[Journey] Ranking AI Sites For Passive Income
- Thread starter spectrejoe
- Start date
100%. Even nofollow links work from hq sites (not from the Reddit though).
Assclown
Regular Member
- Apr 20, 2018
- 407
- 316
How are you going to automate if you dont use some form of prompt template?
I beleive this AI stuff is best done on scale without giving a shit on quality.
- Sep 25, 2013
- 4,698
- 2,694
Not too sure, never heard about it but maybe ur right. Anyways I always use wordpress regardless.
I'm not super sure about nofollow but maybe you're right. I do use reddit a lot for promoting my other businesses and its great for traffic.
I think he means doing things over multiple prompts instead of a single prompt.
And yea you can definitely infinitely scale with terrible quality and just make money while it lasts but I'm not a fan of that type of thing. Plus since goal is selling website eventually I wouldn't be able to sell it unless it was a HQ website with long term profit reports
chrischris15
Registered Member
- May 5, 2014
- 95
- 55
I have multiple prompt templates. The large context window just lets me add maybe put 3 prompts into one request. That gives it more info at once to think about.
- Sep 25, 2013
- 4,698
- 2,694
It's starting to pick up one of the websites it seems
- Sep 25, 2013
- 4,698
- 2,694
I avoid high competitive niches and niches that involve YMYL like Health and Finance.
I am going for whatever random niches I can think of that has a wide enough variety net of potential keywords, I do a keyword sweep (gather like 10k keywords mix of "best xxx", "how to", "can xxx", "why xxx", etc.... And if I find at least 40 keywords worth targetting I go ahead with the website.
I'm not worried about RPM and how much the product prices are in Amazon, I intend to have plenty of websites so even if one website makes me only 10$ a day, that's fine, it'll compound over time.
- Sep 25, 2013
- 4,698
- 2,694
Update #5
Quick Summary:
Sites still fresh, not much going on, indexing is going well, impressions slowly rising. Post count rising quite quickly. Made a new website. Updated my Automation Framework/pipeline.
Made Site #3
This site I don't have big hopes for because it's quite a competitive niche and the keywords are a bit limited, although I did find A LOT of very similar keywords with some potential, I will target those at the end as I think I will have some insane topical authority after targetting literally 100 of those variations lol. I still decided to go with the niche despite not being satisfied with it because why not? Literally costs me 10$ to buy a domain and takes me a day to do keyword research and then everything is automated anyways, worst case scenario I lose 10$ and a day worth of work lol.
Also decided to do an experiment with this site.
Previously I have been targetting extremely low competition keywords and as those ran out I started targetting medium and then higher difficulty keywords for my other sites. That's a pretty common method for most brand new websites and it does make sense so you get some traffic going early on by going for the low hanging fruit.
However I decided to instead try something a bit different and add articles for the top 10 hardest keywords as they'll act as Pillar articles, the reasoning is simple. As I mentioned before my AI add internal linking to older posts which means that:
Older Posts will have significantly more internal links pointed towards them (thus likelier to rank higher) while "new" articles will have no internal links pointed to them. So the idea is:
- First 10 posts are high difficulty high traffic keywords
- All subsequent articles will be lower difficulty which will be giving internal links to the 10 "pillar" articles and hopefully we can get those to rank.
If I forget to properly document / check how well this experiment went, if you remember 3-4 months from now, please do tag me and I'll give it a look in terms of results and post them here compared to the traditional strategy.
Improved AI Automation Framework / Pipeline
I was taking ~1 hour per post previously before I started optimizing things, this is due to articles being 3-4k words long + 3-4 images on each + adding a bunch of delays as fail safes to ensure everything was properly finished before proceeding to the next step. This led to each article taking a long time. (In reality it took 40 minutes but I had to micro manage the next keywords, etc so sometimes I wasn't looking and wasted a lot of time.)
Now I made it so I can insert all my 3 links + the keywords + categories and the AI will do everything for all sites without me micro managing + managed to test and reduce the internal delays while keeping failure rate at 0% (so far). Average Post duration decreased from 50-60 minutes per to ~25-30 minutes per post + 0 micromanaging from my side besides the initial startup which takes 2 minutes. This allows to better focus on other parts like outreach & more keyword research.
Further Keyword Research
Another week or 2 and was gonna run out of keywords for the first 2 sites so went ahead and got some more keywords to use. Might have to go at it again in a few weeks but that's fine. This time focused more on commercial keywords ("best xxx") so I can properly monetize them with amazon. Didn't actually add any amazon links though, I'll wait until they start getting visitors then I'll update them with the amazon links.
