👉 What You’ll Gain
- Learn how lean teams can scale like enterprises by layering AI into workflows.
- Understand how AI avatars, schedulers, and onboarding assistants reduce workload while building audience and revenue.
- See why GEO (Generative Engine Optimization) is the next frontier for creators, SaaS, and service businesses.
- Discover new risks like AI phishing and the precautions needed to protect your workflows.
- Gain insight into where billions in AI funding are flowing—robotics, chips, biotech—and what it signals for the future.
🤝 Biggest Takeaway to Implement:
Pick one AI workflow—like a scheduler, onboarding assistant, or inbox triage—and implement it in your business this month. Start small, prove its value, and layer on additional agents over time.
Free PageMotor and Practical AI Updates:
Episode 8 Reading-optimized Transcript
0:00 Introduction: Intro to Practical AI Episode 8
The episode opens with energy and excitement. Practical AI is back with another session of deep dives into real-world applications of artificial intelligence. The hosts emphasize their mission: they scan countless newsletters, shows, and reports each week, filter the hype, and surface the most practical applications. Their goal is to equip listeners with ideas they can apply immediately—or at least make them the most interesting person at the dinner table.
This episode promises to be one of the most hands-on yet, highlighting specific AI workflows that smart builders are already implementing to run lean, high-output operations. Rather than focusing on flashy headlines, the discussion zeroes in on “workflow nuggets” that can be adapted to everyday businesses—even the less glamorous ones.
0:56 Episode Theme: What smart AI builders are doing differently
The theme of the episode centers on how forward-thinking founders are leveraging AI to run leaner, faster, and more efficient operations. The key question is: what are these people doing differently that others can copy?
The focus isn’t on vague hype or sweeping claims. Instead, the conversation explores tactical implementations—everything from content creation and scheduling to inbox management and onboarding. The aim is to reveal the practical playbooks behind AI-powered growth.
1:51 Segment 1: Scaling with AI: Rowan Chung’s Rundown AI runs a 50-employee operation with 15 using AI tools
The first example comes from Rowan Chung, founder of The Rundown AI, a popular AI newsletter with more than a million subscribers. Rowan scaled his media company into a multimillion-dollar operation with just 15 employees, effectively running like a 50-person company thanks to AI.
His strategy relies heavily on autonomous agents woven into daily workflows—everything from content creation to scheduling and inbox triage. Rowan is not simply delegating tasks to AI; he’s building a company culture where AI acts as a core operating layer.
A key point: despite the power of AI, Rowan still works hard. It’s not “set and forget.” The AI removes repetitive drudgery, but human insight drives the strategy and quality.
5:14 AI Avatars & Audience Building
A major growth driver for The Rundown AI has been audience building through short-form video content. Rowan “cloned” himself by creating an AI-generated avatar trained on his writing and speaking style. With this avatar, he produces videos without being physically present.
His small team storyboards, refines, and enhances content using Claude and other tools. They then use the avatar to deliver concise clips optimized for Instagram and other platforms. The results are striking: in one year, they went from zero to 150,000 Instagram followers.
This approach mirrors Gary Vaynerchuk’s relentless content machine—but with a fraction of the team. Gary Vee might have ten people capturing and editing his daily content; Rowan achieves a similar effect with two. For someone more introverted than a Gary Vee type, this avatar-driven system provides a scalable way to build presence without constant personal appearances.
The takeaway is that audience building is essential for modern creators and businesses, and AI avatars lower the barrier dramatically.
11:17 Tool 1: Voice-to-Post Agent
Rowan uses a “voice-to-post” workflow to batch-create content. On walks, he records voice notes—brain dumps of ideas, stories, or insights. These recordings are then transcribed (via Whisper or similar) and fed into Claude.
The AI is trained on his tone and past viral posts, producing refined drafts for social content. Every Sunday, he batches ideas this way, schedules posts, and maintains a consistent publishing cadence.
The brilliance lies in the framing: AI helps structure his raw thoughts into engaging social-ready snippets. This makes content creation less about sitting in front of a blank page and more about capturing ideas naturally on the move.
13:45 Tool 2: AI Editor-in-Chief
For newsletters, Rowan uses Claude as an “editor-in-chief.” His team drafts content, then Claude provides structure, refines messaging, and sharpens calls to action.
Crucially, they don’t blindly copy-paste outputs. Instead, they iterate—comparing Claude’s suggestions with their own and selectively adopting improvements. This avoids the “AI slop” problem and ensures the final content retains a human brand voice while benefiting from AI-driven polish.
It’s less about generating content wholesale and more about enhancing quality at scale.
15:48 Tool 3: Scheduler Agent (Lindy AI)
Perhaps the most impactful tool in Rowan’s stack is his AI scheduler, powered by Lindy AI.
Here’s how it works:
- An AI email address is set up to handle scheduling requests.
- When a new inquiry arrives, the AI instantly proposes available meeting slots.
- If the prospect picks a time, the bot books it, sends a confirmation, and updates the calendar.
- If rescheduling is needed, it handles that too—automatically.
This is more personal than sending a Calendly link and avoids the stigma some perceive with scheduling links. For service businesses, the ability to respond instantly with meeting options is transformative.
The lesson: scheduling is a universal pain point, and AI can automate it in a way that feels both human and immediate.
21:00 Tool 4: Onboarding Assistant
Rowan’s company uses an “Ask GPT First” policy. All SOPs, Loom videos, and HR documents are trained into a custom GPT.
New hires are encouraged to query the AI before pinging colleagues. This reduces repetitive onboarding questions, speeds up ramp time, and ensures institutional knowledge compounds over time.
Whenever a question isn’t in the system, it becomes an opportunity to enrich the knowledge base—turning onboarding into a continuous loop of improvement.
