Popular AI Advice You Should Never Follow (And What to Do Instead)

Popular AI Advice You Should NEVER Follow
https://www.youtube.com/watch/4JMH-4kKMC8

00:00:00.400 Everybody is giving you AI advice right
00:00:02.480 now, and most of it will hurt you. We’re
00:00:05.040 going to go over 15 pieces of advice
00:00:06.960 that sound smart, but will set you back
00:00:09.519 years. And I’ll give you the actual
00:00:11.920 advice you should listen to instead.
00:00:13.679 Number one, fire your team and replace
00:00:15.759 them with AI agents. Why this is advice?
00:00:18.800 Because having people on your team makes
00:00:20.560 the business fun. Why would you get rid
00:00:22.320 of them, you dingdong? Business is about
00:00:24.640 people. AI can’t build relationships. It
00:00:27.599 can’t feel the energy in the room. Your
00:00:29.679 team is your competitive advantage. AI
00:00:32.159 should make them 10 times better, not
00:00:34.160 replace them. Better advice, what I
00:00:36.079 would tell people is use AI to be your
00:00:38.160 co-pilot for your team. Encourage them,
00:00:40.480 train them, show them what they can do
00:00:42.480 so that they can do their best work, not
00:00:44.480 replace them. Number two, automate
00:00:47.039 everything in your business immediately.
00:00:49.280 Why this is advice? If your process is
00:00:52.239 broken, automating a broken process just
00:00:54.960 makes it broken faster. It makes it I’ve
00:00:57.920 seen founders waste 6 months automating
00:01:00.000 a sales process that shouldn’t exist
00:01:01.840 because it didn’t work in the first
00:01:03.039 place and they’re like I’m going to use
00:01:04.400 AI to make the sales automated. It’s
00:01:06.400 like you can’t even sell. Better advice
00:01:08.720 would be fix your process manually
00:01:10.880 first. Elon Musk did this with the Model
00:01:12.960 3 production line and then he
00:01:14.400 overoptimized it and automated it and
00:01:16.240 then he came back and he added people so
00:01:17.920 that he could learn. Once you know what
00:01:20.240 actually works, then you start using AI
00:01:22.479 to automate the biggest bottlenecks.
00:01:24.240 Number three, let AI handle all customer
00:01:27.040 service to save you money. This is bad
00:01:29.200 advice. Customer conversations are gold
00:01:31.920 mines. It’s where the golden nuggets
00:01:33.600 live. You got to go mine for it. That’s
00:01:35.360 where you learn what’s broken. You learn
00:01:37.040 what features to build. You learn what
00:01:38.560 problems to solve. Cut that off and
00:01:40.159 you’re flying blind. The better advice
00:01:41.840 would be use AI for the simple
00:01:44.000 repetitive questions or response times.
00:01:46.079 Have it answer calls, answer questions,
00:01:47.920 what time are you open, easy stuff, but
00:01:50.159 keep humans in the loop for the human
00:01:52.000 touch points. Business isn’t B2B,
00:01:54.320 business to business. It’s H to H, human
00:01:56.799 to human. Number four, don’t bother
00:01:59.280 learning how AI works. Why this is
00:02:01.920 advice? If you don’t know how the AI
00:02:04.399 thinks, you won’t know its limitation.
00:02:06.479 You won’t know how to craft it. You
00:02:07.759 won’t know how to massage it. You won’t
00:02:09.119 know how to direct it. I’ve seen
00:02:10.399 founders make decisions based on
00:02:12.080 completely made up data because they
00:02:13.920 didn’t understand how the AI makes up
00:02:16.239 the answer and it made up the data. It
00:02:19.040 hallucinated. So improved advice is
00:02:21.280 learn the AI fundamentals. That way you
00:02:23.680 know what is possible and what isn’t.
00:02:25.520 And here’s a pro tip. Ask AI to teach
00:02:27.680 you AI. You can ask it how does a
00:02:29.920 language model work? Make it simple. Use
00:02:31.599 a metaphor. Use a simile and watch it
00:02:33.760 teach you so that you can then learn how
00:02:35.680 to craft it and design it to make it
00:02:37.680 even better. Number five, build
00:02:39.519 everything with AI from day one. This is
00:02:41.920 why that’s bad advice. Most people don’t
00:02:43.920 even know what they’re building yet. So
00:02:45.599 I recommend all the time, start manual.
