
Ready to build an
AI-fluent team?
The cohort
This isn't a course. It's a cohort.
Up to 20 operators from twenty different teams, all building in the same room for four weeks. Your team member walks in solo. They walk out with peers.
The cohort is part of the value. Stuck on an agent design? Someone in the room solved it last week. Brand voice not landing? Someone else just figured that out. The little wins compound across twenty people instead of one.
Who this is for
Most agencies don't need every team member in this cohort. They need the right ones. Here's how to pick.
This is for the team member who:
- Spends real hours each week on recurring marketing work (briefs, reports, content, research, creative concepts)
- Uses ChatGPT or Claude casually but hasn't built anything beyond a clever prompt
- Wants to be measurably faster on Monday morning, not next quarter
- Can hold their own in client conversations
- Is willing to bring real client data, brand assets, and live workflows to class
This isn't for:
- Someone looking for a certification course with no hands-on work
- Someone who wants a fully managed AI solution built for them
- Pure developers looking for a coding deep-dive (this is operator-grade, not developer-grade)
- Someone unwilling to bring real client data and live workflows to the room
Common roles include
If they sit in your marketing, operations, or creative roles and produce recurring deliverables, they're a candidate.
Before the first session
Nothing exotic. This is a working cohort, not a research lab. If your team member can already use Claude, Google Workspace, and a Chrome browser, they're 90% of the way there.
Tools to have ready
- Active Claude account (Pro or Team tier for Cowork access)
- Google Workspace or equivalent (Gmail, Drive, Docs)
- Obsidian (free download)
- Two-monitor setup recommended (one for class, one for build)
Time commitment
- 15 to 20 minutes of pre-session video each Monday
- 4 hours per week of live session time (Tuesday and Thursday, 2 hours each)
- 1 to 2 hours of independent build time between live sessions
What your team member brings to class
- A real recurring task in their current role they want to automate
- Real client data: brand voice documents, current briefs, past performance, campaign context
- Willingness to demo their build in front of peers at the Week 4 Show & Tell
The full curriculum
The four-week arc, broken down to the session level. If your team member wants to know exactly what they're building when, this is the map.
- Claude vs ChatGPT vs Gemini vs Perplexity, honest comparison module
- Three specialized workspaces: Research, Content, Operations
- Claude Cowork Projects deep training (architecture, naming, routing)
- MCP overview: connectors, marketplace, when to use each
- Operator vocabulary glossary: Artifacts, Skills, Projects, Agents, Stack
- Workspace model expandable: 3 to 4 or 5 for complex roles
- RCTFX in production: system instructions, agent briefs, automation triggers
- 5 to 10 production-grade system instruction templates for the moments that matter most
- Natural-conversation playbook: context-loaded vs. structured prompt decision framework
- Deploy first agent for a real, non-fictional recurring task
- Marketing agent catalog: 5 roles × time saved per week
- 4 automation types: Scheduled, Skills, Artifacts, Multi-step
- Deploy 2 automations: daily briefing + weekly digest
- Knowledge base architecture: brand + client + performance data
- 2 knowledge bases loaded, tested, and verified on-brand
- MCP dedicated module: Slack, Notion, Figma, Higgsfield + creative tools
- Artifacts as live refreshing dashboards
- Governance rules: what to never load into a knowledge base
- Full 6-layer Marketing Operator Stack assembled with final checklist
- Build front-end HTML artifact live with Claude Cowork
- Deploy artifact to live URL (GitHub Pages or equivalent)
- 3 governance rules written and documented
- 3-minute Show & Tell live demo rehearsal and delivery
- Complete operational playbook finalized
The show & tell
The cohort closes with a live Show & Tell in Week 4. Every participant demos their stack in three minutes. Live builds, real outputs, real data.
You're on the invite list.
You'll see exactly what your team member built: the agent running on schedule, the automation saving hours every week, the HTML artifact deployed to a live URL they can already share with clients.
Not a recap of what they learned. A working demo of what they ship.