23:26 Tool 5: Inbox Triage Agent
Email triage is another major drain on productivity. Rowan uses AI to filter spam, prioritize important senders, and surface only high-value messages.
Unlike Gmail filters, this approach uses action-based logic. It can tag team members in Slack, escalate VIP emails, or route items into workflows. The result: less inbox dread, more focus on what matters.
25:27 Tool 6: AI Tutor Agent
Rowan also leverages tools like Perplexity’s Comet browser to transform content into searchable knowledge bases.
Podcasts, YouTube videos, and newsletters become evergreen, explorable assets. For education and training, this is powerful: content doesn’t just sit passively—it turns into an interactive learning layer.
The broader implication is that YouTube and other video platforms contain vast unindexed value. AI makes this content searchable, creating opportunities for new educational paradigms.
27:28 Wrap-Up on AI Stack
Rowan’s stack of agents—including avatars, Claude, Whisper, Lindy, and Perplexity—costs about $500/month.
Key points from his experience:
- Disclosure matters: Audiences accept avatars if you’re upfront. Transparency can even increase trust.
- AI ≠ autopilot: Agents handle about 90% of the work, but human oversight ensures quality.
- Start small: Pick one or two workflows to implement first rather than trying to adopt everything at once.
The larger lesson is that lean teams can achieve outsized results by layering AI into core workflows while maintaining human guidance.
31:17 Q&A: AI for Developers
The conversation shifts to a listener question: is AI positive or negative for software developers?
The hosts emphasize context. In large companies, many AI initiatives fail because teams attempt to shoehorn AI into legacy systems. These efforts collapse under the weight of technical debt. By contrast, startups building AI-native products from scratch are finding much greater success.
For developers, the opportunity lies in rethinking products from the ground up. AI is not just a tool for coding faster—it’s a foundation for entirely new workflows. While AI will automate procedural tasks, it elevates the importance of higher-level thinking, synthesis, and creativity. Developers who embrace this shift can build cleaner, more innovative products without legacy baggage.
35:00 Segment 2: Why AI MVPs Fail
The discussion turns to a common trap: AI minimum viable products (MVPs) often fail not because teams move too slowly, but because they move too fast without clarity.
Founders rush to ship AI features without scoping properly, validating use cases, or defining success metrics. The result: wasted sprints and shallow products that never gain traction.
The lesson is to balance velocity with discipline. Moving quickly is valuable, but only if paired with thoughtful design, clear goals, and measurable outcomes. Otherwise, speed just accelerates failure.
39:58 News Roundup
Three major stories stand out in the week’s AI news:
- Microsoft’s $80B bet on AI infrastructure, signaling how much money is still flowing into compute.
- AI salaries rising 40%, intensifying the war for talent. Skilled AI practitioners are commanding unprecedented compensation.
- xAI’s shift to specialists, laying off generalists to double down on domain experts for training models.
The theme across these stories is specificity. Whether it’s infrastructure, talent, or training, the AI field increasingly rewards specialization over generalization.
43:28 Segment 3: MCP Protocol
The Model Context Protocol (MCP) is introduced as a new way to integrate AI into tools. Described as a “USB-C hub for AI,” MCP provides a standardized connector for LLMs to access tool-specific knowledge and workflows.
Rather than stuffing prompts into ChatGPT, MCP lets software expose context-specific data that AIs can query directly. HubSpot, for example, has already integrated MCP into its CRM, enabling AI to run queries like “show deals stalled for more than 30 days.”
This standard could transform how AI integrates into real-world workflows—making it less about jumping between chat windows and more about seamless in-tool intelligence.
53:17 Segment 4: Practical GEO
Generative Engine Optimization (GEO) is positioned as the future of SEO. Instead of optimizing purely for search engines, businesses must now optimize for AI-driven search platforms like ChatGPT, Perplexity, and Gemini.
The playbook includes:
- Creating narrowly focused, highly specific content.
- Supporting it with social mentions and short-form video.
- Structuring pages with schema, open graph data, and images.
The key insight is that AI-driven traffic, though smaller in volume today, converts at much higher rates—sometimes 10–40% compared to 1–2% for traditional SEO. Businesses selling high-ticket products, SaaS tools, or services should begin optimizing for AI search now to gain early advantage.
1:11:40 Segment 5: AI Phishing Risks
With AI integrated into calendars, inboxes, and productivity tools, new attack vectors are emerging. Researchers recently showed how a poisoned calendar invite could trick ChatGPT into leaking private data.
Precautions include:
- Disconnecting unnecessary integrations.
- Disabling automatic calendar invites.
- Reviewing AI permissions carefully before approving.
The broader lesson: every integration expands the attack surface. Businesses must weigh convenience against security as they embrace AI-driven workflows.
1:15:16 Segment 6: AI Funding
The week’s funding highlights:
- Figure AI raised $1B for humanoid robots.
- Groq raised $750M for ultra-fast AI chips.
- Laya Sciences raised $235M to accelerate biotech with AI.
- Ultra Green AI raised $188M for surgical applications.
- Dina Robotics raised $120M for robotic arms.
In total, $3.4B flowed into AI companies across 54 deals, accounting for 42% of all global venture funding. The big themes: robotics, chips, biotech, and AI security.
1:20:33 Closing: Key takeaways
The hosts close by underscoring the practical focus of the episode. With so much AI noise in the market, the real winners are builders who find specific, hands-on ways to improve workflows today.
Two major balls to keep an eye on:
- MCP Protocol – enabling seamless AI integration into tools.
- AI workflow agents – automating scheduling, onboarding, triage, and more in practical, immediate ways.
The message is clear: this is not theory. Practical playbooks exist today. The opportunity lies in picking one, applying it, and compounding results over time.