00:02:48.400 Learn what works. Go into a company,
00:02:51.120 into a team, reverse engineer it, then
00:02:53.519 add AI to scale what’s proven based on
00:02:56.000 what you learn. Better advice would be
00:02:58.000 figure out what works manually first and
00:03:00.480 then accelerate the process with AI.
00:03:02.640 Don’t replace it, then use AI to scale
00:03:05.280 it. Number six, build your strategy
00:03:07.680 around the latest AI. This is why this
00:03:10.000 is horrible advice. Not too long ago, AI
00:03:12.640 changed every 3 months, then it was
00:03:14.879 every month. Now it’s every week and
00:03:16.480 sometimes every day. Learning what’s hot
00:03:18.400 today will put you in a position to
00:03:20.080 waste a lot of time learning about
00:03:21.680 something that’s not even relevant
00:03:22.879 tomorrow. Better advice is to focus on
00:03:25.120 solving problems that existed for years,
00:03:28.000 pains that people have. As Jeff Bezos
00:03:30.239 often said about Amazon,
00:03:31.519 >> think about the things that are not
00:03:33.040 going to change over 10 years. And those
00:03:35.440 are probably the big things. 10 years
00:03:37.040 from now, customers are still going to
00:03:38.159 want low prices. They’re still going to
00:03:39.680 want fast delivery.
00:03:40.879 >> There’s never going to be a day where
00:03:41.920 the customer doesn’t want their package
00:03:43.440 faster or for the website to load
00:03:45.280 faster. So use AI today for what works
00:03:48.400 best today and go deep. Master the skill
00:03:51.440 of one platform so that way as the
00:03:53.280 platforms evolve, you’ll already know
00:03:55.120 how the whole suite of tools work. Focus
00:03:57.360 on solving problems that have existed
00:03:59.040 for years or will always exist. Think
00:04:01.280 about pain. Number seven, let AI make
00:04:04.319 all your business decisions. Bad advice.
00:04:06.720 So bad. Here’s the deal. AI still
00:04:09.040 hallucinates and it makes mistakes. If
00:04:10.879 you think about the concept of how AI
00:04:12.799 works where it’s predicting the next
00:04:14.319 word, by definition, it’ll always give
00:04:17.120 you the most probabilistic
00:04:19.519 answer, the median watered down answer,
00:04:23.520 no matter what. And the best companies
00:04:25.440 were built off innovation. The best
00:04:27.040 companies were built off intuition. You
00:04:28.720 shouldn’t trust it with million-dollar
00:04:30.400 problems yet. Better advice would be use
00:04:33.199 AI for insights and analysis. That’s
00:04:35.360 what it can do better than anybody. then
00:04:37.199 verify it with people who’ve
00:04:38.479 successfully made those decisions to
00:04:40.160 inform if it’s the right strategy. So
00:04:43.040 use it to do research and tell you and
00:04:45.120 then take that blueprint, show it to
00:04:46.880 somebody who’s been there before. It
00:04:47.840 says, "Does this make sense?" They go,
00:04:49.360 "Wow, that’s really smart." But you need
00:04:51.680 to be the director. Let the AI do the
00:04:54.160 task. Now, before I move on to the next
00:04:55.919 piece of terrible AI advice, if you’re a
00:04:57.919 business owner, you’re trying to scale
00:04:59.199 past seven figures and you want to
00:05:00.800 implement AI into every single
00:05:02.560 department in your business, just go
00:05:04.000 find me on Instagram, Dan Martell to Ela
00:05:06.000 Martell, and DM me the word AI business,
00:05:08.160 and I’ll send you over my full guide to
00:05:09.840 do just that. Number eight, replace all
00:05:12.080 your brainstorming with AI. Bad, bad
00:05:15.520 advice. It’s so crazy. AI can only
00:05:18.320 reference what’s existed before. So, by
00:05:20.479 definition, it’ll only allow you to
00:05:22.240 brainstorm things that other people have
00:05:24.160 talked to in the past. So, if you’re
00:05:25.840 really starting from nothing and you
00:05:27.440 want just to understand what’s going on,
00:05:28.960 that’s cool. But brainstorming requires
00:05:30.880 creativity. Brainstorming is about
00:05:32.560 finding these nuances. When you think of
00:05:34.800 like Henry Ford learning from the
00:05:36.639 meatackers in Chicago to create the
00:05:38.479 production line, which is how modern
00:05:40.080 cars are still made today, but borrowed
00:05:42.160 from inspiration from a completely
00:05:43.840 different industry. You can’t do that
00:05:45.759 with AI because it’s only going to give
00:05:47.360 you things that are probably going to
00:05:48.960 work. based on what’s worked in the
00:05:50.560 past, not find new opportunities. One of
00:05:52.800 the things that AI can’t do is have
00:05:54.720 vision. That is your job to see a future
00:05:57.520 that doesn’t exist yet, but should.