Further Link Outreach
Did some more outreaching to get links for my Site #2 which I figured would be "easier". Personalized message for the niche, tried to keep it casual, included the owners name at the start so they know it's not pure spam...
30 Emails sent ~5 failed to be delivered (prob fake / dead mails?) Out of the 25 delivered, 10 have been seen and so far 0 replies
I swear links & outreach has to be the most cancerous part of SEO, I never really did this shit for any of my previous websites and I've always managed to succeed so I don't think i'll need them at all but I had time in my hands so decided to try my hand at it. Think I'll try again some more and then I'll quit if no success.
Some Stats
Site 1:
Posts: 56
Site 2:
Posts: 38
Quick Summary:
Sites still fresh, not much going on, indexing is going well, impressions slowly rising. Post count rising quite quickly. Made a new website. Updated my Automation Framework/pipeline.
Made Site #3
This site I don't have big hopes for because it's quite a competitive niche and the keywords are a bit limited, although I did find A LOT of very similar keywords with some potential, I will target those at the end as I think I will have some insane topical authority after targetting literally 100 of those variations lol. I still decided to go with the niche despite not being satisfied with it because why not? Literally costs me 10$ to buy a domain and takes me a day to do keyword research and then everything is automated anyways, worst case scenario I lose 10$ and a day worth of work lol.
Also decided to do an experiment with this site.
Previously I have been targetting extremely low competition keywords and as those ran out I started targetting medium and then higher difficulty keywords for my other sites. That's a pretty common method for most brand new websites and it does make sense so you get some traffic going early on by going for the low hanging fruit.
However I decided to instead try something a bit different and add articles for the top 10 hardest keywords as they'll act as Pillar articles, the reasoning is simple. As I mentioned before my AI add internal linking to older posts which means that:
Older Posts will have significantly more internal links pointed towards them (thus likelier to rank higher) while "new" articles will have no internal links pointed to them. So the idea is:
- First 10 posts are high difficulty high traffic keywords
- All subsequent articles will be lower difficulty which will be giving internal links to the 10 "pillar" articles and hopefully we can get those to rank.
If I forget to properly document / check how well this experiment went, if you remember 3-4 months from now, please do tag me and I'll give it a look in terms of results and post them here compared to the traditional strategy.
Improved AI Automation Framework / Pipeline
I was taking ~1 hour per post previously before I started optimizing things, this is due to articles being 3-4k words long + 3-4 images on each + adding a bunch of delays as fail safes to ensure everything was properly finished before proceeding to the next step. This led to each article taking a long time. (In reality it took 40 minutes but I had to micro manage the next keywords, etc so sometimes I wasn't looking and wasted a lot of time.)
Now I made it so I can insert all my 3 links + the keywords + categories and the AI will do everything for all sites without me micro managing + managed to test and reduce the internal delays while keeping failure rate at 0% (so far). Average Post duration decreased from 50-60 minutes per to ~25-30 minutes per post + 0 micromanaging from my side besides the initial startup which takes 2 minutes. This allows to better focus on other parts like outreach & more keyword research.
Further Keyword Research
Another week or 2 and was gonna run out of keywords for the first 2 sites so went ahead and got some more keywords to use. Might have to go at it again in a few weeks but that's fine. This time focused more on commercial keywords ("best xxx") so I can properly monetize them with amazon. Didn't actually add any amazon links though, I'll wait until they start getting visitors then I'll update them with the amazon links.
Further Link Outreach
Did some more outreaching to get links for my Site #2 which I figured would be "easier". Personalized message for the niche, tried to keep it casual, included the owners name at the start so they know it's not pure spam...
30 Emails sent ~5 failed to be delivered (prob fake / dead mails?) Out of the 25 delivered, 10 have been seen and so far 0 replies
I swear links & outreach has to be the most cancerous part of SEO, I never really did this shit for any of my previous websites and I've always managed to succeed so I don't think i'll need them at all but I had time in my hands so decided to try my hand at it. Think I'll try again some more and then I'll quit if no success.
