Yes, the idea of RAG is basically what you said. But you cannot just vectorise code from GitHub and then expect your RAG algorithm to come up with something sensible, it's more complicated than that. If you wanna dive in more detail about the technical side of things let me know - happy to help
Someone suggested a tool called datalayer, which is essentially a RAG in that it sits between you and the coding agent, forcing the agent to "think" through whatever docs you have added to it. Bottom line is that it will have more context for the project when you give it commands, which is a major problem. I haven't tested it yet.
Have you thought about the fact that you’ll need to do everything manually to get good results? I mean, you’ll have to prepare many email templates using your client’s brand guidelines (logos, colors, text, etc.). Of course, you could use some AI to help with that, but if you want it to look professional and achieve the best results, you’ll need to do most of the work by hand. This will become an anchor if you want to scale the process.
Yeah no actually. I send only plaintext emails and unless I get an ecom client I would suggest they also do only plain text emails. No logos or any of that nonsense. And I am not setting this up to just be a glorified freelancer, ideally I would have contractors doing the work for me. I have some lined up, I just need to close the deals. Any of this AI stuff that I am doing is pretty much only for in house use.
I go through out your whole journey and it quite impressive and you are getting also success. But one thing is bothering me about your marketing journey. What marketing you are following?
like is it BHW sales thread or other?
I would like know your marketing strategies because after launching you got 10k in sales where bigger companies are doing also same methods to follow.
and honestly your determination is just pumping my adrenaline and motivating me. Keep it up. Thanks for your journey sharing with uss man
Not sure what you mean G I haven't made a dollar yet from this let alone $10k. My goal is $10-15k/mo, but I obviously am nowhere near that.
Speaking of which, let's have a little update.
So basically, I have leaned into the email marketing. I ran a campaign to 18k leads in Nevada and Utah (I wanted small states first so I could scrape all biz leads), hit them with a personalized "hard sell" campaign, and let it run for a week or two so it hit around 12k of those.
From all that I only had 1 booked call. Kind of disappointing as I expected more, but I think what I need to do is change my approach or my lead gen strategy. I think what I may do is position it in a way that I just want to interview business owners in X industry and if they're free for a call. Nobody likes to get pitched, everyone likes to be interviewed. So I am currently in the process of setting up a new "interview" style campaign with a set of 15k leads that I will begin firing off today once the list is done getting verified.
Second, what I think I would like to do possibly is scrape LinkedIn posts for engagement, then make a list of those people and message them. "Hey I saw your comment on blahblah's post, thought I'd reach out..." That seems to do well from what I've read, but the downsides are that it's harder to scale because 1) there are fewer people who engage with linkedin posts, and 2) pulling emails from linkedin profiles adds an extra step that costs money.
Right now, I can use apify to scrape 50k leads for around $60. If I run those through a verifier, let's say 30% are good. I have a very affordable verifier that allows unlimited verifications for $10/mo, up to 100k/day. So $70 for let's say 20k verified leads, very decent.
Compare that to scraping Linkedin for posts (requires api credits or phantombuster subscription). Limited to like 100 profiles per day or something like htat. Then I need to run that through another thing that gets the emails, not cheap at all. Cheapest I found was Anymailfinder which is $400 for 12k credits, but costs 2 credits per verification. So 6k leads for $400.
6k for $400 vs 20k for $70. Not a hard choice, roughly 20x cheaper for the non personalized ones. Yes results will be worse, but I honestly at this point have so many domains and mailboxes that I don't even know how many I have. I am probably close to being able to send 10k per day.
I honestly really just want to get enough leads in to book my schedule and that of any salespeople I have.
I also looked into other options for setting appointments as well, such as building a predictive dialer and hosting it on AWS. Long story short, cold calling campaigns with VOIP are much harder to set up these days. I looked into SMS marketing (I have a lot of phones/sims from another project), and would be able to send around 100 cold SMS per day. Which honestly would probably be a good use of my time, but I kind of gave up when I couldn't figure out a way to drip them out automatically from a central dashboard on a computer. Like if I have to sit there and hold phones and press buttons with my thumbs I kind of lose interest. Maybe I could just do it anyway, even with 1-2 phones. Really would prefer something that just lets me pop in the sim, download an app, connect to a dashboard, load in some leads, add a spintext message, and let it drip throughout the day. Does that exist anywhere? Idk, maybe I'll check.
That said, my most immediate plan is to just get a buttload of leads and hit them with cold email to ask for an interview. If I can get 4-5 a day I'm happy, but I think the results will be much better. I've also coded an AI SDR that can reel leads in and book appointments from Instnatly directly. It's not perfect but with some minor tweaks should be fine.
Idk if I mnetioend it but I also have a database reactivation campaign (also a hard sell) that I am running with around 10k leads. I'm getting some responses, but nothing like what I would like. Very low response rates which is odd. I think it's the leads tbh, but I could be wrong. Reason I say that is that they are not targeted at all. I literally just chose a state and pulled all the leads from that state. Whereas the first campaign I ran was targeted towards dentists and solar companies.
So the real question becomes, how am I going to be able to scrape all these leads and verify their emails to feed the machine? Well, I am building a scraper + bulk email verifier that will scrape leads from Google Maps, verify the emails via API with a cheap solution, and then import them into an Instantly campaign to be emailed. Still working on it but hope to be finished with it by the end of the weekend.
So bottom line, here is the plan:
Continue warming domains/emails
Finish scraping + verifying + importing solution
Set up SMS dashboard thing for cold SMS marketing
I think that's basically it. However I will say that one more thing I think I need is to have more ready made solutions for customers that we can sell. Right now we are doing it on a custom basis, meaning we do a call with a customer, feed the transcript to a specialized GPT that generates an AI solution plan + proposal, and then send that to the customer. But I think having some kind of suite of 3-4 core services that we offer would be good too. Just to give them a taste before getting on the phone, or in situations where lets' say I meet someone in the real world and want to tell them what we can do. Having 3-4 ready offerings will sound better than "let's do a call so I can feed our transcript to AI and have it tell me what we should do for you."
