Activity
Mon
Wed
Fri
Sun
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
What is this?
Less
More
4 contributions to Useful AI • AI For Business
I stopped replying with emails when leads said "interested." Tested this instead.
Normally when someone replies to my cold email saying: "What's the cost?" or "Send more info" I would sit down, write a nice email reply, drop my link, and hit send. And 60% of them would just ghost me. I decided to test what the big outbound shops do: The Multi-Channel Speed-to-Lead bridge. Yesterday at 11:30 AM, a founder replied to my cold email: "Sure, what are your rates?" Instead of typing back an email... I immediately grabbed his direct number from the lead sheet and called him on OpenPhone 3 minutes later. He picked up on the 2nd ring. I said: "Hey [Name], just saw your email pop into my inbox asking about our rates, figured I’d call you while you’re at your desk." He laughed, said "Damn, that was fast", and we talked for 6 minutes about their after-hours setup. Booked him for a full walkthrough tomorrow morning. I used to lose so many warm replies to email lag. Calling them within 5 minutes while they are literally looking at your email changed everything for my booking rate today.
1 like • 3d
Calling while your email is still open on his screen is a smart bit of timing. Do you only call when they ask about price, or also on the vaguer "send more info" replies?
How to give an AI better examples so the answers improve
When an AI answer misses the mark, the fix is often a better example, not more instructions. Show it one or two samples of what good looks like, such as a reply you liked or a format you prefer. Say what makes them good in a sentence. If you also have a bad example, show it and say what is wrong with it. Keep the examples short and similar to the task you want done. Test with a few real cases. If a pattern keeps failing, change the example, not the whole request. What example worked best for you?
0 likes • 4d
Changing the example instead of rewriting the whole request is the part most people skip, they just keep stacking more rules on top. Do you find one good example works better than two, or does a second one stop it from copying the first too closely?
How to double check an AI answer in two minutes
AI answers sound confident whether they are right or not, so a quick check is worth the habit. Start by asking what would have to be true for the answer to be correct. Then pick the one claim the rest depends on and verify only that. A name, a number, a date or a rule is usually enough. Next, ask the same question a different way and see if the answer holds. If it changes, treat it as a guess. Finally, look for anything that sounds oddly specific with no source. Those are the details most worth confirming. Two minutes of checking beats an hour of fixing a mistake that went out under your name. What do you always double check?
0 likes • 11d
The what would have to be true question is a nice one, it forces the answer to show its assumptions. Do you ever ask the model to argue against its own answer too?
Five boring tasks in a small business that AI can take off your plate
The best first uses of AI are rarely exciting. They are the small repeated tasks that eat an hour here and there. Five good ones: Drafting replies to common emails, with a person checking before sending Summarising meeting or call notes into three clear action points Turning a rough list of updates into a short weekly report Writing the first version of product or service descriptions Sorting incoming enquiries by topic so the right person sees them first Keep a human in the loop for anything a customer will read, at least at the start. Check the first twenty outputs and write down what you had to fix. That list tells you how to improve the instructions. Which boring task would you hand over first?
0 likes • 18d
Sorting incoming enquiries would be my first pick, a misrouted email can sit for a day before anyone notices. Writing down what you fixed in the first twenty outputs is a smart habit. Do you keep that list per task or in one doc?
1-4 of 4
Anmol Sidhu
1
4 points to level up
@anmol-sidhu-7176
CEO @ NWG. If you want more booked calls from real outreach, 👉Join us: Outreach with Claude

Active 30m ago
Joined Oct 13, 2026
Powered by