Some Stats
Site 1:
Posts: 56
Site 2:
Posts: 38
RunItBackROI
Newbie
- Mar 13, 2026
- 26
- 9
15 years in and you're jumping back into seo just cause you're bored — respect that honestly. curious though, the software niche you settled into where you barely work once a month, was that something you built up through seo originally or a completely different acquisition channel?
angryPixel
Newbie
- Mar 13, 2026
- 20
- 6
curious what the software dev niche looks like when you basically set it and forget it for a month at a time — like is that mostly saas tools or more info products? asking because coming from a startup background i'm wondering if there's some overlap between that low-maintenance model and what you're building with the ai sites
- Sep 25, 2013
- 4,698
- 2,694
I get customers from youtube & SEO yea. The competition for these niches is generally pretty low (at most 3-4 sites) but there's a big player in the niche which usually dominates the top result. I image he's probably making 7k-10k/month easily in this dead niche
- Sep 25, 2013
- 4,698
- 2,694
It's just a subscription based software. I coded it to do something for the users & requires minimal maintenance. Fixing things that get broken when stuff updates & thats it. I can't go into details otherwise people will instantly figure out what it is. It's a pretty well known niche but no one actually targets it for some reason
- Sep 25, 2013
- 4,698
- 2,694
It's that time of the week for the....
Update #6!
So, like the last few updates, nothing too exciting happened, just consistent posts every day and that's about it. They aren't even a month old yet so I don't expect ANYTHING until we're at least 2-3 months in, although hopefully if I manage to land some links it can be sped up a tiny bit.
In this update I'll talk about:
- The issues I'm running into (Outreach & figuring out if worth going for niches with low CPC when the intention is to monetize with ads)
- Site #4
- General updates on the framework
- What I've learned so far
Issues I've run into
Issue #1:
So as mentioned before, outreach has been a huge pain in the ass, I've actually been doing some outreach for my main website that's not relevant to this journey and proposing link exchanges / guest postings. So far only got 1 positive answer out of 50 mails lol. I did manage to increase my open rate but the actual replies is atrocious.
To fix this issue I will be working on:
- Making an automated pipeline to find potential sites to contact using ChatGPT
- Prompt engineering to find generate highly customized titles & mail bodies
- Automatically send them so I can send 30+ a day without lifting a finger (Success rate is so low that I don't give a fuck anymore if it works, at the end of the day it's a numbers game)
I don't understand why people are so reluctant with trading links or trading guest posts when it's literally a win-win condition, there's literally 0 negative side effects unless you're trading with a website that is penalized lol. It just makes no sense in my POV to turn down a win-win condition.
Issue #2:
I know a niche with virtually unlimited amount of extremely low competition keywords with traffic but the issue is that the CPC for them is extremely low (between 0 and 0.5$). I wonder if it's even worth investing my time in chasing that, will the ad revenue even be worth or will I need a million visitors to make 1000$/month on one of these websites lol. I've been trying to find people who tried this type of thing in the past but it's been hard.
Made site #4!
Found another alright niche with a decent mix of transactional/commercial & informational keywords to target. Made the website and got it running this week.
I'm not super confident on this site but at this point I don't think I'm confident on any of the sites, guess i'm just really fucking negative but I never let that stop me lol. I've managed to make money with AI websites in YMYL niches before with extremely high competition, I'm sure I'll be able to make every single one of these websites become profitable eventually. It's a matter of time, consistency & doing things right.
General Updates
Updated the automation framework to require even less management on my end + fixed some edge case issues (ChatGPT sometimes does things different for literally no fucking reason other than because it felt like it) so I always read my articles and try to find these issues so I can patch them.
I also added some fail safes because due to how massive the articles are & how many images I'm generating per article (3-4) ChatGPT sometimes gets bottlenecked or just pauses when generating image, it was uploading without the images before and so I've added some fail safes to just chill & restart the process after a little bit. Quality is non-negotiable even if it takes longer to complete the articles.
What I've learned so far
Your prompts can be MASSIVE. Just for reference my first prompt has a total of 3,362 words 21,074 characters.
My second prompt (which is the main prompt) has 4,306 words 28,812 characters.
Yes, I'm quite literally writing the bible.
I've always seen people saying "keep prompts short, split it into multiple prompts, etc..." while it's true that splitting things into multiple prompts is recommended and I do it too, I mostly do it when the objective of the response is different.
IE: My prompt 1 is to retrieve me an outline of what the article is meant to be like. While Prompt 2 is the guidelines for the article writing to follow.
And you might be wondering... "That's massive, ChatGPT won't even listen to half the guidelines you ask it to if it's so much" the thing is that it DOES, I can literally add an extra line for example "Don't use the word 'funny'" and it will never use that word ever again. It has become insanely good at comprehending & following massive guidelines.
Stats
Site 1:
Posts: 77
Site 2:
Posts: 59
Site 3:
It's taking a while to index the pages so I started asking google to index 3 per day for a few days.
