Can you make a video (or do you already have) an AI autoresponder for instant cold email campaigns?
I’m not sure I’ll make a video, but I’ve always been skeptical about setting up an autoresponder that replies instantly. I get it—if a message comes in at 4 a.m., I won’t be around to answer, but during business hours a lead is money on a silver platter, and you should respond yourself right away. Even a 1 % edge over an AI in response quality is the whole business. I haven’t covered this in a video because I don’t want people to waste time on a system that often underperforms and adds complexity. Still, I’ll walk you through a simple setup: grab an instant webhook (e.g., from Make.com), which will give you the lead’s email, name, company, website, phone, and the reply text. You don’t need an extra API call; the webhook includes the email body. You can then feed that data into a language model—most people try Retrieval‑Augmented Generation (RAG), but I’d just use a prompt with a model like GPT‑4, prepend a few thousand tokens describing your business, products, and example replies. The model can generate a response and send it via Gmail. In my opinion, this isn’t a great idea; it’s far more effective to answer the inquiries yourself. A single lead can be worth $5‑10 k, and using AI could drop your close rate by about 10 % if it makes mistakes. That risk isn’t worth it.
Yeah, unfortunately unless you're slotting into a team that does cold email specifically, there's little reason to just work with you to build out and manage the system—you might work as a consultant but not handle the whole management side. It just doesn't work in practice; you've got to play the cold email game. People don't realize the layers involved and don't treat cold email leads like gold, so you can't just sell the system and have them self-manage it. Unfortunately, you do it, train them on it, and over several months they can learn to operate it and then you can pass it off. But if you're doing a guarantee offer—which I do—then best to treat you as an API endpoint: they give you money, you give them the result. To make sure the client doesn't see you stop sending emails, I'd assume they'd worry after seeing 10 uninterested emails. Yes, absolutely. So I use AI to filter those out: when I get a reply, I use an instantly received webhook or smart lead reply webhook, filter it with AI to see if it's interested, and if so, pass it through to their Slack with a formatted message saying, 'Hey, we just generated you a new lead. Here's their information.' I do the same for every booked call; the Slack channel ends up being nothing but wins. Hopefully that makes sense. All right, just a little under the weather again. Full cold email delivery—I've made videos on this before, though I don't know which one. Thanks for the value. Do you currently use only ClickUp to set up CRM? When you build a custom CRM, do you build a sub base as a back end? Yeah, it's almost entirely ClickUp; I don't use Superbase as a back end—it's a custom database, and I don't really care about that. Is Lovable secure enough to use your client solutions, not just as a prototype? No, it's definitely not secure enough to use as a client solution on its own. We see this because a lot of these vibe-coded apps are getting hacked—like the T app, which released hundreds of thousands of women's profile pictures, email addresses, locations, and photos online. This happens because these systems are often vibe-coded, and we're seeing more of them get hacked. So I wouldn't use a vibe-coded app to deploy a live solution with a database, user login, etc. If it's just a simple front end visualizing non-valuable data, that's fine, but if you want to make somebody an app, don't do that.
You know, I actually did this exact same thing using a service called Be Human a while ago. I'm not affiliated with them and don't use their service anymore; I'm not even sure if they're still around, but it was a pretty cool service. You could record a video, say the person's name in it, and the AI would find that point, blur over your mouth, and use a training sample of your voice to say the name. For example, I'd say, 'Hey, really cool seeing you. Thought I'd record a quick video just putting this in front of you to see if you get some value out of it.' The video would then say, 'Hey, Peter.' The issue back then was that the models weren't good at distorting your face, so my hack was to grab the Loom video recording bubble (or equivalent) and say, 'Hey, Peter.' While the video was moving fast, you couldn't see my face, and if someone paused on the name to double-check, they'd see me halfway through a transition, making it unclear. Some of the best sales and lead gen cold email campaigns a couple years ago used this approach and flew under the radar. You could create and record the video, give it total plausible deniability, make it seem super customized, and then pass it through. If you're a hacker nerd wanting to build your own pipeline, you could do this for about 5 cents per 200 videos by renting GPUs on a service like RunPod, running a video name‑change model, or using whatever model Be Human uses, treating it like an API point. So yeah, I'm not going to discourage you.
I think you are looking for automation of the wrong thing. While it's important to automate scalable approaches, and cold email is certainly one such scalable approach, the problem is you're looking to automate things before you've even done them. You should always do something at least once before you automate it — preferably quite a few times — because unless you've done the thing at least once, you will not know what you do not know. My recommendation is to go through this process manually. Do lead gen by cold emailing and cold DMing. Follow a guide from me or somebody else, but do it yourself. It'll automatically keep the budget as low as possible. After you've done it for anywhere between 30 to 90 days, you'll know what goes into cold emailing and cold DMing, and you'll be far more likely to make a successful automation — one that doesn't just automate work that makes you zero, but one that automates work you've proven has meaningfully generated revenue before. It is unsatisfying to hear that answer, but it is the truth.
Either way, my recommendation is to containerize your entire offer into an ongoing service where the client pays you a set amount and you return a booked‑meeting package. This is essentially lead generation, not just handing over an AI automation system. Clients often don’t know how to turn an opportunity into a lead or appreciate the value of leads, so they need you to build the system for them. You can take the cold‑email templates and systems you already use to acquire clients and duplicate them for the client—highly leveraged, scalable, and you can add AI‑based auto‑reply features. Containerizing means bundling all the steps (Apollo search, lead scraping, Google Drive upload, etc.) into a single service where the client only sees the money‑in, money‑out exchange and a multiple on their investment. Many agencies discover this approach organically; it’s worth systematizing.