Meher Patel is a serial entrepreneur with exits across hospitality, healthcare, and digital media — each in a completely different industry, each built from the ground up. He founded Neon Digital, a performance-first advertising agency, and then built what very few agencies ever achieve: a SaaS platform that outgrew the agency itself.

Hector AI now processes over $350 million in ad spend across Amazon and marketplace advertising, with 1,000+ users on the platform — and in under 18 months, has earned 3 global recognitions including the Amazon Ads Innovation Award, the Amazon Partner Award, and a Top 20 Global Amazon Ads Advanced Partner ranking.

Today, Meher is building what he believes will become the foundational intelligence layer of the agentic ecommerce era — Hector MCP: the most advanced, context-rich, token-optimized model context protocol purpose-built for Amazon advertising, designed so that every serious AI agent, every autonomous workflow, and every future-ready brand that wants to win on Amazon will have no choice but to be powered by it.

Highlight Bullets

> Here’s a glimpse of what you would learn….

In this episode of the Ecomm Breakthrough Podcast, host Josh Hadley speaks with Meher Patel, founder of Neon Digital and Hector AI, about the future of Amazon advertising. Meher explains how AI and MCP (Model Context Protocol) technology are transforming ad optimization by providing crucial context to raw Amazon data. He emphasizes that sellers should document their ad processes, learn to communicate effectively with AI, and decide whether to build custom AI workflows or use existing tools. The key takeaway: success with AI-driven advertising requires continuous refinement and treating AI as a knowledgeable, context-aware team member.

Here are the 3 action items that Josh identified from this episode:

Timestamps:

00:00:58 Introduction to the Future of Amazon Ads
The host introduces the topic: autonomous, AI-powered decision-making for Amazon advertising, moving beyond simple optimization.

00:01:13 Guest Introduction: Meher Patel
The host introduces Meher Patel, detailing his entrepreneurial background, his agency Neon Digital, and his SaaS platform, Hector AI.

00:02:49 The Problem with Early AI Ad Tools
Discussion on how early AI advertising tools often failed sellers, contrasting with the positive results from newer, more advanced software.

00:04:10 Prediction for Amazon Advertising
Meher predicts Amazon will rapidly release new AI-powered features, but sellers must learn how to properly utilize this infrastructure.

00:08:46 The Importance of Context in AI
AI is only as good as the context it’s given; without it, AI recommendations are generic and potentially harmful.

00:10:04 How Smart Sellers Should Prepare for AI
Sellers must learn to ask the right questions and feed AI the right data with the proper context to get valuable results.

00:12:07 Why Raw Data Isn’t Enough
Uploading raw Amazon reports to an AI lacks the necessary context, leading to “garbage out” optimization strategies.

00:12:42 The Role of an MCP (Model Context Protocol)
An MCP provides the necessary context and data connections, acting as an intelligent layer between raw data and the AI model.

00:18:57 Amazon’s MCP API Limitations
Amazon’s own MCP is just an API, requiring sellers to build their own infrastructure, which is inefficient and token-heavy.

00:21:48 Top Strategies: Building Custom AI Agents
The best strategy is for brands to build their own custom AI agents and workflows based on their unique strategies.

00:24:32 Unlocking Custom Workflows with AI Agents
AI agent workflows allow sellers to build bespoke optimization systems, unlike one-size-fits-all SaaS platforms.

00:27:10 How to Create an AI Agent Workflow
Record your optimization process, use an LLM to create an SOP, and then build an AI agent to execute it.

00:28:06 The Reality of AI Implementation
Building a reliable AI agent is a gradual process of refinement and setting up guardrails, not a weekend project.

00:29:21 Automating Agent Creation
Using connectors like Make.com within an LLM allows you to create and schedule automated workflows by simply describing them.

00:31:08 The Timeframe for Building an AI System
Building a truly autonomous system is a long-term journey of refinement; the key skill to learn is communicating with AI.

00:33:57 Becoming an AI Orchestrator
Sellers must become orchestrators, designing and managing multiple small, independent AI agents to perform specific, connected tasks.

00:35:56 The Future: Loaning vs. Building AI Agents
Sellers will choose between “renting” cookie-cutter AI agents or “building” custom ones that act as a competitive moat.

00:38:29 Are You a Brand Owner or a SaaS Provider?
A warning for sellers: building your own AI tools means you are entering the SaaS business, which requires significant technical resources.

00:41:13 The Shift from Prompt to Context Engineering
The new challenge is context engineering: ensuring the right data and tools are used efficiently to avoid token exhaustion and errors.

00:42:55 Three Actionable Takeaways
The host summarizes three key actions: document processes with video, use an MCP for context, and decide your role (brand/SaaS).

00:47:25 Most Influential Book
Meher shares that the biography of Steve Jobs has been his most influential book due to its lessons on focus.

00:48:25 Favorite AI Tool
Meher recommends WhisperFlow for voice-to-text communication with AI, which has eliminated his need to type when using Claude.

00:49:23 Most Respected Person in E-commerce
Meher names Jeff Cohen as someone he admires for his deep, hands-on knowledge of the Amazon and retail media ecosystem.

Resources mentioned in this episode:

Tools and Websites 
“Hector AI“: “00:02:14”
“Lovable AI“: “00:26:00”
“A10“: “00:30:04”
“Claude AI“: “00:30:56”
“Whisper Flow AI“: “00:48:25”Books 
“Steve Jobs by Isaac“: “00:48:29”Influential People 
“Ezra Firestone“: “00:01:13”
“Kevin King“: “00:01:13”
“Michael E. Gerber“: “00:01:13”
“Jeffrey Cohen“: “00:49:23”

This episode is brought to you by eComm Breakthrough Consulting where I help seven-figure e-commerce owners grow to eight figures.

I started my business in 2015 and grew it to an eight-figure brand in seven years.

I made mistakes along the way that made the path to eight figures longer. At times I doubted whether our business could even survive and become a real brand. I wish I would have had a guide to help me grow faster and avoid the stumbling blocks.