Posts: 31
Site 4:
Posts: 14 (Did 8 on the first day because fuck it)
Update #6!
So, like the last few updates, nothing too exciting happened, just consistent posts every day and that's about it. They aren't even a month old yet so I don't expect ANYTHING until we're at least 2-3 months in, although hopefully if I manage to land some links it can be sped up a tiny bit.
In this update I'll talk about:
- The issues I'm running into (Outreach & figuring out if worth going for niches with low CPC when the intention is to monetize with ads)
- Site #4
- General updates on the framework
- What I've learned so far
Issues I've run into
Issue #1:
So as mentioned before, outreach has been a huge pain in the ass, I've actually been doing some outreach for my main website that's not relevant to this journey and proposing link exchanges / guest postings. So far only got 1 positive answer out of 50 mails lol. I did manage to increase my open rate but the actual replies is atrocious.
To fix this issue I will be working on:
- Making an automated pipeline to find potential sites to contact using ChatGPT
- Prompt engineering to find generate highly customized titles & mail bodies
- Automatically send them so I can send 30+ a day without lifting a finger (Success rate is so low that I don't give a fuck anymore if it works, at the end of the day it's a numbers game)
I don't understand why people are so reluctant with trading links or trading guest posts when it's literally a win-win condition, there's literally 0 negative side effects unless you're trading with a website that is penalized lol. It just makes no sense in my POV to turn down a win-win condition.
Issue #2:
I know a niche with virtually unlimited amount of extremely low competition keywords with traffic but the issue is that the CPC for them is extremely low (between 0 and 0.5$). I wonder if it's even worth investing my time in chasing that, will the ad revenue even be worth or will I need a million visitors to make 1000$/month on one of these websites lol. I've been trying to find people who tried this type of thing in the past but it's been hard.
Made site #4!
Found another alright niche with a decent mix of transactional/commercial & informational keywords to target. Made the website and got it running this week.
I'm not super confident on this site but at this point I don't think I'm confident on any of the sites, guess i'm just really fucking negative but I never let that stop me lol. I've managed to make money with AI websites in YMYL niches before with extremely high competition, I'm sure I'll be able to make every single one of these websites become profitable eventually. It's a matter of time, consistency & doing things right.
General Updates
Updated the automation framework to require even less management on my end + fixed some edge case issues (ChatGPT sometimes does things different for literally no fucking reason other than because it felt like it) so I always read my articles and try to find these issues so I can patch them.
I also added some fail safes because due to how massive the articles are & how many images I'm generating per article (3-4) ChatGPT sometimes gets bottlenecked or just pauses when generating image, it was uploading without the images before and so I've added some fail safes to just chill & restart the process after a little bit. Quality is non-negotiable even if it takes longer to complete the articles.
What I've learned so far
Your prompts can be MASSIVE. Just for reference my first prompt has a total of 3,362 words 21,074 characters.
My second prompt (which is the main prompt) has 4,306 words 28,812 characters.
Yes, I'm quite literally writing the bible.
I've always seen people saying "keep prompts short, split it into multiple prompts, etc..." while it's true that splitting things into multiple prompts is recommended and I do it too, I mostly do it when the objective of the response is different.
IE: My prompt 1 is to retrieve me an outline of what the article is meant to be like. While Prompt 2 is the guidelines for the article writing to follow.
And you might be wondering... "That's massive, ChatGPT won't even listen to half the guidelines you ask it to if it's so much" the thing is that it DOES, I can literally add an extra line for example "Don't use the word 'funny'" and it will never use that word ever again. It has become insanely good at comprehending & following massive guidelines.
Stats
Site 1:
Posts: 77
Site 2:
Posts: 59
Site 3:
It's taking a while to index the pages so I started asking google to index 3 per day for a few days.
Posts: 31
Site 4:
Posts: 14 (Did 8 on the first day because fuck it)
- Sep 25, 2013
- 4,698
- 2,694
It's been 2 weeks since last update, I had family over so didn't have time to do much this past week.
Update #7
Not much happened, growth is still pretty slow on the Google side of things for the most part as I'm sure I've long since hit the crawl budget as they're taking a while to crawl & index pages, so I've also decided to slow down the article posting.
Instead I started & completed my Outreach AI tool and so far I've sent ~500 mails (~120 for each website besides for site #4, and I have also been using it for my business site that's not part of this journey).
Results from the Outreach + Site Stats
Site 1:
Posts: 90
120 Mails sent
2 Replies - 1 Saying he'll have a look and let me know, waiting on that and the 2nd one we actually exchanged links from a DR 27 website, not bad
+1 Link
Site 2:
Can't show more referrers because it'll give away the niche and maybe my site lol.
