Latest / The Empire of Wealth / I_ve Done _1B In Ecom Sales. These Are The 4 AI Frameworks I_d Focus On Now

Transcript

I_ve Done _1B In Ecom Sales. These Are The 4 AI Frameworks I_d Focus On Now · The Empire of Wealth

• 0:00
  Speaker 1: In this video, I'm going to show you the only
• 0:02
  four ways I'd use AI to make money in e-commerce
• 0:05
  right now. And I want to be really clear. I
• 0:07
  do not mean four AI tools. Tools are the trap
• 0:10
  everyone is falling into. I mean four places inside a
• 0:13
  real e-commerce brand where AI can really move revenue, margin,
• 0:18
  or improve speed without blowing something up. Because over the
• 0:21
  past 12 months, I've tried putting AI into so many
• 0:24
  e-commerce brands, from brands just starting out all the way
• 0:27
  to my own brand that's doing over $ 100 million a year.
• 0:31
  I've tested things like AI ad accounts, email optimization, optimizing
• 0:35
  product pages, customer service, finance reports, 3PL invoices, creative.
• 0:41
  Speaker 2: And so much more.
• 0:43
  Speaker 1: And the weird thing I learned is that the most
• 0:45
  profitable AI workflow is probably not an AI media buyer.
• 0:49
  It's also not some magic AI store builder. It's probably
• 0:52
  something so boring most founders will skip. Now here's the issue.
• 0:56
  Most people ask what AI tool should I use? Should
• 1:00
  I use Chord, ChatGPT or Gemini?
• 1:02
  Speaker 2: Or should I use.
• 1:02
  Speaker 1: Something random on Twitter, a random app that someone says
• 1:05
  that they can automate an entire company?
• 1:08
  Speaker 2: Almost always, that is the wrong question.
• 1:10
  Speaker 1: The better question is, where is my e-commerce brand already
• 1:14
  leaking money? And how can AI help me fix that
• 1:17
  one step within that workflow? Because AI does not make
• 1:20
  money because you bought a tool. It makes money when
• 1:23
  you can connect it to a business metric such as
• 1:26
  cost per creative or email as a percentage of revenue.
• 1:30
  Speaker 2: Then get AI to fix that instantly. That is the game.
• 1:33
  Speaker 1: Now, the first way I've used AI to make money
• 1:35
  in e-commerce is by optimizing email automations. And I'm being
• 1:39
  really specific here. I do not mean use AI to
• 1:43
  write a random campaign. Most brands already know they can
• 1:46
  use ChatGPT for a Father's Day email or a Black
• 1:48
  Friday email or whatever campaign they're about to send.
• 1:51
  Speaker 2: That is fine, but that's not where I'd want to start.
• 1:53
  Speaker 1: The place I would want to start is automations because
• 1:56
  for a lot of brands, that is where the majority
• 1:58
  of sends are actually happening.
• 2:00
  Speaker 2: I'm talking about.
• 2:00
  Speaker 1: Your welcome flow, abandoned cart, browse abandonment, post-purchase, win back, replenishments, VIP,
• 2:07
  all of these flows that are sitting there sending every
• 2:09
  single day. And the crazy part is a lot of
• 2:11
  brands barely optimize these automations. They build the flow once,
• 2:15
  maybe update it every now and then, and then it
• 2:17
  just sits there making a large percentage of revenue in
• 2:19
  the background, but it's not optimized. This is the perfect
• 2:23
  place for AI. Now, quick disclaimer before I say any
• 2:26
  of this. This whole video, if you're connecting something to AI,
• 2:30
  there is some level of risk. So make sure you're
• 2:32
  comfortable with the level of access that you give it.
• 2:34
  Speaker 2: But for email, what I would do is this.
• 2:36
  Speaker 1: Step one is I would use code code or codex
• 2:39
  and connect it through an MCP server or API into
• 2:42
  the email provider that you're actually using. If that's too
• 2:45
  technical for you, you can just use screenshots and actually
• 2:48
  paste everything into one of these tools, but I would
• 2:50
  much rather it be automatically able to pull any of
• 2:53
  this data that you actually want. I'd get it to
• 2:55
  look at every major automation and report back on revenue place,
• 2:59
  order rate, click rate, conversion rate, unsub rate, revenue per recipient,
• 3:04
  and where people are actually dropping off in the flow.