00:05:59.680 Better advice would be to use AI to
00:06:01.919 spark ideas, to research things, to
00:06:04.240 expand on concepts, to test crazy ideas
00:06:07.039 in the real world from a validation
00:06:09.280 point of view, not to be the driver of
00:06:11.360 innovation. Number nine, feed AI as much
00:06:14.319 data as possible to get the best
00:06:16.240 results. This is bad advice even though
00:06:18.240 it sounds so great on the surface is
00:06:20.160 because AI has this concept called
00:06:22.400 context. And if you give it too much
00:06:24.560 information, what happens is you have
00:06:26.080 context fraud. Too much data confuses
00:06:28.319 the AI. It messes it up. It doesn’t
00:06:30.319 understand what’s relevant. It’s like
00:06:31.759 giving somebody 10 maps when they ask
00:06:34.000 for simple directions. You need to use
00:06:35.840 the most concise, cleanest context. It
00:06:38.479 actually solves your problem. If you
00:06:40.240 don’t get the context right, you’ll get
00:06:42.080 context rot. Number 10, leave AI to your
00:06:45.199 IT guys. It’s just bad advice. Saying
00:06:48.400 that you don’t need to learn AI because
00:06:50.319 your job isn’t AI and leave that to the
00:06:52.800 technical person on your team is like
00:06:54.639 saying that you don’t want to use a car
00:06:56.319 because you’re not a mechanic. AI today
00:06:58.639 is literally programmed using English.
00:07:02.000 So if you’re hearing the words coming
00:07:03.759 out of my mouth, I know you know how to
00:07:06.319 use AI because you can talk. better
00:07:08.800 advice would be, hey, let’s train
00:07:10.479 everyone on how to use AI to do what
00:07:12.880 they need 10 times more, 10 times
00:07:15.039 faster, 10 times better. Like I said, AI
00:07:17.599 is for humans. That’s why it’s coded in
00:07:19.440 English. That is the first time in the
00:07:20.960 history of technology where a technology
00:07:23.440 is programmed in English. So, don’t just
00:07:25.919 leave it to the technical guys.
00:07:27.440 Everybody should learn this. Number 11,
00:07:29.759 start by picking the coolest AI tools.
00:07:32.639 This is advice because cool technology
00:07:35.199 doesn’t solve problems. Cool technology
00:07:37.280 is just cool technology. Nobody wakes up
00:07:39.120 and says, "I need some AI." They go, "I
00:07:41.199 need to solve a problem." So, if you’re
00:07:42.720 always just picking the coolest AI
00:07:44.400 tools, then yeah, you’ll know a lot
00:07:45.840 about tools, but you won’t actually know
00:07:47.360 how to solve problems. Better advice and
00:07:49.120 what I tell my team is to find the
00:07:50.720 problem that’s worth solving. Figure out
00:07:52.400 what’s the sequence of solutions and
00:07:55.199 then find the AI to solve it. I use a
00:07:57.599 framework called the theory of
00:07:59.120 constraints. You can ask AI about it and
00:08:01.520 it’ll teach it to you. And that is how
00:08:03.520 you choose the right problems to solve
00:08:05.360 with AI. So start with the right problem
00:08:07.840 that you’re having and then go find the
00:08:09.520 right tool to solve it. Number 12, use
00:08:12.000 only one AI tool for everything. If you
00:08:15.280 think that one tool can do everything,
00:08:17.280 then you’ll have the most watered down
00:08:18.960 solution to your problems. Do you only
00:08:21.039 have one social media app on your phone
00:08:22.800 or do you have several? Why? Because
00:08:24.400 each one does something great. It’s the
00:08:26.479 same thing with AI and the tools. Each
00:08:28.479 tool does a thing better than the rest
00:08:30.160 of them. So you need to know which tool
00:08:31.759 to use for the right moment. So if
00:08:33.599 you’re using chat GBT for everything,
00:08:35.760 here’s my advice for you. Build a
00:08:37.919 toolkit of specialized AI. Match the
00:08:40.958 right tool to the right problem. You
