Couverture de How We Replaced Employees With 65 AI Agents (Build Your Own vs Off-The-Shelf)|E15

How We Replaced Employees With 65 AI Agents (Build Your Own vs Off-The-Shelf)|E15

How We Replaced Employees With 65 AI Agents (Build Your Own vs Off-The-Shelf)|E15

Écouter gratuitement

Voir les détails

Offres de saison | 0,99 € par mois pendant les 3 premiers mois

5,99 €/mois par la suite. Des conditions s’appliquent. Annulation mensuelle possible.
À propos de ce contenu audio
What happens when you replace your office with 65 AI agents that never sleep? In this episode, Mickael Gibrael and Amer Batal reveal how they’re using AI to run their businesses — and why it all comes down to buying back your time.In this episode of The Cash Poor Asset Rich Podcast, hosts Mickael Gibrael and commercial real estate investor Amer Batal step into the fast-moving world of AI agents — the 24/7 digital workers quietly transforming how they run their companies. This is a candid, practical conversation from two business owners (not AI experts) learning it in real time, and it’s really about one thing: leverage and time.Amer shares how, over four months, he built 65 custom AI agents running on a virtual private server that stays on 24/7/365. Powered by Claude, these agents each handle a specific task — listening to and drafting email responses, syncing and updating his contacts, designing campaigns and offering memorandums, and performing full deal underwriting. He walks through one agent in detail: when a broker emails an offering memorandum, the agent treats it as a starting point (not gospel), requests the documents it needs to validate the numbers, reads PDFs, Excel, and scanned files via OCR, builds the financial model, and even drafts a pricing memo and LOI — handing off to other agents in a coordinated ecosystem that never conflicts. That single underwriting agent, he explains, saves three to four days of work per deal, with returns that compound infinitely because it runs forever.Mickael took a different route. He hired a developer to build a resellable “chief of staff” software with pre-built connectors to Slack, email, and his bank, but found it buggy and never-ending to maintain. Recognizing how fast AI is moving, he pivoted to off-the-shelf 24/7 agents — first Muse (free), then OpenAI’s Dot (with 4,000+ connectors) — transferring his entire “world” between platforms using an exported memory.md file nearly 200,000 words long. Both hosts stress a critical point: neither approach is perfect, both still hallucinate and require human review, and nobody should turn on fully autonomous emails or payments yet — Amer keeps all money transfers and communications on manual approval, and warns that exposed API keys are “the bloodline to everything in your world.”The conversation turns philosophical as they debate the future. Mickael predicts AGI could arrive by the end of Q1 next year — a tool you’d trust more than your own decisions — and recounts Elon Musk’s chilling answer that AI will need humans “as much as humans need chimpanzees.” They explore the never-ending nature of AI development (bug tickets, constant retraining, new agents), why Mickael hired a human chief of staff to “operate the machine” so he only makes high-level decisions, and why AI’s computing and error-catching can already exceed a human’s once a solid SOP is in place.Ultimately, both land on the same truth: it’s not about how hard you work — “a dog digging a hole works hard too” — it’s about outcomes. Volume plus leverage equals outcome, whether that leverage comes from well-trained humans or well-coded AI agents. Their advice for most listeners: don’t build from scratch — start using tools like Muse today, let them handle your personal tasks first, and multiply your time.In this episode, you’ll learn:How 65 custom AI agents can replace an entire officeHow one underwriting agent saves 3–4 days of work per dealHow multiple agents hand off tasks in a coordinated ecosystemWhy most people use AI as a glorified search engineThe serious security risk of exposed API keysWhy you should keep money and email autonomy on manual approvalWhy AI development is a never-ending taskThe trade-offs of build-your-own vs off-the-shelf tools (Muse, Dot)How to transfer your “world” between AI platforms with a memory fileWhy Mickael hired a chief of staff to operate the machineWhy the hosts believe AGI could arrive within monthsWhy the real ROI of AI is time, not moneyTimestamps:(00:00) The Rise Of 24/7 AI Agents(02:00) 65 Agents Running A Business Around The Clock(06:00) One Agent That Does A Full Deal Underwriting(10:00) How Much Time Does One Agent Really Save?(14:00) How The Agents Hand Off To Each Other(16:00) Why Most People Use AI As A Glorified Search Engine(20:00) The Security Risk Nobody Talks About (API Keys)(24:00) Why AI Development Is Never-Ending(26:00) Why Mickael Hired A Chief Of Staff To “Operate The Machine”(30:00) Is AGI Coming By Q1 Next Year?(34:00) The Elon Musk “Chimpanzee” Answer That’s Scary(36:00) Build Your Own vs Off-The-Shelf (Muse & Dot)(44:00) The Real Lesson: Volume + Leverage = OutcomeKey takeaways:Agents replace tasks, not just assist. 65 custom agents following clear SOPs can take over the work of an entire office — no salaries, sick days, or retraining.The time savings compound. One underwriting agent saves 3–4 days per ...
adbl_web_anon_alc_button_suppression_t1
Aucun commentaire pour le moment