Why AI Is Hurting Your Reinstatement Odds

[00:00:00] Hey everybody. Welcome back to Seller Performance Solutions. Chris McCabe here, former seller performance investigator at Amazon, or Specialist as I think they said back in the day going into trends we're seeing, appeal problems, appeal mistakes. It's what I'm gonna focus on today. But with taking the angle on some of the AI content that we're seeing.

And also just some of the comments I'm seeing in our inbox in the messages I get, whether it's email, WhatsApp, or even just phone calls, that revolve around a dependence or an over-reliance on. AI slash chatt. I've talked about this, before, but I want to focus specifically on the idea that people are using a tool like chat GPT to scrape info off the internet.

That's what it does. That's what it's for. And by definition, I think it's dated information. It's not necessarily a hundred percent up to date information, which is [00:01:00] one reason a lot of sellers tell me that they sent in a plan of action to Amazon. Using some sort of yardstick from two, three years ago that they found online just by using AI tools and POAs or a plan of action didn't even apply to their appeal.

Well, that's because it's out of date information and the problem there being, they're just taking what they find online or on ChatGPT and copying and pasting it into whatever strategy they're using, whatever plan to appeal they've mentally concocted. Without vetting it, without tweaking it, without editing it, or without understanding that they might be basing their strategy on either out of date information or inaccurate information.

ChatGPT is not a crystal ball. You still have to apply some brain work to how you're gonna use it, why you're gonna use it, when you're going to use it, what topics it'll be best for. And I just see an increasing number, I assume it'll eventually be daily, but an increasing number of people every week [00:02:00] coming to us saying, well, I researched the viability of an appeal on ChatGPT, and it looked like the odds were low based on what I was seeing there.

Well, of course, in that type of case, people are more likely to post online. Because they have a problem, because they've gotten stuck because they submitted some poor appeals previously that didn't work out. It doesn't mean that that's a raw number of the odds or percentage of success. So that's one major problem I want to talk about this week is the over reliance on AI to determine the odds or likelihood of success of an appeal for certain types of compliance appeals for different kinds of ASIN takedowns, even account suspension type problems. Definitely the wrong time to just turn your brain off and let AI do the thinking for you, because the content that's going to get scraped using any AI tool is more likely to pick up info about people who are stuck, people who appealed poorly, people who didn't understand the nature of [00:03:00] the problem, people who might have posted on seller help or, the seller forums. In my experience, which is, 11 or so years of doing it our way, as ecommerceChris, but also for five to six years before that at Amazon, in my experience, the people who go make these posts or go to these types of resources are the ones who need it the most and who are the most confused or have the least valuable case study material for how they got into the fix, how they got deactivated, why their appeals are getting denied. The types of research they do, whether it's calling account health or otherwise, talking to their account managers, their SAS core, the research into why their appeals were rejected, or how a subsequent appeal could be improved, those are the ones most likely to be posting inaccurate information that you shouldn't be relying upon. And if you're not vetting it appropriately, if you're not attaching the right [00:04:00] filter that takes bad info and puts it in the bad bucket, worthwhile info, puts it in the good bucket, then you're wasting your time and you're relying on tools that aren't quite there yet.

Maybe 2027-2028, this will be different. I don't see it trending that way this year, and I don't see the behaviors around ChatGPT use as it applies to appeals work. I don't see that too promising at the moment because people are simply mentally copying and pasting these concepts around, maybe even copying and pasting text that they find from an appeal.

So it's not just doing ChatGPT research to find out if it's worth appealing. I think most sellers, especially, most brands that we talk to or that I talk to, they've already decided that it's worth it to appeal. That's just one example. Understanding the odds of an appeal or the odds of success ChatGPT is definitely not worth, spending your time on for that, for the appeals themselves.

The second thing I want talk about [00:05:00] today is the commonality or the likelihood that I hear from a seller that they've done ChatGPT appeals research and they found good content. In terms of what can go to into an appeal that is analogous to everyone in the last few years who would find templates, right?

