If I Had to Get My Brand Cited by AI, I'd Build This Grok Bot — Transcript

Full transcript

1. 0:00I searched AppSumo for AEO. Zero
2. 0:02results. Meanwhile, over 30
3. 0:04venture-backed AEO tools have over 200
4. 0:07million in funding spread between them.
5. 0:09Profound alone has over 58 million in
6. 0:11funding from Sequoia and Kleiner
7. 0:12Perkins. I scraped AppSumo directly and
8. 0:15classic SEO tools have customer reviews
9. 0:17in the range of at least 130 all the way
10. 0:20up to over 950. When I look at AI search
11. 0:22platforms on AppSumo, [music] the number
12. 0:24of customer reviews falls in the range
13. 0:25of 32 to 49. Now, lifetime buyers can be
14. 0:28very practical when they find a software
15. 0:30that they really like, and it makes
16. 0:31sense. They focus on getting the most
17. 0:33value that they can. So, instead of
18. 0:34trying to buy the answer, I decided to
19. 0:36build it instead. I'm going to show you
20. 0:38my seven-bot rock bot system. [music]
21. 0:41Talk about a handful. Now, this system
22. 0:42is strategically designed to get you
23. 0:44cited by ChatGPT. I'll share the cost to
24. 0:46run it and, of course, any key findings.
25. 0:47[music]
26. 0:48Now, the reality is most of you should
27. 0:49probably stop watching and maybe focus
28. 0:51on fixing your product instead. However,
29. 0:53let me give you context as to why you
30. 0:55should listen to me and not the guy
31. 0:56trying to sell you a $99 dashboard. Now,
32. 0:58of course, before I built anything, I
33. 1:00did a pretty comprehensive research
34. 1:01pass. [music] I used Firecrawl for
35. 1:03indexing sources. I used Appify for
36. 1:05analyzing Reddit commentaries. This
37. 1:07resulted in analyzing over 142 different
38. 1:10posts and 14 different subreddits. We
39. 1:12also analyzed every major vendor study
40. 1:14that's been released in the last few
41. 1:15years. [music]
42. 1:16We also analyzed Profound's 680 million
43. 1:19citation data set, Peaks AI 30 million
44. 1:21sources,
45. 1:22>> [music]
46. 1:22>> Conductor's 238,212
47. 1:25different prompts, Umbrellas 120,000
48. 1:28mentions across five different
49. 1:30verticals. And if that wasn't enough, I
50. 1:32also scraped AppSumo's marketplace
51. 1:34myself, which nobody else had done
52. 1:36previously. But, here's my contrarian
53. 1:37take and warning. It's probably going to
54. 1:40annoy people who sell AEO services. The
55. 1:42dashboard is not the work. You see, this
56. 1:45entire category isn't failing because
57. 1:47the products are bad. It's failing
58. 1:49because the very thing they measure gets
59. 1:51fixed by product quality, review volume,
60. 1:54and PR. And none of that is a software
61. 1:57feature. So, a tracker hands you a
62. 1:59percentage that you cannot act on. What
63. 2:01I'm about to build and show you does not
64. 2:04tracked. It does the work. Now, with all
65. 2:06of that being said, let's get into it.
66. 2:08The first thing you need to realize is
67. 2:10that you can't build agents against
68. 2:12[music] a strategy you don't have. So,
69. 2:14three findings fast and then we build.
70. 2:16Finding number one, mentions beat
71. 2:19backlinks by about 3x. Ahrefs studied
72. 2:2275,000 different brands. Branded web
73. 2:25mentions correlate with 0.664
74. 2:28with AI overview visibility. Backlinks
75. 2:30correlate at 0.218. That's the That's
76. 2:33the whole structural shift in one
77. 2:35number. [music] Classic SEO is going to
78. 2:37reward what's on your domain. AI answers
79. 2:40is going to reward what the rest of the
80. 2:41internet thinks about you. A commenter
81. 2:44in the Reddit AEO thread put it better
82. 2:46than any vendor blog I've ever read.
83. 2:48Rankings reward your own domain. The AI
84. 2:51answers reward what everyone else says
85. 2:53about you.