If you’ve hit a plateau and want to know the next steps to take your business to the next level, then email me at josh@ecommbreakthrough.com and in your subject line say “strategy audit” for the chance to win a $10,000 comprehensive business strategy audit at no cost!

Transcript Area:

Meher Patel 00:00:00  And that’s exactly what an aging tech workflow is going to unlock for every 100 sellers. They can now rely on building a workflow that they desire, which is important for their business. Unlike the SaaS platform, which is built for everybody and they have no restrictions, no limitations of any any aspect of this rule, it doesn’t work because I don’t have this aspect here. I don’t have the ranking gear, I don’t have the search query performance talk to my search term yet. So it’s it’s it’s not there because the platform is not built it, but MCP unlocks that.

MC 00:00:45  Welcome to the Ecomm Breakthrough podcast. Are you ready to unlock the full potential and growth in your business? You’ve already crossed seven figures in sales, but the challenge is knowing how to take your business to the next level.

Josh Hadley 00:00:58  Most Amazon sellers are still optimizing ads the same way they did 3 to 5 years ago. But what if the future of Amazon advertising isn’t just optimization, it’s autonomous decision making powered by AI? That’s the playbook that we’re going to be diving into today.

Josh Hadley 00:01:13  Welcome to the Econ Breakthrough Podcast. I’m your host, Josh Hadley. I scaled my own brand from 0 to 8 figures in sales, and now my mission is to take it to over nine figures on my journey to nine figures. I bring you the unfiltered conversations with the smartest minds in e-commerce. Past guests include Ezra Firestone, Kevin King, and Michael E Gerber, author of The Myth. Today, I am excited to welcome Meher Patel. Meher is a serial entrepreneur with exits across hospitality, healthcare and digital media, each in a completely different industry, each built from the ground up. He founded Neon Digital, a performance first advertising agency, and then built what very few agencies ever achieve a SaaS platform that outgrew the agency itself, Hector AI now processes over 350 million in ad spend across Amazon and other marketplaces, with over a thousand users on the platform and in under 18 months has earned three global recognitions, including the Amazon Ads Innovation Award, the Amazon Partner Award, and a top 20 Global Amazon Ads Advanced Partner Ranking today.

Josh Hadley 00:02:14  Maher is building what he believes will become the foundational intelligence layer of a genetic e-commerce era. Hector MCP, the most advanced, context rich, token optimized model context protocol purpose built for Amazon advertising design so that every serious AI agent, every autonomous workflow, and every future ready brand that wants to win on Amazon will have no choice but to be powered by it. With that introduction, welcome to the show, Maher.

Meher Patel 00:02:39  Thanks a lot, Josh. I think as I said, I mean, you did a fantastic, you know, introduction. I really loved it. It’s just so much better when you say it. So thank you for doing that.

Josh Hadley 00:02:49  It’s it’s always good to hear somebody else tout you up. So you’ve got a fantastic background. And what’s even more impressive is that you’ve built this, you know, this software business that again, outgrew your agency, which obviously means you are doing something right. And what I love is that, in the past, I think like maybe 3 or 4 years ago, there were some advertising software tools that claimed to be AI optimized advertising tools that, frankly speaking, it sounded like it wrecked most of the sellers ad accounts whenever they utilized it.

Josh Hadley 00:03:23  However, on the flip side, I’m actually hearing positive things from users from your software. But I want to dive into this my hair, because our listeners I know are asking this question. Everybody A wants to be at the cutting edge of AI in Amazon advertising, but B they’re not exactly sure like how they should be doing it. I think the challenge with Amazon advertising is, is that it’s very nuanced, right? Like, yes, you can set up rules in automations. However, there’s many different scenarios where it’s like, well, sometimes this and maybe sometimes that, which is where it gets really challenging. So the question I want to ask you, my hair is like, what is your prediction of what Amazon advertising looks like and the best practices for it over the next 12 to 18 months?

Meher Patel 00:04:10  Well, that’s a good question, Josh. I think the way Amazon is, is building and in the way it has launched new features. I’m sure you’ve been, you know, looking at this, that there was a time Amazon used to launch new features once in three months, once in six months.

Meher Patel 00:04:25  But now with the help of AI, Amazon is launching new features. Every month. You go on the dashboard, you wake up early morning and you open your dashboard and you see something new popped up on your screen. There’s a different navigation, there’s a different feature which Amazon has just launched all of a sudden. And that’s because Amazon believes that has access to that sort of a dev team, which is powered with AI. And obviously, as we discussed in past, that whatever Amazon is going to build is going to build it for the entire category for all the Amazon sellers. So there’s going to be a certain amount of, fair play that Amazon is going to create to make sure that every brand leverages the AI to the extent that they can job, you know, that they can enable their their growth on Amazon. But still the responsibility of of a brand remains that how well are they utilizing the entire AI infrastructure that Amazon is going to. Amazon is going to power them with. So that being said, I think the most important part that Amazon sellers has to understand is AI is definitely changing the entire landscape in the way the Amazon ads ecosystem is going to function, but it’s not up to them to start building the AI from the scratch, because it is just the launch of AI at this moment.

Meher Patel 00:05:41  There’s a there’s a large amount of infrastructure change in infrastructure. Building from the bottom up would be required. So the FOMO is nice, the excitement is great, but it’s not something that every Amazon seller should look into building from scratch. There are opportunities in the market where a sellers sellers can go up and look out for platforms, look out for an MCP and try their best, see what’s the what’s working, what’s not working, instead of building it from scratch. Because that’s not what Amazon seller is, is, is actually supposed to do. Right. So that’s not their core. That’s not their main responsibility. Their main responsibility is to grow their brand and not to build a platform. And, one thing that gives us that opportunity to build the platform the way it has been accepted and adopted is because we are an agency ourselves. So we understand the pain to the deepest level when we run those campaigns ourselves. Just like you’re an Amazon seller yourself. I mean, you know what it takes. And you know the pain an Amazon seller goes through.