Posts: 72
120 Mails sent
4 Replies - 2 Negative replies, basically not interested & 2 Positives, one has already linked to one of my articles with DR 24 & the other one is still in the process of doing so.
+1 Links, Potentially +2 Links.
Site 3:
Posts: 56
120 Mails sent
4 Replies - 1 Negative, 2 50/50 (Interested but not that interested, still in talks) and 1 Positive which we're currently in the process of exchanging links.
+1 Potential Links (not yet confirmed).
Site 4:
Posts: 42
Overall not too bad, it's automated for the most part other than the actual replies which I do manually to ensure the process runs well after they reply to the initial AI outreach. The open rate is pretty good but the actual reply rate is atrocious unfortunately.
I have also made it so that every few days it will send a single follow-up (only once, I don't want to pester people) if they have not replied to the initial outreach, I've seen case studies of follow-ups working well so lets see if it makes it work for me. Currently in the process of sending them as they take a while, will have to rework the follow-up pipeline to speed it up for the future.
The only thing worrying me is that a couple of the sites suddenly lost nearly all impressions in Google (Like from 200 impresisons down to 10, etc...) which is weird, I would expect it to be a penalty of some kind but its not like I'm really doing anything wrong so it feels weird that Google would penalize a website with no bad links, informational content, etc.
It's also weird how some sites getting picked up really fast on Bing while others have yet to get more than a single article indexed on Bing. And Google indexed is extremely slow, hopefully with a few more links & time it'll catch up.
For me it's still in the baby stage, I'm gonna let it sit for another month or so and re-evaluate how they look + maybe I'll begin monetizing them.
Update #7
Not much happened, growth is still pretty slow on the Google side of things for the most part as I'm sure I've long since hit the crawl budget as they're taking a while to crawl & index pages, so I've also decided to slow down the article posting.
Instead I started & completed my Outreach AI tool and so far I've sent ~500 mails (~120 for each website besides for site #4, and I have also been using it for my business site that's not part of this journey).
Results from the Outreach + Site Stats
Site 1:
Posts: 90
120 Mails sent
2 Replies - 1 Saying he'll have a look and let me know, waiting on that and the 2nd one we actually exchanged links from a DR 27 website, not bad
+1 Link
Site 2:
Can't show more referrers because it'll give away the niche and maybe my site lol.
Posts: 72
120 Mails sent
4 Replies - 2 Negative replies, basically not interested & 2 Positives, one has already linked to one of my articles with DR 24 & the other one is still in the process of doing so.
+1 Links, Potentially +2 Links.
Site 3:
Posts: 56
120 Mails sent
4 Replies - 1 Negative, 2 50/50 (Interested but not that interested, still in talks) and 1 Positive which we're currently in the process of exchanging links.
+1 Potential Links (not yet confirmed).
Site 4:
Posts: 42
Overall not too bad, it's automated for the most part other than the actual replies which I do manually to ensure the process runs well after they reply to the initial AI outreach. The open rate is pretty good but the actual reply rate is atrocious unfortunately.
I have also made it so that every few days it will send a single follow-up (only once, I don't want to pester people) if they have not replied to the initial outreach, I've seen case studies of follow-ups working well so lets see if it makes it work for me. Currently in the process of sending them as they take a while, will have to rework the follow-up pipeline to speed it up for the future.
The only thing worrying me is that a couple of the sites suddenly lost nearly all impressions in Google (Like from 200 impresisons down to 10, etc...) which is weird, I would expect it to be a penalty of some kind but its not like I'm really doing anything wrong so it feels weird that Google would penalize a website with no bad links, informational content, etc.
It's also weird how some sites getting picked up really fast on Bing while others have yet to get more than a single article indexed on Bing. And Google indexed is extremely slow, hopefully with a few more links & time it'll catch up.
For me it's still in the baby stage, I'm gonna let it sit for another month or so and re-evaluate how they look + maybe I'll begin monetizing them.
- Jun 15, 2021
- 225
- 21
you have to focus on information gain and data infection within your automated workflow to ensure long term idexing and avoid thin content
- Sep 25, 2013
- 4,698
- 2,694
Same hosting but they are in different cloudflare accounts so should have different ips I believe.
- Sep 25, 2013
- 4,698
- 2,694
I tend to keep an eye on the content produced and it looks fine to me so far
bottingisfine
Power Member
- Apr 13, 2022
- 665
- 269
I hate backlinks too