• 3:06
  Then step two, I would get it to produce a
• 3:08
  report every single month, not just some giant dashboard that
• 3:12
  nobody ever reads or not just a one-off that you
• 3:15
  don't keep coming back to. I want it to come
• 3:17
  back and basically say, this month you only tested two automations,
• 3:20
  Your abandoned cart flow is doing okay, but email two
• 3:23
  is getting opened and not click. Your welcome flow has
• 3:26
  not had a proper split test in months. And if
• 3:28
  I was going to test one thing, this is exactly
• 3:31
  what I would test first. This section or this subject line.
• 3:34
  That's the first job here for AI is analysis. It's
• 3:36
  almost always the first job for AI. Before it writes anything,
• 3:40
  it should tell you where the split test should actually go.
• 3:42
  Because this is where most teams waste time and actually
• 3:45
  AI can often be better than the team member analyzing this.
• 3:48
  Often team members will split test the email they feel
• 3:50
  like changing, not the email where the test has the
• 3:52
  highest chance of making more money. So the AI should
• 3:55
  come back and say the abandoned cart second email is
• 3:58
  getting open but not clicked. Try testing the hero section
• 4:01
  and the objection handling or even putting an offer in
• 4:03
  this email. Now the second part is where AI can
• 4:05
  actually help create things. Once it's chosen the best split test,
• 4:09
  then I would ask it to rewrite the section being tested.
• 4:12
  If it's a subject line test, it might give you
• 4:13
  10 subject lines that you can choose from. If it's
• 4:16
  a hero section, it'll rewrite the hero and give you
• 4:18
  three of those options. What's really cool that a lot
• 4:20
  of people don't understand is you can actually write these
• 4:23
  emails in HTML in something like Codex or Cloud Code
• 4:26
  and paste that in your email provider and get beautifully
• 4:29
  designed emails that are very very easy for the AI
• 4:32
  to split test. Now this is a massive sticking point
• 4:34
  for brands that maybe don't have a designer or not
• 4:37
  doing email optimization because of the design element. They don't
• 4:40
  want to hire someone from Upwork. This is a great
• 4:43
  way to actually do it. You can actually just ask
• 4:45
  Cloud Code or Codex to design beautiful HTML emails and
• 4:48
  you can split test those two things. But again, I
• 4:51
  really think humans should be reviewing all of this before
• 4:53
  it goes live. I don't want AI randomly changing live automations.
• 4:57
  I want AI to tell me where the money is
• 4:59
  every single Monday and then create the test for me
• 5:02
  and actually roll it out and actually then report back
• 5:05
  on what the test did. Now, obviously, that's just a
• 5:07
  very quick overview of how it can work with email,
• 5:10
  but let's not forget this. If you're ever stuck with
• 5:13
  the process that I just described, all you need to
• 5:15
  do is actually ask the AI itself, give it the
• 5:18
  objective of what I just described of creating that report
• 5:21
  every single week, creating the emails and let it work backwards.
• 5:25
  And whenever you get stuck, just screenshot it and ask
• 5:27
  the AI, what's the next step? For example, if you're
• 5:30
  confused about what CLI is, MCP is, APIs are, then
• 5:34
  you just need to ask AI and say, what would
• 5:36
  you recommend to get this level of access? You can
• 5:38
  even ask it, what's the security risks of doing this?
• 5:41
  I don't want you to listen to any of the
• 5:43
  ideas that I have today and say, I'm not technical
• 5:45
  enough to actually do this without actually going into the
• 5:48
  AI and actually trying. The second way I would use
• 5:51
  AI to make money in e-commerce right now is add
• 5:53
  creative intelligence. And I want to remove one idea straight
• 5:56
  away for beginner. I do not think the most interesting
• 5:59
  use of AI in Facebook ads right now is letting
• 6:02
  it make budget decisions. There are a lot of people
• 6:04
  right now trying to use AI to say, increase the budget,
• 6:07
  turn this ad set off, decrease the budget if these
• 6:10
  things are actually.