Meher Patel 00:06:45  So when you talk to an Amazon seller, you actually know what to ask. You actually can figure out what’s working and what’s not working because you’ve been there. I mean, you are the Amazon seller. So similarly, if a tool is built by someone who has actually taken and is going through the pain, the outcome of the technology which is built out of it is going to be completely different than the ones who are just building the tool, because the one who’s just building the tool is just building the tool. He’s not he’s not taking the pain. He’s not going through that pain. So the passion in which the tool is being built matters. And I think that’s something that we have seen and the features that we’ve built, we’ve actually not we don’t go out looking for what is built outside, but we go out looking what is happening on the floor. What is it that that guy who is optimizing that campaign is going through, and how our tech or how AI now is going to help them get better optimized or be more efficient? I think the brands should look out for what’s not just what Amazon is building, but also look out for what’s being built with the infrastructure that Amazon has enabled.

Meher Patel 00:07:53  So that’s more powerful. I mean, that’s that’s that’s growing at a faster pace, I believe.

Josh Hadley 00:07:56  Yeah. And I think it continues to change. You know, month over month as new AI tools get deployed. But to to summarize what you just talked about there is, you know, over the next 12 to 18 months, my prediction as well is that Amazon is going to come out with AI optimized campaigns and bidding structures and things like that that are going to make the let’s call it the unrefined or the uneducated Amazon seller feel like they are using AI and they’re going to feel like, oh, well, surely because Amazon has said that these bids are AI optimized and Amazon has my back. I think we all see through that at this point in time. Like Amazon, at no point in history has truly had a seller’s back. It’s kind of like, well, if you don’t if you don’t succeed, there’s just another seller right around the corner that’s ready to take your spot on that page one listing.

Meher Patel 00:08:45  All right.

Josh Hadley 00:08:46  And so what that means is that you cannot rely on any of the optimization tools that Amazon is, quote unquote, going to give you for free. Right. Here’s the here’s the important thing that you touched on. It’s all based on like the context that you give it. An AI is only as good as like it understands your systems, your processes and your goals. Because if it doesn’t have any of that context sitting behind it, it’s kind of like going to ChatGPT opening it up on some like a library computer has no context or history about you. You’re not even logging in, you’re just using the free aspect of ChatGPT. And then you say, hey, what do you think my life is going to look like over the next ten years? And it’s like, and then it’s going to it’s going to confidently spit out a response for you, but it has absolutely no idea who you are. But it’s going to randomly say things about you. And so that’s what’s going to happen if you just allow AI to just like basically run you off a cliff.

Meher Patel 00:09:43  Indeed.

Josh Hadley 00:09:43  And so my question to you, my hair is like, so what do the smart Amazon sellers do knowing that hey, AI is where things are going. Like what are what should be the steps that a smart Amazon seller starts taking today? To be prepared for that shift to where you know, their Amazon ads are actually being optimized by AI agents?

Meher Patel 00:10:04  So great, great great question, Josh. You know, one of the few things that when we launched Hector MCP and when we started using Hector MCP and we started offering it to a few beta clients, you know, more than anything, they were like dumbstruck, like, what do I do now? What is the question that I should ask? Because the world of Amazon sellers have always been assisted with a SaaS platform. What it means is that whenever you open a platform, you have an assistance to navigate what you want to see, and you’re actually seeing what is shown to you. You’re not asking what to see. So you’ve always been guided on the SaaS platform, but the moment that is taken away, you have a blank screen in front of you and all you have is just ask me a question and I’ll answer you.

Meher Patel 00:10:55  So like you said, that, you know, it is about I mean, I keep saying that in doubt. It’s basically if you are not going to assist the AI to drive the right results from you, from from what you’re asking, it’s going to come up with something that you that you’re going to go in the wrong direction. So the point is that consider the AI your best and the most intelligent person that you could hire next to you. Consider having a McKinsey or a large consulting company. The topmost guy, the most intelligent data scientist next to you. Now that guy is with you. All you need to do is ask him and feed him with the right data. And the most important thing is the right context. Now that’s conversation. When you start this conversation, that person that AI is only going to spit out based on how enriched your data is and how powerful your question is, right? So if the data and if the question is matched is marrying, it can talk to each other.

Meher Patel 00:12:00  Well, that best friend of yours, which is the AI, is going to give you the best answer. And that’s the world that.

Josh Hadley 00:12:07  Are you telling me, are you telling me that if I just download a random, you know, Amazon advertising report and if I just upload that to either cloud or ChatGPT today, I’m going to get some kind of like really random answers. If I can download or upload the Excel sheet, I’m going to get random optimization, like it’s going to give me optimization strategies. But are you telling me like it’s lacking context? And so any of the response I get back is just going to be kind of like garbage out? It’s like I’ve got a good report, but like it’s just missing context. Is that what you’re saying?

Meher Patel 00:12:42  Indeed. So any data with no context is as good as throwing it. You know, throwing, dumping it to the best friend. Right? Dumping it to the guy who’s the most intelligent guy. So you’ve given him data, but you’ve not explained him.

Meher Patel 00:12:58  What that data comes from. Is that a keyword data? Is there a search from data? Is there an advertised Ace in data? Is this advertised being done in a single keyword? Single campaign? What has been your structure? Is this search term coming directly from the keyword? What is a search term? What is a keyword? All of this information lies in silos across the web. And based on what model, what tool are you using and based on how deep is your research? When you’re asking this question, it’s going to give you an answer. So that varies. Now that’s exactly what a good MCP is going to take care of. He’s going to make sure that that person, that consulting guy who’s going to be backing you, who’s going to be next to you all the time, Dedicated for you. Already knows what an Amazon seller needs. Already knows that this data comes from where and what. It means that when you see a search term and when you see a search query from the search query performance report, it’s two different terms altogether.

Meher Patel 00:13:57  And when you see a search term, it doesn’t have advertising spans. But when you see search query report it I mean it has advertising spans when you see search term. But when you see the search query report, it doesn’t have an advertising expense. But the search query performance report is the right search term that you can marry with with the search term report. So all of this context, all of this cross linkage is doesn’t exist. Now. Even if you wanted to exist you have to start taming your AI. You have to start asking him first. Hey, do you know anything about the search query performance report? Then he’s like, all right, let me just tell you, what do I know from the web search that I’m doing? And what is the search query performance report? Then you if you’ve done only a web search, and if you’ve not done the deep research, it’s going to give you a different answer, by the way. And once you’ve done that, then you say, hey, how is it relevant with this version of report? Then he’s going to go another deep research and then you will say that, hey, can you just marry them both together and tell me if that makes sense to see both of those report together.