We always talked about use of generic copy and paste templates. Sellers, were kind of trading them around. Sometimes services were just pedaling templates. Most of those have kind of receded into the background or gone away at this point. Thank God. But they were doing the same thing, right? AI's kind of replaced what the Chop shoppy, copy and paste services were doing with templates, which were just pre-written documents or pre-written appeals, and they were adjusting the slightly, that probably won't go away entirely for a while.

Sellers are still doing it themselves, and some sellers just take a previous appeal they wrote in-house and copy and paste it. Probably using an outdated format, which is a really bad way to do it, of course, [00:06:00] but making minor changes or tweaks and then just kind of throwing it at the wall to see what sticks, which as you know try that once fine.

But as you know, we argue against that because we don't see that you have an limited amount of time or you're losing revenue every day pretty much. Why take the risk to the pounding that it's gonna do on your account health. Why take the risk that you're gonna lose all the additional ASIN in revenue by just kind of winging it or experimenting?

I understand it's cheaper to do it that way and sometimes it's better to do it yourself than to hire one of these copycat generic content services. I mean, there aren't that many of them left anymore. They tend to be lawyers or say that they're lawyers and they're not, and I get that.

But they're still out there. And if you're doing the same mental approach or strategy as you did with the templates, but just swapping it out for ai, you're really just kind of robbing Peter to pay [00:07:00] Paul at that point. You're not improving your luck or your odds, you're just swapping one weaker strategy and potentially lazier strategy for another.

And when I say lazy, I don't use that word lightly. It's not that there's no blood, sweat, and tears going into the assessments and the strategies and the appeals and reworking somebody else's appeal. I'm saying it's kind of more intellectually lazy. Like you're assuming that something you find on ChatGPT is gold. You're assuming it's accurate, you're making assumptions you can't afford to make if it's a high revenue ASIN that Amazon is about to suspend or has already suspended. God forbid you do ChatGPT appeals with account reinstatements. I've seen, I don't know, maybe an uptick of 10, 20, 30% more people using ChatGPT to write an account reinstatement appeal than I did maybe this time last year. I'm not a fool. I understand why that is. People are AI all the [00:08:00] time. Every time you look on LinkedIn, WhatsApp groups, masterminds, Facebook strategy groups, whatever it might be, it's AI all the time.

I understand why people are embracing it. I understand why people are tinkering with it. But when cooler heads prevail, as anyone with any expertise in AI will tell you it's the quality of the prompts you're using. It's the quality of the information you're getting back. How do you assess it?

How do you examine it? Do you question it? Do you fact check any of it? I know there isn't enough fact checking going on because people are putting it in front of me, and I'm sort of becoming the defacto fact checker for a lot of appeal related content or question and answer and comments that people are finding on ChatGPT.

And it's low enough quality for me to be concerned about it, to make a podcast episode about it. And we're in January right now. You're going to be hearing me talk more about this on LinkedIn, on the podcast in [00:09:00] Suspension Trend videos, and just in anecdotal experiences reading people's appeals, which is what I do at least half of my day.

So, after, 16 plus 17 plus years of doing it, I would say that I have the experience to see what Amazon's more likely to accept and what they're not going to waste time on. So maybe I'll conclude this week's episode with just, why are you motivating Amazonians to ignore you by sending them something that they can tell is ChatGPT with their own, interpretive tools.

Or they can just tell it by eyeballing it. I can tell it's ChatGPT because people are. Copying and pasting, maybe not the entire appeal, but a lot of it, and you can just tell by how it's written. You can tell if it's been tweaked from previous ai. We've all been on LinkedIn. We've all seen AI posts that have been slightly tweaked, but it's mostly ai.

