The operational reference.
Slack channels you should join. The bi-weekly briefing. The AI Show & Tell submission. Frequently asked questions. People to ping when you're stuck. Bookmark this page.
Five channels. Each has one purpose.
Channel structure was rebuilt for Q2 to remove overlap. Join the ones that match your role.
| Channel | Purpose | Audience | Posting |
|---|---|---|---|
| #ai-enablement | General AI help, beginner questions, where to start | Everyone | Anyone |
| #ai-news | Feature launches, show & tell highlights, newsletter mirror | Everyone | Sid + champions |
| #ai-alerts | Outages, tool status, breaking changes | Everyone | Automated + Sid |
| #ai-champions | Closed channel — peer Q&A, async support, weekly check-ins | Champions only | Champions + Sid |
| Team channels | Local team-specific questions; safe space for beginners | Per team | Team members |
If you only join one: #ai-enablement. That's the catch-all and the fastest path to help.
The AI Task Force Briefing.
Lands in your inbox every other Wednesday. Six sections. Designed to be skimmable in 90 seconds and useful for two weeks.
Metrics dashboard
MAU, top teams, efficiency leaderboard snapshot. The state of adoption.
Feature spotlight
What's new in the last 2 weeks — skills, agents, tools, MCP servers.
Champion corner
One champion featured per issue with a quote and what they shipped.
Tips & tricks
One actionable prompt or workflow shortcut per issue.
Use case spotlight
Real example — problem, AI solution, time saved.
What's coming next
Preview of upcoming launches, events, hackathons.
First issue lands May 13
Distribution: company-wide email + posted in #ai-news. Issue 0 spotlights Alyson Sprague's 57 custom skills as the archetypal champion case.
Submit something you built.
Quarterly showcase of AI work happening across Samba. Open to anyone — champions, ICs, managers, execs. Small wins go in the newsletter; big wins get All-Hands stage time.
What counts as a submission?
Anything — an agent, a skill, a workflow, a prompt, an MCP server, a way you saved 3 hours a week. No use case is too small. Done is better than polished.
How it gets featured
4-question form, ~5 minutes to fill. Sid + Madeline review. Small wins land in the bi-weekly newsletter; big wins get a slot at All-Hands or the exec offsite.
One Sid, three time zones — here's how we handle that.
Samba has people across NA, EU, and APAC. Sid is one person in PT. We handle the gap with a mixed model — synchronous office hours per region (champion-led where Sid can't be in the room), plus an async backbone everyone gets regardless of time zone.
| Region | Time | Frequency | Lead | Status |
|---|---|---|---|---|
| NA / Pacific | Wed 11 AM PT | Weekly | Sid Dani | Launching May 2026 |
| EU / UK + Poland + Portugal | Wed 4 PM BST (= 8 AM PT) | Bi-weekly | TBD — Wave 2 recruit needed | Recruiting champion |
| APAC / Australia | Thu 11 AM AEST (= Wed 6 PM PT) | Bi-weekly | Yasmin Sanders | Pending Yasmin confirmation |
Async backbone
Sessions are recorded and posted to #ai-news with a written recap. If you can't
attend live, you don't miss anything substantive. Questions land in #ai-enablement
and get answered within the day — by Sid, a champion, or another active user.
We need a champion
Most of our never-used employees are in EU (Porto, London, Hamburg, Poland). Champion-led regional office hours is the targeted intervention. If you're in EU and want to lead this — DM Sid. Wave 2 nominations are open.
Honest about what we don't have yet: Hiring more ATF headcount for full regional coverage is a 2026-Q3+ conversation, not a Q2 launch answer. The champion model is what we're using until then. Tradeoff is real — champions run office hours on top of their day jobs, so cadence is bi-weekly not weekly outside NA.
The questions everyone actually asks.
Is Claude reading my data? Where does it go?
Samba's enterprise plan means your conversations are not used to train models. Data stays within Samba's tenant. Connectors authenticate per-user — Claude sees only what your account has access to. For specific compliance questions, contact Madeline Brown.
Can I trust the output? What about hallucinations?
Claude is excellent at structured tasks (writing, analysis, coding, summarization) and weaker at recall of specific facts. Treat it like a smart, motivated junior colleague — verify anything load-bearing. The Anthropic Academy "Foundations" course covers this in depth.
How is this different from ChatGPT?
Same general category. The differences that matter at Samba: enterprise plan with data protections, deeper agentic capabilities (Cowork, Claude Code), MCP for tool integration, and the ecosystem of skills + connectors we've built. Use what works for you — but the Samba investment is in Claude.
I've never used AI for work. Where do I start?
Three steps: (1) Open claude.ai and log in with your @samba.com account. (2) Take the Anthropic Academy "Claude 101" course — it's free and 30 minutes. (3) Pick one painful task you do every week and try Claude on it. If you get stuck, post in #ai-enablement.
I'm worried about job security. What's the message here?
AI doesn't replace people — it changes what people work on. The goal of this program is to make every Samba employee more effective, not to reduce headcount. Practically: people who learn to use these tools well become the ones who shape what Samba builds. The faster you skill up, the more leverage you have.
How do I get a Claude seat? Is there a cost to me or my team?
Most Samba employees already have a seat assigned. Check claude.ai with your @samba.com login — if it works, you're set. If you don't have access, ping #ai-enablement or DM Sid. Cost to your team: zero. The program is centrally funded.
Who do I ask if I'm stuck on something specific?
Default: post in #ai-enablement. Champions and the AI Task Force watch that channel. For technical depth or building something specific, DM Sid Dani directly. For organizational questions (champion participation, manager conflicts), DM Madeline Brown.
What's the difference between Claude.ai, Cowork, and Claude Code?
Claude.ai is the browser app — chat, projects, file uploads. Use for writing, analysis, research. Cowork is a desktop GUI for agentic work — multi-step tasks, file editing, automation. Same engine as Claude Code, easier interface. Claude Code is the engineering CLI — terminal-based, the daily-driver for our engineers.
People, not ticketing systems.
Sid Dani
Lead AI Product Manager · AI Task Force
For: tool setup, prompt help, MCP requests, harness configuration, training questions, "is this possible?" questions, escalations from #ai-enablement.
sid.dani@samba.tv · DM on Slack
Madeline Brown
Chief of Staff to the CEO · Office of the CEO
For: champion program participation, manager conflicts, executive sponsorship, accountability layer questions, anything where the answer is org-shaped, not tool-shaped.
madeline.brown@samba.tv · DM on Slack