Meher Patel 00:14:51  All right. So now you are training that guy. Now that training, if you don’t do it, if you don’t pass this journey before you dump the data and then expect the answer, if you don’t train him well, he’s going to give you abstract answers. And that’s the context is what we as an as a responsible MCP has to build. I keep telling my people that what’s an MCP? Imagine MCP is is a wrapper on a chocolate. And that wrapper can talk to any LM. It depends on how enriched and how I would say context built that wrapper is. It is going to talk in that manner with the LMS. The LMS could be your ChatGPT cloud Copilot Gemini, any of those things. Now the point. The second aspect to that is that is your wrapper very thick. That means that it is context bloating. There is too much data because, you know, like you’re just doing the research and research and you keep throwing that research into it. And then tell that every time when I ask you for soup and search term, make sure you go through this.

Meher Patel 00:15:58  What we just learned before you give me an answer. Now, if you’re going to ask them to go through that before answering that which search term was good for me, your entire context window on that cloud system is going to crash by itself, because it doesn’t have the the token for you to give that answer to you with taking all that context, taking all that data together and then respond to you. So that’s a big mess right now. And and the good part is we’ve gone through this journey and this is not this is not the first three months of us building MCP. We’ve been building MCP from last nine months. So we understand what is going to come the way to an Amazon seller. And that’s why we keep telling an Amazon seller, if you really want to do justice to your business, it’s about you. How do you best utilize the MCP then? How do you start building the MCP? So don’t get into building the MCP, but start looking at how do you use the MCP? MCP is just an enabler, but the creative mind and the the sellers, the way they want to optimize their campaign still lies with them.

Meher Patel 00:17:02  And now with the MCP, you have less dependency on a SaaS platform, which has always been an assistant to you, to see what you were shown.

Josh Hadley 00:17:11  Yeah, I love those insights. And I think, like the overall takeaway, this is similar to the conversation I had with with Ellis Whitehead from Data Brill’s. And he talked about how it, you know, the first version of like ChatGPT, everybody was saying, you know, hey, if you feed it a battery prompt, you’re going to get a bad response back out. Well guess what? Now every like that’s still true, and prompt engineering has come a long way. But the llms have also come a long way at the same token. But here we are at this intersection where, you know, there are businesses being built with AI agents right now that are operating independently. However, in order to do that, just like you talked about here, where you have to set up the data that it is analyzing or making decisions off of properly.

Josh Hadley 00:18:01  And this is the mistake. I’ve spoke to so many sellers and I’ve listened to sat in on many of these conversations. Most of them are just trying to like API directly into Amazon, and then they get the API and then like, oh, well. And then I just like tie the API into cloud, and cloud runs things and it’s like, okay, that’s fine, but you’re missing a massive data gap here. And what you talked about is that MCP layer needs to sit in between that API connection to Amazon or whatever tool you’re using so that your data. Honestly, it’s almost like you you need to become familiar with, like the technical terms like data warehousing and adding context. This is what this like number and variable means in this column on this spreadsheet. This is what the data means on this column. This is what the data means. Here’s the context. Here’s the why behind it. Yeah. And most people aren’t doing that work. And that’s I guess, what you’ve built. Is that what you’re saying?

Meher Patel 00:18:56  Indeed, indeed.

Meher Patel 00:18:57  In fact, the Amazon’s own MCP is actually not an MCP. It’s an MCP APIs. That means that it’s not a plug in play. It’s just an I mean, the one word which is changing the way the MCP is supposed to be. What is an MCP? The Gmail has an MCP. The click up has an MCP. The slack has an MCP. What does it mean? You go to the cloud, you go to ChatGPT. You toggle it on. You sign in and you’re done. You can ask Claude to read your email. It’s going to read your email. But that’s not what Amazon’s ads. MCP is Amazon’s ads. MCP is an Amazon ads MCP API’s. So cloud is able to read your Gmail. Because the Gmail has stored your data. But if even if you connect your Amazon ads MCP, which is an API, it is only going to request for the data. And when you say how is my performance for my keyword X, it’s going to download all the performance of the last 30 days for that keyword.

Meher Patel 00:19:52  Impression clicks and all of those things. For all you know it is just firstly, you will not understand what is performance. The second, even if you understand it’s going to download so much of data that your token is going to get exhausted. And he doesn’t understand this because you Amazon is not storing the data, Amazon API is just pulling the data. And you need tokens to even pull that data. Right. So and when you’re pulling it it’s going to pull every key every matrix. Right. And it doesn’t have a filter option. So it’s going to bring all those keywords. And from there it’s going to find your keyword that you ask for as a performance. So that’s like a very bad practice that somebody who doesn’t have the right MCP direct plug in MCP means it has to be a plugin. It cannot be something, which is you have to start building it. It has to be just toggle in, sign in and you’re in. You can ask your data. You can filter it. There’s going to be less you know management.

Meher Patel 00:20:47  It’s going to give you response in a very quick manner because the context layer is enriched. It understands where it has to go. It understands where he has to find those things. It’s not even just the columns which and the rows which has to be understood. It is about the entire report, and it is about how this report talks to the other report. How an SKP report talks to a search term report. And they both have its own context, which has to be married at on top when there is a cross link required. So those things don’t exist. Amazon Ads API at this moment has zero contacts. I mean, because it’s in the first phase, right? But that’s a very raw data. But that’s good. I mean, nobody has even launched an MCP APIs. No platform whatsoever has launched an MCP API, but Amazon has at least done that. But it’s not a it’s not for Amazon sellers to go and build that. I mean, the API existed long back for everything, but does an Amazon seller has to go and build an API? They’ve used tools, right? Which has empowered them and enable them.

Meher Patel 00:21:46  And that’s the point right now as well.