I think the [00:10:00] same principle applies here, but with a LinkedIn post, there's much less at stake, right? With an Amazon appeal for reinstatement or to counter a performance notification. That's kind of my closing question for this week. Why are you motivating Amazon to not take you seriously? Why are you motivating whoever's reviewing your appeal?

I'm not saying it's a top-notch employee. I'm not saying it's somebody that won't necessarily be laid off or worried about being laid off in the near future. And I'm not saying Amazon's doing impeccable work and they couldn't improve. What I am asking is why are you pushing pressuring, prodding them to reinstate you but you're only halfway there on the content? And that's putting it charitably. Some of the appeals I see are so obviously composed from a multitude of sources, a mixed bag of good and bad, or it's just so obviously derived from AI that you're motivating the people, the audience that you're sending your [00:11:00] appeal to.

The Amazonians on the other side to ignore it, to ignore you, to send you a copy and paste generic response, which is right at their fingertips, right? If they know they're not going to accept it after reading two or three lines or the first paragraph. That's why so many sellers complain about getting generic denials.

You're giving them an excuse to use those more, and that's not what you want to do. It's hurting you. It's hurting other sellers who are putting time, effort, and energy into properly constitute appeals and you're making it easier for Amazonians to kind of punted away and not do anything, and you're kind of numbing their brains to the point where they can just knee jerk, reject everything.

That's the ultimate consequence in my humble opinion. The real long-term damage. Forget the short-term damage of having one or two appeals denied that you really wanted to see accepted. The long-term damage is you're perpetuating this cycle of garbage in, [00:12:00] garbage out sellers sending in bad appeals, Amazon sending bad responses out, and everyone sort of just psychologically gets accustomed to rejection and denials.

And I would not recommend that you take the idea that AI is reading all the appeals. So your AI is just talking to their AI and there's no point putting any time or work or any effort into it because they're just either A, being pessimistic and saying they're going to reject it no matter what, or B, assuming that no one's reading it, we're not there yet.

Come back in 2027 or 2028 when you know that their AI is reading your AI. That's not what's happening in a lot of the escalation cases that we look at because a lot of people just doing run of the mill seller central appeals aren't necessarily bringing us in or hiring us on at that stage. I wish they would.

They're looking at escalations and a lot of people are scraping the internet with tools like ChatGPT about [00:13:00] escalation appeal content. So imagine putting together, even if you waste it by emailing Jeff, which I'll be talking about in another episode soon, even if you waste the appeal, if you're sending any kind of escalation anywhere at Amazon, whether it's an escalation email queue or to a person or to a team that you've dealt with previously or you heard about, over AIing that type of an appeal where escalation means you are almost definitely getting it in front of a person and you're treating them with the discourtesy and the ignorance and the flippancy of I know a person's probably reading this and I know their manager level or maybe even senior management or above, but I'm still going to send them the same copy and paste drivel that I send through Seller Central. You're guaranteeing yourself failure at that point, and you at that point, you really only have yourself at to blame.

So that is my question for anyone who wants to post in comments or get back to us with some [00:14:00] personal experiences or feedback in this arena. Why aren't you taking it with a grain of salt? Why are you assuming that just because ChatGPT is a common tool to use, that it's gonna be golden for Amazon appeal or even appeal denial content.

I don't understand it, so it's a real question. I'm not just being sarcastic for once. Definitely throw that at me. Show me some examples, because I have a feeling that if I come back on this topic in two months, there's going to be more negative experiences and more help to pay in terms of sellers getting rejected multiple times before I even get eyes on it and tell them to stop doing the ChatGPT or watching it turn into some sort of a knee jerk addiction in terms of getting information on Amazon appeals and Amazon escalation processes because we are not at the point now where you can just sort of mentally offload that entire process onto AI [00:15:00] and hang out and just wait until you get a response back because that's not the way this type of strategy works.

So questions on that, let us know. As always, thanks again. For listening in on Seller Performance Solutions,, and I'll be back soon with another wonderful topic. Good luck. Take care.