• 6:11
  Speaker 2: happening in the ad account. And I get why that
• 6:13
  feels exciting.
• 6:14
  Speaker 1: But truthfully, Facebook has had rules and actually automated rules
• 6:18
  and campaign budget optimization from a long, long time ago
• 6:21
  that will probably satisfy a lot of this criteria. The
• 6:24
  big and simple opportunity with using AI and Facebook is
• 6:27
  to understand what creative is actually working and then making
• 6:30
  it way easier to get more of that creative into
• 6:33
  the account.
• 6:34
  Speaker 2: Think about the average ad account.
• 6:35
  Speaker 1: There might be hundreds of ads, different naming conventions, UGC
• 6:39
  founder videos, statics, product demos, before and afters, all of
• 6:42
  this kind of stuff.
• 6:42
  Speaker 2: And it's all messy.
• 6:43
  Speaker 1: And most teams are still basically guessing. They know the
• 6:46
  average CPA and ROAS, but they don't really know what
• 6:49
  creative pattern is driving the account because the learning is
• 6:52
  buried inside random naming conventions, spreadsheets, screenshots, Slack messages, and
• 6:57
  someone saying, I feel like this.
• 6:58
  Speaker 2: Work, let's do a little bit more.
• 7:00
  Speaker 1: So the first thing I would do is fix the
• 7:02
  naming convention enough that AI can actually read the account.
• 7:05
  I would then pull the last 30, 60 or 90
• 7:08
  days with spend, purchases, ROAS, even the hook type, the
• 7:11
  format and the avatar, as well as the asset type.
• 7:14
  And if the naming convention is messy, that is actually
• 7:17
  a great first job for the AI to actually go
• 7:19
  in and fix all of that.
• 7:20
  Speaker 2: Then you go in and.
• 7:21
  Speaker 1: Ask it to help classify the account into something useful.
• 7:24
  So you might want to understand It did founder-led product
• 7:28
  demos really work versus UDC ads. Is the before an
• 7:31
  ad comparison really working? I wanted to come back and
• 7:34
  say the ad account is spending most efficiently on these
• 7:37
  types of formats. Statics are getting spanned but not converting.
• 7:41
  And the last three winners all use this problem solution
• 7:44
  hook in the first three seconds.
• 7:45
  Speaker 2: That is actually useful.
• 7:47
  Speaker 1: Then step two where AI is really good with Facebook
• 7:49
  is creative volume. You should get the AI to look
• 7:52
  at how many new creatives were introduced over a period
• 7:55
  of time versus the amount of spend going through the
• 7:57
  account or the results the account is getting. Because if
• 8:00
  you're spending a lot and barely introducing new creatives, you're
• 8:03
  probably just waiting for the fatigue to actually hit you
• 8:05
  in the face. So I want the report to say
• 8:08
  you spent this much in the past 30 days, introduced
• 8:10
  this many creatives and based on your current spend, you
• 8:13
  probably need to introduce this many more tests based on
• 8:17
  the rest of the market or competitors introducing way, way
• 8:20
  more creatives. a kind of analysis a founder can actually use.
• 8:24
  And then step three is competitor and adjacent niche research.
• 8:27
  Speaker 2: But this is where I'd be careful.
• 8:28
  Speaker 1: The Facebook ads library is really useful and there are
• 8:31
  tools that can help you look at ads that have
• 8:33
  been running for a long time and appear to get
• 8:35
  a lot of reach because you can now sort by
• 8:37
  impressions on the Facebook ad account.
• 8:39
  Speaker 2: But meta is strict around automated data collection.
• 8:42
  Speaker 1: So I would not connect random scraping tools to my
• 8:44
  business manager or build some aggressive scraper and pretend there
• 8:47
  is no risk to my actual account. A safer way
• 8:50
  of doing this, but there is still some risk, is
• 8:52
  actually using approved APIs where they actually apply, use reputable tools,
• 8:56
  keep permissions as narrow as possible and reduce the platform risk.