Josh Hadley 00:21:48  So what are your what are your recommendations for sellers then. So we’ve talked about yes. Use an MCP because it provides the context. It helps you with kind of like token exhaustion so that you know you can efficiently utilize a tool like Claude, maybe Claude Cowher, Claude code, etc.. So tell me, like what do you think are the best strategies the top brands are going to be using? Is it, hey, you know, connect to Hector MCP so that you’re getting the Amazon data and then you create your own custom agents based on your existing workflows that you have today that begin running. Like, what do you think that looks like and how would you recommend sellers set that up for themselves?

Meher Patel 00:22:27  Fantastic. Great question Josh, because everybody is right now feeling overwhelmed and excited with the access to MCP, with the access to ChatGPT cloud data, uploading that data, they’re seeing some some insights, which is obviously, you know, very broad insights and stuff like that.

Meher Patel 00:22:45  But there is not some richness in that. But that’s that’s just table stake right now. I mean, how far are you going to keep experiencing talking to your data that’s not making anything efficient? That’s not I mean, you’re just you’re just just enjoying that excitement right now for the new tool which is available in the market. But the real efficiency will come when you build your own agents, when you’re going to build your own workflows, when you’re going to start taking. So imagine yourself that you have an Amazon ads manager in your organization and you’re an Amazon seller. You have run those campaigns yourself, and then you teach your Amazon ads manager of what he needs to do, even if he has that knowledge. You will guide him with a strategy. Now, when you are building that strategy, that strategy stays with you. The way I manage my campaigns for my clients is probably different. And you have an approach. I have an approach. The fundamental approach would be same, but to go forward with it, it could be different.

Josh Hadley 00:23:49  Now, if you ask 100 Amazon sellers the best way to optimize PPC, you will get 100 different answers. And guess what? And it works. It works for different brands. And here’s the interesting part. I almost feel like one of the most difficult things to kind of like automate in terms of AI right now, honestly, is Amazon advertising because the nuances are absolutely incredible compared to routine tasks like, hey, go check a flight, check the price when it dips below this, buy it for me. Like that’s easy automation like it’s very set variables. Not a whole lot of nuance happens there. But it’s the jungle of Amazon. So continue on. But like, you’re right, like ask 100 different people. There’s 100 different ways to optimize those on ads.

Meher Patel 00:24:32  Absolutely. And that’s exactly what an agent workflow is going to unlock. For every 100 sellers, they can now rely on building a workflow that they desire, which is important for their business. Unlike the SaaS platform, which is build it for everybody and they have no restrictions, no limitations of any, any aspect of this rule does not work because I don’t have this aspect here.

Meher Patel 00:25:01  I don’t have the ranking gear, I don’t have the search query performance talk to my search term yet. So it’s it’s it’s not there because the platform is not built it. But MCP unlocks that. When you have an MCP, it actually understands all the siloed data. If you have done that Richmond, once it does that, you can build your own workflow. It depends where you want to build it. You can build it on any tent. You can build it on cloud. I mean, code works. You can do schedules, you can do you can use any agent workflow building platform. But the one important thing that you need to have is going to be that MCP, which is able to connect the dots for you and build that journey for you, and then you are less dependent on any platform which has built that journey for you, or which has built that platform, platform navigation for you. You build your own navigation. You build your own custom dashboard. You build your own visualization.

Meher Patel 00:26:00  I mean, you connect your MCP to lovable. You can make your own SaaS platform, by the way. I mean, that’s for real, Josh. I mean, if and Hector MCP connects to lovable, which makes their own dashboards and even an app with an Hector MCP, you can build your own app. And very soon, maybe in the next couple of months, I can take you to a live demo of how you can make your own app with your own data shown on your phone whenever you open, and you can launch it too. I mean, there are tools which allows you to launch your build and launch your own app. All it needs is an MCP plug in to that platform that you can build on.

Josh Hadley 00:26:38  Brilliant. So ultimately it allows you. And here’s like if we were to boil this down to the basic steps. Anybody that’s currently optimizing their own ad campaigns or they currently have given it to a VA or somebody or even like an agency. The most important thing for you to do is like if you’re working with an agency today, like understand their workflow, understand how are they currently optimizing the ads, and then this is what you’re going to do, whether you’re optimizing it, a Vas, optimizing it, it’s all going to be the same process.

Josh Hadley 00:27:10  Go do the thing yourself first. Record a loom video while you do that, and while you record the loom video, talk out loud. Then guess what you can do. You can download the transcript. You can even upload the loom video. Give it to Claude and say, hey, this is step one of my optimization process. Tell like, give me the SOP and the written instructions for an AI agent to go execute this and ask me as many questions as you need to to be able to ensure that this AI agent does this every single time without fault. Right. And so you build this like step by step. Here’s I think, and then you just you do this on repeat my hair. This is the challenge that I think most entrepreneurs are running into. And where most people are failing with the AI implementation because they’re so used to just going to ChatGPT and asking one question and getting all the answers. Well, what I just said is going to take you months, a few weeks of time.

Meher Patel 00:28:05  Yeah.

Josh Hadley 00:28:06  Because you’re building it, right. Like you’re building it step by step by step. That’s custom to you, custom to your brand. But most people just want to say, well, surely all I have to do is just set up an open account and connect it to Amazon and just let it do its thing. And over the weekend, I should have this thing running and I call BS on all of that because I don’t think that’s the way AI is supposed to operate. And and frankly speaking, those that are doing that and think that they’ve got it set up over the weekend, I love it because I also know secretly AI is actually probably ruining their business and starting to run it off a cliff because like, there’s no way the AI is AI is very confident in its decision making, but you’ve eliminated all of these, like micro steps that are going to allow you to have so much more context, actually make the right decisions, and you have to set up guardrail after guardrail after guardrail on on any of your AI agents to ensure like they don’t cross this step.

Josh Hadley 00:29:04  And so if you try to have them do everything in one, like, hey, optimize my ads and improve my ranking, and while we’re at it, launch this new product. If you ask that all in one prompt, like, good grief, yes, it gets done, but not in the right way.

Meher Patel 00:29:19  Right?

Josh Hadley 00:29:19  Anything you would add to that?