86. 2:54>> [music]
87. 2:55>> So, the work shifts off your own site,
88. 2:57which means the agents I build cannot
89. 2:59just optimize pages on your site. They
90. 3:01have to go find and shape the
91. 3:03third-party testimony. Finding number
92. 3:05two,
93. 3:06reviews are a gate, not a gradient. Now,
94. 3:08this is the one that reframed everything
95. 3:10for me. From SOCi's 2026 Local
96. 3:13Visibility Index, AI platforms use
97. 3:16reviews as a confidence threshold, not a
98. 3:19ranking gradient. Businesses that get
99. 3:21recommended average 4.3 stars on
100. 3:24ChatGPT,
101. 3:254.1 on Perplexity, and 3.9 [music] on
102. 3:28Gemini. And locations near 3.4 stars
103. 3:32with [music] sub-5% review response
104. 3:34rates are effectively invisible in
105. 3:37AI-generated local recommendations. Not
106. 3:40ranked lower, they're excluded entirely.
107. 3:43Google's local [music] three-pack will
108. 3:45recommend 35.9%
109. 3:47of locations. ChatGPT will recommend
110. 3:501.2. AI is roughly 30 times more
111. 3:54selective and [music] fewer than half
112. 3:56the brands winning traditional local
113. 3:58search show up in AI recommendations at
114. 4:01all. Or in other words, your rankings
115. 4:04are not transferable. So, if you're
116. 4:06sitting below 4.0 with a bad response
117. 4:09rate, [music] close this video right
118. 4:11now. No agent I build is going to fix
119. 4:13that.
120. 4:15Only better services do. I'd rather tell
121. 4:17you that honestly up front before you
122. 4:19spend the time to watch the rest of this
123. 4:21video. Finding number three, volume
124. 4:23beats perfection. Uberall ran five
125. 4:26studies across five verticals, [music]
126. 4:27five frontier models, and 120,000 plus
127. 4:31mentions. Review volume predicts AI
128. 4:34mentions better than star rating in most
129. 4:36categories. A practice with 1,500
130. 4:38reviews at 4.4 is going to beat one with
131. 4:4220 reviews at 4.7. And here's the single
132. 4:45biggest measured lift in that entire
133. 4:47data set. Restaurants on zero review
134. 4:49platforms beyond Google saw a 51.2%
135. 4:53mention rate. At four platforms, that's
136. 4:5589.6. Now, one more from a subreddit
137. 4:58thread called the local SEO. A pizzeria
138. 5:01>> [music]
139. 5:01>> with 4.7 stars, thousands of ratings,
140. 5:04ranked number two in the Google map
141. 5:06pack, completely ignored by every LLM
142. 5:09for the term best pizza near me. So, the
143. 5:13competitor that's going to win has lower
144. 5:15reviews and lower ratings. The
145. 5:17difference is difference is that the
146. 5:19client's reviews say great [music]
147. 5:21pizza. The competitor's reviews are
148. 5:23paragraphs about the wood-fired crust
149. 5:26and the staff. [music] What does this
150. 5:27mean? You see, a star rating is nothing
151. 5:30except a number. A model is going to
152. 5:32need evidence that it can quote. Detail,
153. 5:34volume, spread, response. Those four
154. 5:37things alone are the job entirely. Now,
155. 5:39>> [music]
156. 5:40>> watch me as I hand them to seven
157. 5:41different GrokBot agents. Now, one quick
158. 5:44note on the tool and then we're going to
159. 5:45build. I [music] picked Rockbot for
160. 5:47three reasons, and none of them are that
161. 5:50it's the smartest model. One, every bot
162. 5:52gets its own cloud [music] computer.
163. 5:54That matters here more than almost any
164. 5:56other use case because this work is all
165. 5:59third-party platforms, Yelp,
166. 6:01Trustpilot's directories, and Reddit. I
167. 6:04do not want one mega agent holding my
168. 6:06logins to every platform I touch. Each
169. 6:09bot logs into its own accounts on its
170. 6:11own machine. Two, agent-to-agent
171. 6:14delegation is on by default. This
172. 6:16workflow has real hand-offs. Something
173. 6:19finds a complaint, something else has to
174. 6:21respond to it, and something else has to
175. 6:23log it. Now, that is a team, not a
176. 6:25prompt. Three distinct routines.
177. 6:28Profound found that the average Reddit
178. 6:30post cited by AI in 2025 was posted
179. 6:33roughly a year earlier.