00:08:42.958 might be using Claude to do writing. You
00:08:45.279 might use Gemini to do research because
00:08:47.440 it has 3.5 times more information than
00:08:49.680 any other language model. You might
00:08:51.680 still use Chad GPT for certain things,
00:08:53.920 but you might also find a tool to
00:08:55.519 generate images that are better than all
00:08:57.360 those three I just mentioned. Number 13,
00:08:59.920 use every AI tool that comes out. So,
00:09:02.480 this is advice because you don’t want to
00:09:04.880 be just running around using every new
00:09:06.560 tool that comes out. What you want to do
00:09:08.320 is what I just said, which is find the
00:09:10.080 three to five that you need for your
00:09:12.080 work that solve 90% of your problems,
00:09:14.160 like actually your real problems that go
00:09:16.240 deep so that you can become a master in
00:09:18.080 those tools. I know a lot of people that
00:09:19.839 just get addicted to the new tools. And
00:09:22.000 I just call that productive
00:09:23.200 procrastination. They’re addicted to the
00:09:25.360 newness, but they actually don’t do
00:09:27.040 anything with the tools. In many ways,
00:09:28.959 they are the tool, but um number 14,
00:09:32.080 just copy paste other people’s prompts.
00:09:34.640 So easy advice. Why? They’re outputs are
00:09:38.480 not what you need as a result. If you
00:09:40.399 just keep copying everybody else’s
00:09:41.920 prompts, then you’re just going to look
00:09:43.200 like everybody else. Those prompts work
00:09:45.360 for those people based on their
00:09:46.959 problems, not yours. Generic prompts
00:09:49.120 create generic output. The better advice
00:09:50.959 is understand how to create your own
00:09:53.360 prompts. Understand your process, your
00:09:55.760 inputs, your specific outputs. What
00:09:58.240 problem are you trying to solve? And
00:10:00.480 then create your prompt based on maybe
00:10:02.320 that is inspiration, but have it
00:10:04.160 customize it based on what you need.
00:10:05.760 Number 15, AI will give you a
00:10:07.839 competitive advantage just by using it.
00:10:10.160 Pope advice. It’s so crap. Look, here’s
00:10:12.399 the deal. AI doesn’t give you anything.
00:10:15.040 It’s just a thing you’re using. I know a
00:10:16.959 lot of people that use AI every day and
00:10:18.560 their life looks the exact same because
00:10:20.160 they don’t understand how to actually
00:10:22.160 use it to get their life moving forward.
00:10:24.160 Just using AI won’t get you anywhere.
00:10:26.640 It’s about how you use it. So my advice
00:10:28.800 for you instead would be your advantage
00:10:30.800 comes from combining AI with your unique
00:10:33.440 process, your relationships, your
00:10:35.200 expertise. That’s the moat. That’s where
00:10:37.839 you take you plus AI to create real
00:10:40.399 value. Now, I know what I just gave you
00:10:42.399 feels like a lot and some of it was even
00:10:44.320 controversial. It’s like master one, go
00:10:46.720 learn the other ones, build your
00:10:48.000 toolkit. But here’s the deal. All I need
00:10:49.680 from you right now is one mindset shift
00:10:51.920 that changed the way you think about AI.
00:10:54.320 I just need some feedback to know that
00:10:56.240 you got what I was sharing. So ask
00:10:58.480 yourself, out of all those ideas, which
00:11:00.160 one did you need to hear most today?
00:11:01.839 Leave a comment. I’m going to read them
00:11:03.120 all. And remember, if you want to learn
00:11:04.720 how to implement AI into every
00:11:06.320 department of your business to go even
00:11:08.160 past seven figures, just DM me AI
00:11:10.480 business on Instagram and I’ll send you
00:11:12.000 over my AI implementation playbook. Now,
00:11:14.320 if you want to learn which AI tools you
00:11:16.640 should actually use in 2026, click here
00:11:18.800 and I’ll see you on the other