Meher Patel 00:29:21  Sure. I mean, you’re going in the right direction, Josh. And in fact, you know, there is something. As something else that is very important now is connectors in cloud and ChatGPT. The very recent connector which has got added is knit n and make com. So now you don’t even have to open an A10 to make an intake workflow. You can talk to Claude directly and you can say hey you know what I’m looking at? Building an agent who can go and check my budgets for all my campaigns and see how many of my campaigns are utilized. 80% of the budget and the Roas of that campaign at this moment is about five, and the orders are more than ten.

Meher Patel 00:30:04  If that’s the case, I would like to increase my budget by 50%. Now, if all you do is this. First you make the MCP. Check that. Now this is what I do. I make the MCP check that rule right now. So I want you to go and check all my campaigns. With the budget utilization above 80 and the spins and the rewards is above five and the orders are more than ten. Go and check it right now and let me know how it’s going right now. How many campaigns do you see that you’re going to increase it by 50%? Boom! It’s going to give you a response. With Hector MCP able to understand that thing, it’s going to say these are the five campaigns which is going to require to be changed. Fantastic. Now I want to make a niche and journey right now with the same rule. And I want this to be scheduled every three hours. Boom. Because you’re entertained is connected to cloud or ChatGPT that you’re using. You’ve already identified the schedule.

Meher Patel 00:30:56  You’ve identified what you want to do. You’ll be surprised the journey is so seamlessly created that you don’t need to even learn how to build a workflow. That’s how powerful it is.

Josh Hadley 00:31:08  What? Like what type of timeframe should sellers like be thinking about this in terms of how long do you think it takes them to build a system in a process that could actually run on its own and actually execute well. Are you? Could it be done in a day? Or is this like a period of time where it’s like, look, you’ve probably have hundreds of SOPs as to why you make different decisions. Like, is this something that you think gets built out over a week, a month, a few months? Because I also know the other most important thing is like, you got to QA, check this. Like just because it executes doesn’t mean it’s executing properly. So there’s a lot of like, okay, this worked. This did not work. Now I’ve got to go update the instructions. Oh okay.

Josh Hadley 00:31:50  I got to go ask questions. Why did you do this. So it’s like it is a game of refinement at the end of the day. So like how like is it ever done or are you always having to refine your AI agents. And it’s like it’s like human beings, but you don’t have to deal with like their pet dying or being sick. It’s just like you’re always giving them feedback. You always have to watch and double check their work. Like what? Like give me some of your thoughts on that.

Meher Patel 00:32:14  Oh, fantastic. No. So it’s not a weekend job, it’s not a one month job. But I think that one skill that every Amazon seller has to learn today and they have to start today, is to how to communicate with AI and how to understand the potential capabilities of an AI, how to talk to them, what makes them do when you say something? Because every time you say something, it reacts differently. How often do we go ahead and ask an AI, hey, why did you answer this to me? I wasn’t expecting you to respond like this.

Meher Patel 00:32:50  And then he’s going to say, hey, I thought that this was something that you were looking for. I had this in my memory. Hey, I had this in my files as one of the frameworks that you find. Then you’ll be like, all right, so let me double down on that framework that you’ve got. Can you list it down? Because that’s not what it should be. Then you change that framework and you put the device that framework, put another version and download and upload that into your project. Because once you build that framework upon which you’re going to start building your agents, you’re going to be in a massive autopilot mode very soon. So imagine you’re building an engine, but if you’re building an engine which is going to run, which is going to speed up, you need to take those baby steps for that foundation to be super sexy, it has to be rock solid. So I think more than building an MCP or more than doing any of those things, the most important skill is to know how to talk to an AI, right? Once you learn that the building of AI agents is going to be so smooth and you’re going to do a great job building the agents, right.

Meher Patel 00:33:57  And one important word that I would like everyone to also understand. You’ve got to become an orchestrator. You you’re the one who’s going to build agents and you’re going to you’re the one who’s going to orchestrate that agent. So you’ve got to start visualizing that. what are those small, small tasks that an independent agent should do? Or like you said, that one task, one agent should do five tasks. Are they connected or if they are not connected, you don’t make them be done with one agent. So that’s the time that needs to be invested with a with an A by an Amazon seller. You’re going to waste your time. Every seller is going to waste their time building it rather than leveraging it. So we need to learn how to leverage it, how to use it and then building it. And once you find that right MCP, you understand how cloud or ChatGPT or the LLM that you’ve decided that you want to use and you know how he reacts to your questions and to your MCP. The life is going to be easy, and it’s a journey.

Meher Patel 00:35:04  Like you said, it’s a journey of refinement, but it’s going to come to you very soon. You’re going to become good at it.

Josh Hadley 00:35:08  Well said. And I totally agree with you. And this is my stance on AI. If I want to begin creating my own AI agent to optimize my ads, I don’t expect that to kind of be ready in like to be almost like hands off for probably another 12 months. Yeah, that is the way I’m looking at this. And most people laugh and they’re like, oh, dude, you’re not doing this fast enough. Oh, you need to be spending more time. Like, what are you. Are you kidding me? I’ve got open cloth set up. And again, I call BS on all of this stuff because if you want to architect this properly and you actually want to have a true AI agent optimizing on your behalf based on your rules and your expertise and how you know to dominate the market, it’s going to take time to refine this. And this is ultimately the question.

Josh Hadley 00:35:56  I do think that my prediction, my hair, is this what the future looks like for AI agents? You can either loan an AI agent or rent an AI agent like you rent a house, or you could go buy the house aka or build the house yourself. Right. And so there’s I think that’s the best analogy I could share, is kind of like if you want the custom home of your dreams, if you want the AI agent of your dreams, guess what? You have to figure out the blueprint. You have to. You have to then begin layering in, starting with the plumbing and the foundation and and the walls and the electrical. And you’re going to have to figure all this stuff out, and then you’re going to run into roadblocks. It’s like, oh, this, this had an issue. We got to go correct this. What’s going to take longer? Well, the custom home is going to take longer than the. All right. Well I could go get you a mobile home tomorrow and it’s all yours.