180. 6:36>> [music]
181. 6:37>> 4% date from 2019 or earlier. Old
182. 6:40threads outlive the platforms that they
183. 6:42describe. [music]
184. 6:43So, this is not a one-time audit. You
185. 6:45see, it has to run forever, which means
186. 6:47it has to run without me. And right now,
187. 6:50it's only $20 a month. Now, hold that
188. 6:52number for now because I'm going to be
189. 6:54comparing it to something else later in
190. 6:56this video. Okay. Seven bots. [music]
191. 6:59I'm going to give you the exact name,
192. 7:01the job, the exact description I wrote,
193. 7:03and the routine. Now, please, steal all
194. 7:06of it. Bot number one,
195. 7:07>> [music]
196. 7:07>> Marlow, chief of staff. The first bot is
197. 7:10the delegator. Everything runs through
198. 7:12him, so I'm not scrolling [music] a
199. 7:13sidebar of 12 agents trying to remember
200. 7:16who does what. The description that I
201. 7:18use, you are Marlow, chief of staff for
202. 7:21an AI visibility operation. Before any
203. 7:23task, check whether another bot owns it
204. 7:26and delegate first. Only do the work
205. 7:27yourself if no one specializes in it.
206. 7:30Bring results back here. The mission is
207. 7:32a single number. How often this brand
208. 7:35gets cited across ChatGPT, Perplexity,
209. 7:38Gemini, and Google AI overviews on a
210. 7:41fixed prompt set. Every bot you manage
211. 7:44exists to move that number. Give it a
212. 7:46human name.
213. 7:50It might sound silly. It is not. You
214. 7:52will interact with a thing called
215. 7:53Marlowe more than you will interact with
216. 7:55a thing called Chief of Staff. And
217. 7:57interaction is what makes these systems
218. 7:58[music] get good. Bot two, Iris, the
219. 8:01diagnostic. This is the cheapest
220. 8:03high-value bot in the system. And most
221. 8:06brands have literally never done what it
222. 8:08does. Iris runs a fixed prompt set and
223. 8:11that consists of 20 to 50 different
224. 8:13queries, the same ones every time, split
225. 8:15into three buckets, [music] and the
226. 8:17split matters. The first prompt is
227. 8:19branded prompts. What do people think
228. 8:21about brand?
229. 8:22>> [music]
230. 8:22>> The next one is unbranded prompts. So,
231. 8:25best insert category in insert city.
232. 8:28[music] And then, comparison prompts,
233. 8:30insert brand versus insert competitor.
234. 8:32That split comes from a production data
235. 8:34paper on arXiv tracking 102 different
236. 8:38brands. And I'll be honest with you, I
237. 8:40don't know if arXiv is how you actually
238. 8:42pronounce that. So, [music] feel free to
239. 8:44call me out on that one. Branded prompt
240. 8:45recognition though is 94.2%
241. 8:48on ChatGPT. Unbranded drops to 22.1%.
242. 8:52You see, models know when you're named.
243. 8:55They almost never volunteer you when
244. 8:57you're not. And by prompt category, and
245. 9:00that category is brutal. Comparison
246. 9:02prompts surface brands at 74.6%
247. 9:05of the time. Problem and solution
248. 9:07prompts, 10.8%. The routine,
249. 9:10every [music] 2 weeks, not daily. AI
250. 9:12responses are non-deterministic.
251. 9:15The same query produces different
252. 9:17results 10 minutes apart.
253. 9:20So, daily readings are nothing more than
254. 9:21noise. Iris is responsible for logging
255. 9:24raw counts and dates. This month, your
256. 9:26brand appeared in 47 of 120 test [music]
257. 9:29queries, up from 31. Now, that is honest
258. 9:32reporting. The quote AI visibility
259. 9:34increased 31.4%
260. 9:36is nothing [music] more than theater,
261. 9:38and any vendor showing you AI search
262. 9:40volume or AI keyword difficulty invented
263. 9:43those men
264. 9:45invented those metrics. That is the
265. 9:46whole diagnostic. 10 minutes of setup,
266. 9:49and most brands have never even run it
267. 9:51once. Bot three, Hollis.
268. 9:54The crawl auditor. This is a boring bot,
269. 9:56but it has the highest ROI hour [music]
270. 9:58in the entire build. Cyrus Shepard's
271. 10:00meta-analysis scored 23 factors across
272. 10:0354 studies. [music] Number one was not
273. 10:06content or links. It was URL
274. 10:08accessibility. If the bot can't fetch
275. 10:10and preview the page,
276. 10:12>> [music]
277. 10:12>> nothing else matters. Hollis is
278. 10:14responsible for checking robots.txt
279. 10:16files, CDN rules,
280. 10:19and WAF [music] for CPT bot, chat, CPT
281. 10:22user, quad bot, Perplexity bot, and
282. 10:25Google [music] extended. Then, it reads
283. 10:27server logs to confirm they're actually
284. 10:29fetching. This is important because a
285. 10:31permissive robots.txt file with a
286. 10:34blocking WAF is the most common silent
287. 10:37failure I found. Now, if you don't know
288. 10:39what WAF stands for, that's okay. I
289. 10:41didn't either.
