I haven't lost a customer service fight in seven months

I wonder what kind of life the author has in order to start 9 disputes worth $12k in just 7 months.

The only dispute I had in the last year is some $3 tweezers I ordered on Amazon that never shipped, and Amazon automatically reimbursed the money.

I saved about $65k this way, having Claude guide me through the process of assuming the (very low interest) mortgage on my house instead of having to refinance it under current rates. When I first contacted the mortgage servicer they said my loan wasn't eligible. I used Claude to do the research citing why my loan did qualify, drafting the letters and request packages with all the supporting evidence, directing me to the proper government complaint portals when the lender still tried to stymie the process. Cost me about 5-6 hours of time over all and will save me >$60k over the course of the loan.

I had a dispute with Canada Revenue Agency and won. A good part of a decade ago now, I got into contracting for a couple of years. The accounting lackeys at S_____ W_____ in Richmond, BC, Canada paid me using cash method accounting. They marked down payments to the date they actually paid. That was often late, so for instance December money would end up in January, and since that's when the money changed hands, it was now next year's income. They reported it this way to the CRA (unbeknownst to me, who would not suspect such a practice at this level of business). In one year, November and December payments were paid in January, and accrued to that year's income from their side.

In my self-made accounting system, written in a self-made Lisp, I used accrual accounting, associating all transactions with their correct dates, and remitted taxes accordingly.

This resulted in some amount owing (looking like underreporting).

I used reports from my accounting system to document my case, and they agreed with me: I did my accounting and reporting the right way, and that company should not be using cash-based day-money-changes-hands accounting.

I got 900 euros refunded for a delayed flight with the help of Gemini. It was a two step process first to raise to Turkish airlines who refused, and then fill in a form in Turkish to some ombudsman with a bunch of screenshots/proof and citing the rules. I would not have the cognitive space to deal with it alone.

Mostly this leaves me concerned about an escalating war between companies intentionally introducing friction into their customer service pipelines to discourage expensive customer behavior, and customers increasingly relying on more and more AI automation to deal with that friction.

In a perfect world this would cause the intentional friction to be less and less effective, while laws and regulations stop the companies from implementing more and more aggressive policies/systems, leading them to go back to the fastest possible pipeline to capture those savings if nothing else.

But I am worried that is not the world we live in and I am going to try to return some shoes I bought in the wrong size and be asked to do some integrals before I am allowed to.

I applaud you. Over the past year, more than a few companies have become impossible to deal with over issues of their own causing, and issues that never would have come up in previous years, now do. They just don't seem to care at all anymore, and seem completely checked out.

In the past year, I've had to spend quite a few hours just finding replacement parts and solutions for issues manufacturers were responsible for and would have handled previously. I may have to reevaluate and have AI handle more of these matters.

LLMs are a great equalizer. I had no idea how much of the world runs on ignorance of one's own rights and how much administrative burden is used adversarially. Now my learning is greatly accelerated. I hope it's the same for more people than not and we will end up with a more just society.

A smart but wrong path to solve the problem. We need competitive markets that would force corporations be responsible instead of an extra layer of complexity in the already complex consumer world.

I applaud to the author for creativity and I'd be tempted to try (elements of) it in practice next time, but I don't see anything good coming out of it at scale.

Great content! Inspires me to do the same thing since I am constantly dealing with things falling apart like this. I recently did one manually after putting it of for too long. Get Delta (airline) to go from "here is a 200 dollar bag to replace your 400 dollar bag we damaged" -> "Here is a 450 dollar eCredit".

I would try focusing on building a more unique style of writing, this feels like it went through heavy alignment to Claudes distribution. For the general audience maybe they aren't that sensitive to it yet, but for those that have to use LLMs at work, having to read more of it is grating.

Reminds me of an executive at my company who has sent extremely inflammatory, needlessly escalating, LLM-generated messages to one of our vendors in support threads (for problems she simply does not understand, any prompting she does I'm sure is just to have her AI agent coerce a concession/credit). I recognize that there are companies with scummy billing practices, but people like this don't seem to care that there are human beings on the other end of these things who are scrutinized as though they are interacting with another human being.

Im sure I saved a few thousands of dollars when one of my 3 solar inverters failed, chatgpt read through the documentation, my power configuration and pointed out that i could safely reroute the connections from the failed inverter to one of the two inverters. Seeing they are high voltage lines, I opted to work with the original installers, who were reluctant to do so, and was pushing to just get a brand new model, after a few back and forth with chatgpt providing the exact documentation that it was safe and within the limits of each inverter, they agreed to do so, for a few hundred bucks. (which only took less than an hour)

Yeah. My first real use of LLMs was a couple years back when we had a legal fight. We never would have won without it. Even got the high school principal to apologize. That was delicious.

How many of the 9 ended with an actual human on the other side

Asking because from the seller side a lot of that friction is on purpose, and I'd expect companies to start answering agents with agents

I like you blog and tried to add it to my feed, but it doesn't appear to have an RSS or Atom feed. I know beehiiv supports feeds - any chance you could switch it on?

For some reason, Firefox thinks this article is in Portuguese.

Crazy having to fight with insurance providers to get basic health care.

[flagged]

> Author here. For seven months I've been handing every consumer dispute to a Claude agent instead of eating the cost like I normally would. It builds the case file, pulls the warranty or policy text and the terms of service, works out who actually regulates the company, and plans a "campaign" with fallbacks. Ten unrelated fights (IRS penalties, a car warranty, two insurance denials, a robot mop, grocery stuff), yielded just over $12k back! The biggest single one was an IRS first-time penalty abatement I'd never heard of, $6.8k, eight years into an installment plan.

> To be honest, the agent's first-pass advice is usually bad ("file an FTC complaint"); what works is multi-prong approach until something moves. For the robo-mop it ended up all the way at their own arbitration clause, whose consumer fee rules cost them several times the refund. For the pet insurer it was the state regulator plus two exec emails, and they relented in hours after two months of stonewalling.

> Companies rely on a drop-off funnel, and an agent with infinite patience breaks it - thats the story.

It's flagged and dead but good advice. I've used it a lot for legal disputes (basic contracts).

It helps to make your argument, then have AI steelman a rebuttal, then modify your argument to smash their steelman. Then your initial outreach is your stage 2 argument that bypasses all plausible initial denials.