Josh Hadley 00:36:48  Congratulations. But it’s very cookie cutter, and it just serves this in a purpose. So that is the decision that sellers are going to need to use. And I think it’s all going to come down to prioritization. Is your Amazon ad strategy an actual moat and a defense around your business? If so, maybe you need to go more towards the custom building route for an AI agent. If it’s more of an afterthought and it’s more of like, hey, you know, it’s not my core business. I think this is probably like, if I were to guess out loud here, I would say for big multinational brands, right? The Procter and Gamble’s of the world, things like that. Go ahead and take the cookie cutter Amazon ad strategy that you know, minimizes your ads, optimizes for tacos and ACOs. Just run it. Good. It’s good enough because guess where the core of their business comes from? Retail halo effect from retail onto Amazon or from DTC onto Amazon. So for them like Amazon’s quote unquote an afterthought, even though it can be driving meaningful revenue.

Josh Hadley 00:37:49  But for people that are Amazon first, like I would say, like your ad strategy is like arguably should be one of your moats. And so you’re going to need to go that route.

Meher Patel 00:37:59  indeed. I mean, so well said. I think if you need it and you will need it. And if you believe that your moat is your strategy for your brand and you understand that category? No. Than anyone else. Then the point is, are you ready to build your own genetic workflow? And if you’re ready. Do you know how and how an LLM responds to you? Have you learned that skill? Have you practiced it enough? So that’s super important for for it to come. So for sure.

Josh Hadley 00:38:29  Yeah, I think the final thing before we wrap things up here is every brand owner needs to decide today. Are you a physical products brand owner or are you a SaaS provider. And I ask that question because yes, you can go set up your own AI agents and yes, you can go custom build and code your own dashboards.

Josh Hadley 00:38:54  But guess what’s going to happen? Just like we discussed today. Do you think that that. So here’s the best analogy. And I see this in almost every Facebook group that I’m a part of right now in the Amazon seller community. It’s hey guys, I just built my own dashboard in Claude. I. It took me the weekend, but I built it. I no longer need seller board or hello profit or whatever dashboard it was that they were looking at their metrics on a daily basis. Guys, I no longer need that. Guess what? It was only costing you maybe 100 bucks a month at best. And this dashboard I know will break next month. It will break something from the API is not going to be fed in properly and the dashboard is going to break. So then you’re going to have to spend more time debugging it and dealing with it, because all of those software companies deal with that on a daily basis. You just don’t see it. You just think, I’m paying these guys a hundred bucks a month, and they’re just sitting on their hands like they’re just this is easy business.

Josh Hadley 00:39:50  And it frankly speaking, it’s not. And so I think like it is a rabbit hole that I would caution many sellers from going into. If you are technical and you enjoy that process, just understand like you’re going to need to have like a CTO, you’re going to need to have like software engineers on staff, even if you can use cloud code, because like you’re entering the game of SaaS and you have to ask yourself, do I want to be a SaaS owner or do I just want to be a brand owner, a physical products brand owner? So I think that’s the key thing. Would you allow anything else to add on that 100%?

Meher Patel 00:40:24  Bang on. Josh, I think I don’t want to add any of those things. You said. It’s so well that the seller has to understand this. He has to make a choice. He or she has to make a choice today because it’s going to hurt. If they go that route, they might just not be able to. They have to choose.

Meher Patel 00:40:41  It’s a trade off, right? It might be exciting, but it’s a trade off and they have to take a call.

Josh Hadley 00:40:45  Well, and if you’re going to lean in one direction, you got to go hard in one direction or the other if you’re going to be like, well, no, I use AI for this and I’m halfway in. Okay, well you get half the results. If not, you get like a quarter of the results because it’s not even well optimized. So again, it’s going to be fascinating to see how A.I. changes things. By the time this airs, there’s probably something else that is already been a game changer in the AI space, but.

Meher Patel 00:41:13  We go through it as well. I mean, like I said to you in the past that we first started with the problem of prompt engineering of what to ask, how to ask, how to revise our ask. So that’s prompt, right? So how do you devise the prompt? How do you talk. Right. How well do you talk.

Meher Patel 00:41:29  But now it’s a problem of context engineering, right. Is your context floating is too much. Context is not good because it’s going to hurt the prompt that is going to happen. At the end of the day. The user wants a quick response. The user doesn’t want things to fail. The user doesn’t want hallucination, the user wants the right prompt, goes to the right tool, goes to the right data. So that’s context. That’s context engineering. Which like if you’re doing an orchestration, when there is a prompt, there is a question related to a certain aspect of your business. There are multiple tools inside an MCP. Which tools should I go to, and should I check each tool together to find an answer? Which one should I go to? Because if I have 70 tools, if I’m going to touch all of them to see which one is the right one, I’ve spent 70 attempts. You can imagine the token which is going to get exhausted. You can imagine the window which is going to get exhausted, and that’s the responsibility of an owner to deal with.

Meher Patel 00:42:31  And that’s exactly where you don’t need to get into doing that, because that’s technical. But what you need to do is is it responding well, what is it that it’s understanding? What’s my next strategy that I would have an agent orchestrated for me? So those are the things that needs to be dealt with by an Amazon seller. And you said it’s so right. Yeah.

Josh Hadley 00:42:55  Fantastic. Well, I love to leave our audience with three actionable takeaways from every episode. Here are the three actionable takeaways that I noted. You let me know if I’m missing something. Action item number one. It doesn’t matter where the world of AI goes, one of the most important things that you can start doing today that will be beneficial. Whether you want to fully adopt AI, or whether you’re going to stick your head in the sand and not fully adopt AI. But it is this take, take whatever process or whatever thing you need to outsource or get off your plate and do it 3 to 6 times yourself and record your you doing it, and narrate the process while you do that, just using a tool like loom.