290. 10:43But, it means web application firewall,
291. 10:45and Hollis is explicitly [music]
292. 10:47instructed not to touch LLM text files.
293. 10:50Ahrefs, the tool that I use, analyzed
294. 10:52137,000
295. 10:54different sites.
296. 10:5697%
297. 10:58of LLM text files received zero requests
298. 11:01in a month. Zero AI bots went looking
299. 11:04for files that didn't exist. Google's
300. 11:06Google's very own documentation says
301. 11:09they don't [music] use them. Shepard
302. 11:10scored it 2.0 out of 10.
303. 11:13Dead last of [music] 23 factors. You can
304. 11:16publish one in 5 minutes if you want,
305. 11:18but never put it on an invoice. Bot
306. 11:20four, the listing accuracy bot. Synci
307. 11:22found [music] business profile accuracy
308. 11:24is 68% on ChatGPT and Perplexity versus
309. 11:29100% on Gemini, [music]
310. 11:31which is grounded directly in Google
311. 11:33Maps. Their read is that this gap
312. 11:36explains [music] most of why Gemini
313. 11:39recommends local businesses nearly 10
314. 11:42times more often. Inconsistency does not
315. 11:44demote you. It simply makes the model
316. 11:46lose confidence. Hours, addresses,
317. 11:49phones, categories. Vera is going to
318. 11:52check every surface weekly [music] and
319. 11:54flag any drift. This alone is the
320. 11:56cheapest category and almost nobody ever
321. 11:59makes it. Bot five, Ansel, the review
322. 12:02engine. This is the one that moves the
323. 12:04gate. Three different jobs. Job one,
324. 12:08response coverage. Response rate is an
325. 12:10explicit signal in the SoCi data set.
326. 12:12Ansel drafts a reply to every review,
327. 12:14especially the bad ones. Then, it drops
328. 12:17it into a queue for me. Most
329. 12:19importantly, it never [music] sends. A
330. 12:21public "We fixed this in Q4. Here's what
331. 12:24changed." gives the model newer evidence
332. 12:27to weigh against an old complaint. And
333. 12:29remember, [music]
334. 12:30cited posts average a year old. Job two,
335. 12:33detail extraction. Ansel writes the
336. 12:35review request prompt, not "Please leave
337. 12:37us a review." It asks the customer to
338. 12:40name what they [music] bought, where,
339. 12:42and what specifically happened. That's
340. 12:44the pizzeria lesson, and it's free. Job
341. 12:46three,
342. 12:47>> [music]
343. 12:47>> platform spread. Ansel tracks which
344. 12:49platforms have coverage and which ones
345. 12:51don't. And it's going to drive toward
346. 12:54forward beyond Google. Yelp [music] at
347. 12:561,000 plus reviews hit a 93.3% mention
348. 12:59rate for restaurants. Grocery at 500
349. 13:02plus hit a whopping 100%. OpenAI
350. 13:05licensed Yelp's [music] 330 million
351. 13:07reviews in July of 2026, so that pipe is
352. 13:11now direct. Bot six, Quill, [music]
353. 13:14the crawlable proof bot. Now, here is
354. 13:16the mechanism that most people miss
355. 13:18entirely. While it might hurt to hear
356. 13:19this, AI does not review your Google
357. 13:22reviews. Star ratings in Maps data load
358. 13:25via JavaScript and are largely invisible
359. 13:27to AI crawlers. Reviews reach the models
360. 13:30when they get quoted or referenced on
361. 13:31crawlable third-party pages. So, Quill
362. 13:34builds one page on your site. Ratings,
363. 13:36dated review excerpts, plain HTML in the
364. 13:39main content, not schema, not
365. 13:42JavaScript. You see, if it isn't
366. 13:43crawlable, it doesn't count. Quill's
367. 13:45second job is comparison and alternative
368. 13:48assets. Go back to that 74.6% versus
369. 13:51[music] 10.8% number. If you have no X
370. 13:54verse Y page and no alternatives to X
371. 13:57page, you are absent from the single
372. 13:59[music] prompt category that actually
373. 14:01brands mention. Quill builds and
374. 14:03maintains those. Format rules Quill
375. 14:05follows. [music] Direct answer in the
376. 14:07first 40 to 60 words under a clear
377. 14:10heading. Comparison tables, real FAQ
378. 14:12schema in the main body. SE ranking
379. 14:15>> [music]
380. 14:15>> found FAQ presence nearly doubles
381. 14:18ChatGPT citation likelihood. And Q&A
382. 14:21format hits a 55% perplexity top three
383. 14:25rate against [music] a 31% baseline.