• 9:00
  This is why I was so excited about the tool
• 9:02
  Manus because Facebook seemed to buy that and it was
• 9:04
  an agent that could go in and analyze ad accounts
• 9:06
  and I knew that they were connected, but that's kind
• 9:09
  of up in the air whether that transaction is happening.
• 9:11
  Speaker 2: So you could just make sure.
• 9:12
  Speaker 1: That you connect nothing of your Facebook ad account, use
• 9:15
  something like Codex or Korg Code and just use URLs
• 9:18
  where you don't have to log in, scrape that, not aggressively,
• 9:22
  and then connect it through the Gemini API to actually
• 9:25
  analyze the creative and get really good reports about what
• 9:28
  the industries are doing. The fourth step with Facebook that
• 9:30
  AI can really help with is the actual operations of it.
• 9:33
  A lot of people will be using Google Sheets or
• 9:35
  project management tools like Monday.com or ClickUp to actually manage
• 9:40
  their creative inputs. I've seen a lot of people set
• 9:42
  up systems where all you need to do is upload
• 9:45
  the creative to Google Drive and simply click approved in
• 9:48
  some of these tools. The AI will actually automatically upload
• 9:52
  it as a draft, not spending or just up into
• 9:55
  the ad account. so that media buyers can quickly go
• 9:57
  in and grab the creative rather than someone having to
• 10:00
  manually upload it.
• 10:01
  Speaker 2: The third way I would use AI to.
• 10:03
  Speaker 1: Make money in e-commerce right now is the creator and
• 10:05
  influencer content. If you're in the growth stage, maybe doing $ 100, 000, $ 200, 000
• 10:08
  a year, this might be the most practical tip today.
• 10:13
  because at that stage you need more content, you need
• 10:16
  more creators, you need more people talking about your product
• 10:19
  and you probably do not have a huge team to
• 10:21
  manage all of it. But the part I think most
• 10:23
  brands get wrong is what they're actually trying to measure.
• 10:25
  They measure things like how many creators they message, they
• 10:28
  measure how many people actually reply, maybe they measure how
• 10:31
  many videos came back. In the early stages of reaching
• 10:33
  out to creators, I want to actually know who is
• 10:35
  excited and who is pumped to get their first product
• 10:38
  And who is super transactional? This might be indicated by
• 10:41
  them just sending back a rate card before they've even
• 10:43
  touched the product or even entertained the idea of a collaboration.
• 10:46
  You'll often get this with people that are over 20,000, 30,000 followers.
• 10:51
  Because I don't really want my creator pipeline filled with
• 10:54
  people who don't care about my product.
• 10:55
  Speaker 2: So this is where AI is really useful.
• 10:57
  Speaker 1: Not to spam 1,000 creators with the same fake personalized message.
• 11:02
  Speaker 2: This is how you ruin your brand.
• 11:03
  Speaker 1: What we actually want is a pipeline where AI helps
• 11:06
  one smart team member manage the whole conversation properly. It
• 11:09
  can find the creator handles, summarize where they fit, draft
• 11:13
  the first outreach message, maybe even with a voice message
• 11:16
  from the founder, and then once they reply, it can
• 11:18
  start scoring the conversation. Not just positive or negative, I
• 11:21
  want it to tell me this creator seems genuinely excited
• 11:23
  about the opportunity. Then I would have it all feed
• 11:26
  into a dashboard, into monday.com or ClickUp or Airtable, whatever
• 11:29
  you're actually using. It'll have the creator, the sentiment score,
• 11:32
  the latest conversation summary and the next recommended action all
• 11:36
  sitting in one place. And once the team is happy,
• 11:39
  all they should have to do is change the status
• 11:41
  to approved, then the AI can trigger the next step
• 11:45
  using Shopify's AI toolkit or API. It creates 100% off
• 11:49
  discount code, one use only, capped properly so nothing weird
• 11:52
  can happen, and sends the creator the email with their code.
• 11:55
  Speaker 2: And this is important.
• 11:56
  Speaker 1: Using a discount code means that you don't actually need
• 11:58
  to collect their address manually.
• 11:59
  Speaker 2: It's so much more streamlined. They can then go through
• 12:02
  the normal store.
