Activity
Mon
Wed
Fri
Sun
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
What is this?
Less
More

Owned by Dr. Stephen

Build & grow a profitable Skool membership. Free training for coaches & experts who want more members, clients and recurring revenue.

AI Agents for Coaches
46
5.0 (3)
$9/month

Build AI agents for your coaching business to automate content, sales, client delivery, ops and admin, with practical workflows and live support.

119 contributions to AI Automation Society
Your agent is not expensive because of the model. It is expensive because it retries.
Watch one agent run for an hour and count the retries. A tool call fails, the agent tries again, same input, same failure, three, four, five times. Each retry is a full model call. The bill is not the model price, it is the loop. The fix is boring and it works: a retry budget. Two tries on the same input, then stop and escalate. Not five, not ten, two. And log the failure with the exact input that broke it, so you fix the cause instead of paying for the same failure again tomorrow. An agent that retries forever is not persistent. It is a leak with a chat interface. How many retries do you allow before it stops?
๐Ÿ€
๐Ÿ”ฅ
0 likes โ€ข 23h
id cache the tool results too so a retry doesnt redo the steps that already worked. cuts a lot of the waste on longer chains
๐Ÿ€
๐Ÿ”ฅ
0 likes โ€ข 20h
id also log which step failed each time, its usually the same one over and over
Has an automation of yours ever โ€œworkedโ€ for weeks while doing nothing?
A builder in this group told me his content automation ran clean for four weeks while doing nothing. The model it was pinned to had been retired. No error, nothing red. He only noticed when engagement dropped. Iโ€™m building a small watchdog for exactly that. Your workflow sends one ping when it finishes (โ€œproduced 12 itemsโ€). If runs come back empty, or drop well below normal, or stop arriving, you get an alert. I only see counts and timestamps, never your data. I need 2 builders to try it free for a week. A copy of a workflow with fake data is completely fine, so nothing client-related is at risk. After the week Iโ€™ll post exactly what it caught or didnโ€™t, even if it caught nothing. Two questions: 1. Which of your automations would hurt most if it quietly stopped producing results? 2. Has that ever happened to you, and how long did it take to notice? Comment or DM me if youโ€™re in.
๐Ÿ€
๐Ÿ”ฅ
1 like โ€ข 23h
id also alert when the count drops way under the usual, not just zero. a run that does 3 when it normally does 40 slips past most checks
Hello World!... now give me your money!
Newbie question for people who got their first client I can build things with Claude now, but I have no idea how to get someone to actually pay for one. Did you start with people you know or cold outreach? And what was your first message like? Real examples would help a lot. Thanks in advance
๐Ÿ€
๐Ÿ”ฅ
0 likes โ€ข 23h
warm first for sure. i would message 10 people you know who run a business and just ask whats the most annoying repetitive thing in their week, no pitch
Who Governs the AI That Governs Other AI?
We're building agents that supervise agents. One plans. Others execute. Another monitors performance, assigns work, and decides when something needs escalation. Looks like a well-managed digital organization. But here's the uncomfortable question: What happens when the orchestrator itself makes the wrong call? If the same AI can assign work, interpret policy, approve actions, and judge its own compliance, we've created a governance loop with no independent check. I'd separate three responsibilities: โ†’ Orchestration: Coordinates agents and delegates tasks. โ†’ Enforcement: An independent policy layer checks permissions and blocks unauthorized actions. โ†’ Accountability: Human owners define boundaries, review exceptions, and remain responsible for consequential decisions. I'd also keep immutable audit trails and test the orchestrator, not just the agents it supervises. Because adding another AI supervisor doesn't automatically create better governance. You can have ten layers of AI oversight and still have zero independent accountability. In your multi-agent architecture, who or what has the authority to stop the orchestrator?
Who Governs the AI That Governs Other AI?
๐Ÿ€
๐Ÿ”ฅ
2 likes โ€ข 23h
id give the stop button to a human who doesnt own the output, one place to pause everything. if the orchestrator can override its own stop it isnt really a stop
I stopped selling "AI agents" and started selling 5-minute replies
Nobody buys an agent. They buy the thing that stops them losing leads. I build automations for small service businesses. The pitch that gets ignored: "an AI agent that handles your enquiries." The pitch that gets a reply: "out of your last 50 leads, how many got an answer within 5 minutes?" Almost every owner thinks they reply fast. Then they actually check and it's hours, sometimes the next morning, and the lead has already messaged two competitors. So the build turns out boring: instant reply, one qualifying question, a nudge if it goes quiet, then hand it to a human and stop. No dashboard, no chatbot personality, no "AI transformation." Two things I've learned: 1. Report the before and after number, not the demo. "First reply went from 4 hours to 3 minutes" lands harder than any walkthrough. 2. Judge the bot on the handoff, not the conversation. If a human doesn't take over quickly, the automation made it worse, not better. Question for the room: what's the one number you show a client that they actually believe?
๐Ÿ€
๐Ÿ”ฅ
1 like โ€ข 23h
id lead with how many leads came in after hours last month. owners already suspect they lose those so seeing the real count hits harder than any speed stat
1-10 of 119
๐Ÿ€
๐Ÿ”ฅ
@steve
๐Ÿ† ๐—ฆ๐—ธ๐—ผ๐—ผ๐—น ๐—š๐—ฎ๐—บ๐—ฒ๐˜€ ๐—ช๐—ถ๐—ป๐—ป๐—ฒ๐—ฟ โ€™๐Ÿฎ๐Ÿฒ, ๐—”๐—บ๐—ฏ๐—ฎ๐˜€๐˜€๐—ฎ๐—ฑ๐—ผ๐—ฟ & ๐—”๐—ญ ๐— ๐—ฎ๐˜†๐—ผr. ๐Ÿš€ ๐—ฆ๐—ฐ๐—ฎ๐—น๐—ถ๐—ป๐—ด ๐—•๐˜‚๐˜€๐—ถ๐—ป๐—ฒ๐˜€๐˜€๐—ฒ๐˜€ ๐—ผ๐—ป ๐—ฆ๐—ธ๐—ผ๐—ผ๐—น ๐˜„/ ๐—”๐—œ

Online now
Joined Oct 18, 2026
Sedona, AZ ๐Ÿœ๏ธ
Powered by