By playing, you agree to uListen's Terms and Privacy Policy and YouTube's Terms of Service.
The Calum Johnson ShowAI Runs My Entire Business: EASY Way To Use AI Agents In Your Work, Content & Life (Step by Step!)
Summaries, chapters, transcripts, and FAQs on this page are made by uListen, not YouTube.
CHAPTERS
- 0:00 – 1:47
AI doing real work for you: the “eight hours while sick” promise
Rachel Woods opens with a concrete example of AI handling substantial operational work—without her actively driving—because she had already organized processes into playbooks. Calum frames the core theme: most people focus on building an agent, but the leverage comes from building a playbook first.
- •AI handled inboxes, meeting reschedules, proposal fixes, reporting, and even sending a Stripe link that resulted in payment
- •The system worked because tasks were pre-structured rather than improvised in chat
- •Calum highlights the framing shift from “build my first agent” to “build my first playbook”
- 1:47 – 3:45
For anyone feeling behind: mindset shifts and a workable framework
Calum addresses the anxiety many feel when comparing themselves to online “power users.” Rachel argues most people aren’t as far behind as they think, and that the key is learning the right mental model and structure to get consistent results.
- •Feeling behind often comes from comparing to complex-looking workflows online
- •Rachel emphasizes “mindset shifts” plus a structured method as the unlock
- •She sets expectations: listeners will leave with a framework for getting AI to do more reliably
- 3:45 – 5:47
Rachel’s path into AI: self-taught, Facebook AI research, and language models
Rachel shares her origin story: no engineering degree, self-taught skills, and landing in Facebook’s AI research team. She describes early exposure to the same “flavor” of AI now powering tools like ChatGPT and Claude.
- •She taught herself AI without a formal technical degree
- •Worked on Facebook’s AI research team with experimental tech
- •Early work involved language-model-style systems (not called that yet)
- 5:47 – 9:22
AI + entrepreneurship: from teen business builder to AI-enabled winery platform
Rachel connects her entrepreneurial background with her AI experience, describing how language understanding could offload marketing and communications for small businesses. She explains leaving Facebook, raising funding, and building an AI-enabled platform for winery owners.
- •She’s been an entrepreneur since age 15 and learned marketing through doing
- •The “aha” was pairing language-based AI with business workflows
- •She left Facebook, raised funding, and built an AI-enabled winery platform to help owners market and grow
- 9:22 – 15:05
“Operating with AI”: tools got easier, but the hard part is thinking operationally
Rachel explains why she says “operate with AI”: it’s about structuring work so it runs predictably, like operations in a company. She notes tools have changed in recent months, but success still hinges on organizing work into systems rather than ad-hoc prompting.
- •Recent tool improvements made this approach more doable for everyday users
- •The main bottleneck is not the tool; it’s how you think about structuring work
- •Operations means making work repeatable, non-chaotic, and expectation-driven
- 15:05 – 17:57
The “todaying” routine: one word triggers a day’s worth of AI-driven operations
Rachel walks through the night she got extremely sick and decided not to “push through.” Because she had already structured work into a system of playbooks, she could trigger a main daily routine with a single word and let the system run.
- •She triggered her daily routine playbook by typing “todaying”
- •AI executed a broad set of tasks without needing constant direction
- •The leverage came from prior organization of work into playbooks and routines
- 17:57 – 20:32
Playbooks vs agents: autonomy needs guardrails to be trustworthy
Calum asks for clarity on agents versus playbooks. Rachel explains the agents did the work, but they followed her playbooks—step-by-step instructions that make outputs reliable and reduce back-and-forth correction.
- •Agents are on a spectrum of autonomy; more agency means less need for approval
- •Playbooks provide the structured process that makes agent behavior predictable
- •The goal is trust: fewer corrections, less “fighting” the AI in chat
- 20:32 – 24:33
Business impact: avoiding headcount growth and scaling student support with playbooked feedback
Calum probes the business value of combining agents + playbooks. Rachel shares that playbooks became foundational: they avoided hiring in a department she expected to expand, and they enabled detailed student coaching by playbooking the feedback process while keeping human quality control.
- •Rachel says the business “could not exist without playbooks”
- •She expected to hire 3–5 people in one area but hired zero after playbooking responsibilities
- •They playbooked how to give detailed student feedback, with a human still involved for quality
- 24:33 – 27:35
Stop chasing tools: “own the playbook, rent the tech”
Rachel explains how playbooks reduce tool anxiety and platform switching costs. Once the process is captured as instructions, it can be run across different AI tools that follow instructions, rather than being tied to one vendor or interface.
- •Their philosophy: “own the playbook, rent the tech”
- •Playbooks are portable across tools (e.g., switching between AI platforms)
- •The durable asset is the operational process, not the specific tool UI
- 27:35 – 33:44
What changed recently: skills made playbooks faster to deploy than heavy automation tools
Calum asks what shifted in the last months to energize power users. Rachel says “skills” changed the game by letting people save instruction sets directly in AI tools, reducing the need for slower-to-build automation workflows.