• 12:02
  Speaker 1: Like a customer, choose their product, and actually everything gets
• 12:05
  tracked properly. Then you can also create a second code
• 12:08
  for their followers so they get their one-time product code
• 12:10
  and then they get their public code that they can.
• 12:12
  Speaker 2: Share with their audience.
• 12:13
  Speaker 1: Now the AI can also track what actually happened, whether
• 12:16
  they posted, whether the content came back, whether the code
• 12:18
  drove revenue and whether the content actually turned into a
• 12:21
  powerful ad that performed.
• 12:22
  Speaker 2: And this is where.
• 12:23
  Speaker 1: AI becomes really good because influencer managers are often terrible
• 12:26
  at circling back. The AI should be circling back automatically,
• 12:29
  not in a creepy way, just normal follow-ups like, hey,
• 12:32
  just checking whether the product actually arrived or I would
• 12:34
  love to see did it fit or something like that.
• 12:37
  You could even ask if you posted already, can I
• 12:38
  see the link? The key here is also to have
• 12:41
  a human touch in this process. You don't want the
• 12:43
  AI handling everything. You want to build a really strong
• 12:46
  relationship with 10 to 20 creators, but by automating the
• 12:50
  rest of the process, you can really apply your focus
• 12:52
  to these key creators. The fourth way I would make
• 12:55
  money using AI right now is to actually build an
• 12:57
  AI finance and operations control system. And this is where
• 13:01
  I'd put a few things together. I'd put profit leak audits,
• 13:04
  inventory management, customer service costs, fulfillment, sales commissions, agency invoices,
• 13:11
  finance reports, all of these fit into one category in
• 13:14
  my opinion. It is anything in the business where a
• 13:16
  cost changes based on what actually happened, like a consistent
• 13:20
  percentage cost. If there's a variable cost tied to real activity,
• 13:23
  AI can usually help check whether the numbers that you've
• 13:26
  actually been charged actually make sense. And honestly, this is
• 13:29
  probably one of the biggest opportunities for most e-com brands
• 13:31
  because nobody likes to talk about 3PL invoices, but the
• 13:34
  P & L does not care what is sexy. The
• 13:37
  P & L cares about where the money is going. Now,
• 13:40
  big disclaimer is this is something where like open core
• 13:42
  might actually make sense, like an open agent that's scanning everything.
• 13:45
  Speaker 2: But I would not recommend this, especially for beginners. It
• 13:49
  is not simple.
• 13:50
  Speaker 1: And if you do not understand AI, APIs, permissions, all
• 13:53
  of that kind of stuff, something could break and something
• 13:55
  catastrophic could happen.
• 13:56
  Speaker 2: Personally, I would start much simpler.
• 13:59
  Speaker 1: I would use something like Call Code or Codex, probably
• 14:01
  a cron job, and build a very narrow agent or
• 14:05
  skill that runs the same audit every week or every month.
• 14:08
  The first thing I would probably build is the AICFO
• 14:11
  layer simply for 3PL audits.
• 14:14
  Speaker 2: Because that is a huge cost for most businesses.
• 14:16
  Speaker 1: So with Fulfillment, it pulls your Shopify order data, total orders,
• 14:20
  approximate weight, delivery zones, all of that kind of stuff,
• 14:23
  and it will actually compare it to the 3PL invoice
• 14:26
  that you've actually got. It might say something like, these
• 14:29
  pick fears look higher than expected. This weight ban does
• 14:32
  not make sense. These delivery rates look worse than last month.
• 14:35
  Then a human can go through and check it. You're
• 14:37
  not letting your AI email your 3PL and start a fight.
• 14:40
  You're using it as a point of reference so that
• 14:42
  you don't have to go through line by line in spreadsheets.
• 14:45
  We've actually built this full process in Daily Mentor. You
• 14:48
  get an AI-powered report card every single month so you
• 14:51
  can see all of these key focuses that you should
• 14:54
  do and analyze a ton of financials. So if you're
• 14:56
  interested in that, make sure you click the link and
• 14:59
  apply and we'll help you out with our mentors.