384. 14:28Question-based titles had roughly seven
385. 14:30X more impact for smaller domains than
386. 14:33for big ones. Now, the last part is the
387. 14:36tail.
388. 14:41Structural lever Structural leverage is
389. 14:44inversely proportional to domain [music]
390. 14:46strength. Big brands get cited
391. 14:47regardless. If you're small, format is
392. 14:50your cheapest lever. And SE ranking
393. 14:53found 85% of perplexity cited URLs had
394. 14:57fewer than 50 backlinks. Bot seven.
395. 15:00Sable,
396. 15:01the sentiment watch. This is the last
397. 15:03bot and the reason the whole system has
398. 15:05to exist forever. Profound from over 4
399. 15:08billion [music] tracked citations.
400. 15:09Positive brand sentiment on Reddit is
401. 15:11cited 5% of the time. Negative sentiment
402. 15:14at 6.1%. Models are not filtering for
403. 15:17constructive feedback. Complaints are
404. 15:19going to surface at a slightly higher
405. 15:21rate than complements from threads
406. 15:24averaging a year old. Sable watches for
407. 15:27new negative third-party threads and
408. 15:29pings me within hours,
409. 15:31>> [music]
410. 15:31>> not weeks.
411. 15:36>> [music]
412. 15:38>> Because complaint you answer on day one
413. 15:41is a different artifact than a complaint
414. 15:44that sits for a year and then gets
415. 15:45quoted back to your prospect by ChatGPT.
416. 15:48And here is the line I hardcoded. Sable
417. 15:51does not post. It simply surfaces and
418. 15:53drafts. I always end up deciding because
419. 15:56manufactured mentions are the fastest
420. 15:58way to lose the channel that still pays
421. 16:01your bills. Google's own guidance says
422. 16:03seeking inauthentic mentions isn't as
423. 16:06helpful as it might seem and that their
424. 16:08spam systems intentionally block it. The
425. 16:11risk is asymmetric. If it works, some
426. 16:13modest lift on some engines. If it
427. 16:15fails, you're in scaled content abuse
428. 16:18territory on the engine driving most of
429. 16:20your revenue. And even if you ignore the
430. 16:22ethics, look at the channel stability.
431. 16:25When Reddit sued Perplexity in October
432. 16:272025, Perplexity's Reddit citation share
433. 16:30dropped 86% almost immediately. Reddit's
434. 16:33overall AI citation share fell 50%
435. 16:36between October and January. A channel
436. 16:39that can lose can lose 86% of its value
437. 16:42in four days because of a lawsuit is a
438. 16:45dependency, not a strategy. Seven bots,
439. 16:48$20 [music] a month. For comparison, the
440. 16:50most upvoted honest pricing breakdown I
441. 16:52found in the subreddit [music] thread AI
442. 16:54Search Lab puts content and technical
443. 16:57AEO at $2,000 to $5,000 a month. Genuine
444. 17:01digital PR at $5,000 to $12,000 and real
445. 17:05monitoring at $2,000 [music] to $4,000.
446. 17:08Now the honest part. Because you guys
447. 17:10have been so generous giving me your
448. 17:12time to watch this video, I'm not going
449. 17:14to spend the last little bit of time in
450. 17:15this video over selling anything. These
451. 17:18agents handle crawlability, accuracy,
452. 17:20response coverage, review detail,
453. 17:23platform spread, crawlable proof,
454. 17:25>> [music]
455. 17:25>> comparison assets, and sentiment
456. 17:27monitoring. That is a real amount of
457. 17:29work, and it's an amount of work that at
458. 17:31places I've worked that specialize in
459. 17:33this very topic charge as much as
460. 17:36$50,000 a month on retainer. But, what
461. 17:39they will not do is lift you over a 4.0.