- •Skills let you save instruction sets so the AI can reuse them on demand
- •Rachel’s team stores playbooks as skills so they’re instantly available
- •Compared with building in automation platforms, skills are faster to set up and iterate
- 33:44 – 43:17
Live demo: a project-planning playbook that standardizes execution and populates ClickUp
Rachel shares her screen and runs a project planning playbook inside Claude, showing how it asks clarifying questions, proposes options, drafts timelines/tasks, and even identifies tasks AI can do. The playbook then adds the project structure into ClickUp, making project setup thorough and repeatable.
- •A single project planning playbook helped triple the number of projects her team does
- •The playbook begins with clarifying questions and then recommends an approach
- •It drafts tasks/timelines, flags which tasks AI can take on, and pushes structured output into ClickUp
- 43:17 – 52:04
Content and sales playbooks: LinkedIn post writer + proposal writer workflows
Calum asks to see how playbooks apply outside project planning. Rachel shows a LinkedIn post writer playbook (raw thoughts → questions → reframes → draft in her tone) and a sales proposal playbook (sales call transcript + CRM context → proposal + PDF + emails + CRM updates).
- •LinkedIn playbook turns a brain dump into a polished post that still gets reviewed and edited
- •She includes follow-up questions and reframing to find the most compelling angle
- •Proposal playbook uses a sales call transcript and CRM data to draft proposals and supporting emails, then updates their system
- 52:04 – 54:12
Why “playbook first” beats “agent first”: you’re already using agents, now make them reliable
Rachel argues many people already use agents whenever an AI tool autonomously searches or acts—so the real gap is reliability and control. Playbooks are what turn agent capability into consistent outcomes that match how you want work done.
- •Rachel describes simple AI tool usage as already being “an agent” in practice
- •Playbooks make outputs more dependable and aligned with your standards
- •The key is capturing your step-by-step process so the agent can repeat it accurately
- 54:12 – 57:17
Subagents and autonomy: powerful parallelism, but playbooks keep control
Rachel demonstrates asking Claude to spin up subagents to do research across multiple cities in parallel and compile results. She explains autonomy settings (like reducing permission prompts) and emphasizes that while subagents are easy to invoke, playbooks are how you prevent uncontrolled behavior when tasks become more complex.
- •Subagents are copies that work in parallel, speeding up multi-part tasks
- •Autonomy can be increased by reducing approval/permission stops
- •This is powerful for simple tasks, but complex work needs playbook structure for control
- 57:17 – 1:06:42
Build your first playbook: pick an easy repeatable task and use the 4-part outline
Rachel lays out a practical starting method: don’t hunt for the perfect use case; build momentum. She recommends identifying playbook ideas from processes you already do, your calendar, your to-do list, and your wish list, then outlining with four components before writing in depth.
- •Prioritize momentum: the first playbook is hardest; later ones get much faster
- •Good sources: recurring calendar events, meeting prep/follow-up, repeated to-dos, and “wish list” tasks
- •Four outline components: trigger, inputs, steps, outputs
- 1:06:42 – 1:17:01
Make playbooks better (and avoid “AI slop”): smaller steps, more context, fix the playbook not the chat
Rachel explains common failure modes and iteration mechanics. Steps that are too large often underperform, and unique processes need extra context and examples; when something goes wrong, update the playbook so future runs improve rather than repeatedly correcting one-off chat outputs.
- •If a step contains too much (often signaled by “and”), break it into smaller steps
- •Unique acronyms/processes require more explanation and examples
- •When outputs are wrong, update the playbook so it “remembers,” rather than correcting only in the chat
- 1:17:01 – 1:25:36
From individual playbooks to AI ‘teams’: manager playbooks, handoffs, and the “chief of staff” routine
Rachel describes how multi-agent systems work in practice: like a human organization, with defined roles and designed handoffs. She explains “manager playbooks” that delegate to subordinate playbooks, including her daily “todaying” routine that runs standard checks and triggers other playbooks based on what’s needed.
- •Think like org design: manager playbooks delegate to subordinate playbooks
- •Handoffs work when each playbook’s output is clearly defined as the next input
- •Her daily routine functions like a “chief of staff,” running standard workflows and triggering others based on needs
- 1:25:36 – 1:29:22
How to start this week: steal templates, list 10 playbook ideas, and anchor to your ‘why’
Calum asks what to do immediately after the episode. Rachel suggests starting with an easy win using templates, quickly generating a list of playbook ideas, and picking the easiest to build momentum—while staying grounded in the personal reason you want to operate this way so you stick with it.
- •Use shared templates and start running an existing playbook to learn faster
- •Write down 10 playbook ideas and choose the easiest to build first
- •Clarify your “why” (business growth, time with family, etc.) to sustain the learning curve