Josh Hadley 00:43:37  And then you have some context that you can upload to an LLM of your choice and say, hey, here are six different videos of how I’ve executed this process. Tell me the different nuances, help me create an SOP around this. And so I think like the days of like laboriously documenting and grabbing screenshots to like create an SOP, it’s completely gone out the window and it’s actually going to get even better with these llms that like have more context and understanding. It can help you refine it ten times faster. So that’s action item number one. Whatever it is that you don’t want to be doing or needs to be moved off your plate, begin creating SOPs. And even if you say you want to get into the AI world, that’s fine. Before you can even have an AI agent set up, still document that dang process six times, then give it to the LMS. So that’s action item number one.

Meher Patel 00:44:26  Sure.

Josh Hadley 00:44:27  Action item number two is when you are using an LM, the lazy way of doing this is downloading your.

Josh Hadley 00:44:36  Whether it be Shopify data or just downloading your Amazon keyword report data and just uploading it into an LLM and saying, hey, what should I do? Help me optimize my ads based off of this data? Because and I hope this is the understanding the mindset shift is more than anything in action. Item number two is understand. The context is so vitally important When you are utilizing these AI tools because like they’re going to come back very, very confident in their recommendations. And so you don’t know if it’s trending in the right direction or not. And so all the more reason you need to provide more context. Obviously that’s the reason why you need to have an MCP. That your candy bar analogy is a great analogy. It’s like you don’t know what what chocolate is is inside there if it doesn’t have a wrapper, if you just have ten chocolates just sitting on the table, one of them could be jalapeno flavored and the other one could be sweet. Or, you know, semi-sweet chocolate doesn’t matter. But like they all look the same.

Josh Hadley 00:45:34  And but if you’re missing the context, you’re like, oh, I hate jalapeno. And now my mouth is on fire, right? And like that’s the problem that’s happening with most people. So utilize an MCP. And then my third and final action item here is if you do decide to go down the AI route, right. And you are going to build custom AI agents, just understand that you are going to begin entering the world of what other SaaS providers have been dealing with for for decades at this point. And so understand that if you want to be AI first, you are also like probably going to need a CTO, right? That’s going to be the architect of all of these systems. Your data integrity and your security of that data becomes a paramount issue in your business. And so there’s a lot of different nuances that that go inside of that. So I want brand owners to understand, I think the what the future looks like is there’s going to be AI tools that you’re going to be able to rent out and be able to say, hey, I’m a physical products brand first.

Josh Hadley 00:46:35  Like I’m just leveraging the AI tools without me having to maintain support. And then you’ve got other people, on the other hand, that are going to try to navigate the I’m a physical products brand owner and now also a SaaS provider, even though I am the only client of that SaaS tool. So make that decision now, and I think it will also reduce your FOMO when you see other brand owners saying, like, I built this sexy dashboard and it’s like, great for you. Not a game I’m interested in playing because, like, my time is more valuable than saving, you know, spending 20 hours to go code the dashboard. That saves me $100 a month.

Meher Patel 00:47:10  Yeah, yeah. So well said Josh, I think that’s, that’s a, that’s a lot of juice for, for your audience there. I think even if they do like the first two or even one of them, that’s going to make a lot of difference for them. Fantastic.

Josh Hadley 00:47:22  Well, here are my final three questions.

Josh Hadley 00:47:25  Number one, what has been the most influential book that you’ve read and why?

Meher Patel 00:47:29  I think Steve Jobs by Isaac. And the reason is, I think I don’t have to say this because the world knows this. The way Steve works, right. The way he did what he did. I mean, you can’t build that sort of a company without being relentless, without being so dedicated, without being leading the world aside, but focusing on building something which is impactful. And I think that’s, that’s that’s the ideology that I live with. I believe that if you really want to build something which is impactful, and if you want to do that, you need to sacrifice and you need to stay focused and you need to be relentless in order to achieve that. And that’s a very radical way of doing things. But I think that’s that’s the way I live my life, too. And and I’m just living it right now and I enjoy it.

Josh Hadley 00:48:19  Yeah. Great recommendation. Okay. What’s your favorite AI tool that you’ve been using and how have you been using it?

Meher Patel 00:48:25  So I think one one good tool that I believe that everyone should start using, and it has changed the way I’ve been communicating with AI is whisper Flow.

Meher Patel 00:48:35  It’s outstanding. I don’t remember typing on my laptop in the last 30 odd days that I’ve started Whisper Flow, and I get on like, the longest conversation with Claude. I mean, that’s the place that I use, and it understands me so well that it’s just amazing and it just takes your hands off keyboard and you can paraphrase things and it just does a great job. And I think that combined with anything that you’re doing in AI or whisper flow in it, it’s just going to be like five times more efficient.

Josh Hadley 00:49:14  Great recommendation. Thank you. Final question here for you. Who is somebody that you admire or respect the most in the e-comm space that other people should be following and why?

Meher Patel 00:49:23  I think Jeffrey Jeff Cohen is someone that I have introduced myself to the Amazon world with, and I think he’s always got something more juicy. He is always in the weeds. He understands how this ecosystem and landscape of retail media network and Amazon first as what it was for him. I think he’s he’s he’s got his hands super dirty and he knows things at the from the bottoms up.

Meher Patel 00:49:50  So I look up to anything that he says and I love to listen to him on, especially on LinkedIn if he’s around. Yeah for sure.

Josh Hadley 00:49:58  Amazing recommendations. Well, Maher. This has been a great episode. If people want to learn more about you, your agency, the MCP model, where’s the best place people can go.

Meher Patel 00:50:07  LinkedIn and just DM me on LinkedIn. I’ll be happy to respond. And I love talking and I love co building. I keep saying to everybody that we are not building, we are co building because it’s everything that has been built in Hector. It is actually built because we’ve had problems, so we’ve built it. So we’d love to build anything that a seller wants to do and we’d love to talk. I’m available on LinkedIn. Please DM me. Yeah.

Josh Hadley 00:50:29  All right. Well, thanks again for your time today.

Meher Patel 00:50:32  Thanks a lot, Josh. Nice. Thank you for doing this. And thank you once again. Yeah. Bye bye. See ya.

MC 00:50:37  Thank you for listening.

MC 00:50:38  Visit ecom breakthrough com for more information. If you’ve enjoyed today’s episode, the best way you can show your appreciation is by clicking the subscribe button and quickly leaving a review. See you again next time!