462. 17:41One genuinely bad quarter of customer
463. 17:44experience
464. 17:45drops you below the threshold that
465. 17:47decides whether you're eligible to be
466. 17:49recommended at all. No amount of
467. 17:51optimization will ever bring you back
468. 17:53over it. So, the way I look at it is you
469. 17:55can either spend $20 a month or as much
470. 17:57as $50,000 a month, and it will not
471. 18:00matter. Stated in the caveats on my own
472. 18:02research, stated plainly, most of this
473. 18:04is local business data. So See and
474. 18:07Uberall study multi-location brands, so
475. 18:09the threshold effect is best evidenced
476. 18:11there. And B2B SaaS almost certainly
477. 18:15behaves differently. Now, everything I
478. 18:17quoted is correlation, not causation. No
479. 18:20engine has confirmed a cutoff. And
480. 18:21nearly every study in this space is
481. 18:24vendor funded, including the ones I
482. 18:27used. The AppSumo scrape is the
483. 18:29exception, and it's a proxy for sales,
484. 18:32not sales data. So, here's what I
485. 18:34learned. Everyone is selling a
486. 18:36measurement of a thing that you
487. 18:38ultimately fix with your own work. The
488. 18:4033 tool category isn't failing because
489. 18:43their founders are bad. It's failing
490. 18:45because the whole space is selling you a
491. 18:47measurement of something [music] you fix
492. 18:50with product quality, customer
493. 18:52experience, and PR. And perhaps what I
494. 18:54find most interesting is that the
495. 18:56AppSumo is that the AppSumo customer
496. 18:58figured this out before venture [music]
497. 19:00capital did. You see, venture capital is
498. 19:02far removed from the day-to-day customer
499. 19:04sentiment and operations. [music] I
500. 19:06think the most elegant way to sum all of
501. 19:08this up is with a typical Horschults
502. 19:10quote. Customers ultimately want three
503. 19:12things: a product or service with zero
504. 19:15defects, they want timeliness, and most
505. 19:18importantly, they want to be treated
506. 19:21nicely. If your business doesn't
507. 19:22accomplish those three things, no amount
508. 19:24of technical strategy is going to fix
509. 19:27this issue for you. What I built is not
510. 19:29smarter than profound. It's just pointed
511. 19:31at the [music] work and not the
512. 19:32dashboard." Or said another way, it's
513. 19:35focused on making the customer happy.
514. 19:37After all, they're paying your salary.
515. 19:39So, that's it right there. Seven bots,
516. 19:41$20, all working while you sleep. If you
517. 19:44want the exact bot descriptions, the
518. 19:46routines, and the 30-prompt [music]
519. 19:48diagnostic set that Iris runs, it's all
520. 19:50linked below in the description. And I
521. 19:51promise you, you don't have to pay for a
522. 19:53school community, and you don't even
523. 19:55have to give me your email address. That
524. 19:57is my thank you for watching this video.
525. 19:59All I ask is maybe giving me a like, or
526. 20:01if you didn't like it and have
527. 20:02constructive feedback, feel free to
528. 20:03leave a comment. I'm always trying to
529. 20:04improve. I just want you to actually run
530. 20:06the diagnostic set because most brands
531. 20:09never do. Now, if there's one thing you
532. 20:10do from this video, make it the
533. 20:1210-minute one. Open ChatGPT, Perplexity,
534. 20:15[music] and Google AI mode right now.
535. 20:16Simply type, "What do people think about
536. 20:18my brand?" and "What do people think
537. 20:20about my category?" That costs nothing,
538. 20:21[music]
539. 20:22and it's the entire category in a
540. 20:23nutshell. One quick story. A company
541. 20:25that I used to work at, which does over
542. 20:27a billion dollars in annual revenue, and
543. 20:29has been around for over 30 years, never
544. 20:31had a warranty policy. Never had a
545. 20:33refund policy. And the executive team
546. 20:35knew that they could get away with it.
547. 20:37And for the longest time, they didn't
548. 20:39really care about customer reviews
549. 20:40because they didn't really put much
550. 20:42emphasis on SEO. But now that AEO is
551. 20:46what shows up first, you better believe
552. 20:48they're paying the price. Now, you can
553. 20:49just imagine what they're going through
554. 20:51now. Never ever forget,
555. 20:54the person paying your salary is
556. 20:56ultimately the customer. And I dare say
557. 20:57it's unethical to discount that. But I
558. 20:59think that's all I got for now, guys.
559. 21:00Subscribe if you want to keep hearing
560. 21:01from me. As a small YouTube channel, I
561. 21:03really appreciate you, and I always
562. 21:05will. I hope you guys have a great rest
563. 21:06of your day. Cheers.