AI Sales Secret: How to Turn Dead Leads Into Revenue (Without Spending More on Ads) 💰

Full video transcript

[Music]

What’s going on everybody, Darby here with Gen AI University, joined today by my friend Marcus Anderson, who has developed an AI sales Android bot that helps uncover hidden revenue opportunities in businesses without spending more money on marketing or ads. Wanted to bring on Marcus for a quick interview and a demonstration about his tool, because a lot of you out there that have your own business right now could probably be making a lot more money from the leads you’ve already paid for if they’re just sitting there and you’re not engaging with them or segmenting those lists. Marcus is a master at this process — he’s even built a tool around it that he’ll be sharing with us today.

Marcus: Thanks so much for joining us here, D. Thanks for having me — it’s good to see you again, my friend.

Darby: Absolutely, and I know when we started learning together several years ago, initially the idea of “found money” — you’ve taken that concept and applied it into an AI and automation process that really taps into the empathy and connection and understanding of the people you’re working with. Before we get into technical details, give us the high level of who you are and your background — what you do at We Market Humans, and how you came across AI and started applying it.

Marcus: It’s interesting — I started out in a local advertising agency, running ads for car dealerships and casinos, then moved into video production, then social media, then the typical online-business detours: drop shipping, whatever. Eventually I found my way to a guy you and I both know, Travis SEO, and Ron Lynch and Jordan Hall, and learned the concept of found money — that’s actually where you and I met years ago. A lot of the concepts I learned there apply directly to what we’re doing now with AI and these conversational sales Androids. Honestly, I disregarded AI early on, even though you were at the forefront of it with Jasper and Jarvis at the time — I had no idea I’d end up marrying those two worlds together.

My whole philosophy, from being deep in the found-money world, is that every business doing lead gen spends money on front-end customer acquisition. They all have sales processes, but a lot of them are only converting 5, 10, 15% of those front-end leads into backend programs or high-ticket offers — leaving 80, 85, 90, 95% of the people they paid for on the table. We found that if you add an extra sales process that doesn’t require a lot of plate-spinning, you can eke out extra sales without spending more money, building more funnels, or managing more staff.

Darby: I think it’s good to look at it from that perspective of getting more out of what you’ve already got. Everyone thinks “more leads,” but leads that aren’t converting right now don’t mean they won’t down the line. I’m curious what you’re looking for to identify these opportunities and make it make sense to add AI into the mix, because I’m sure you’ve seen situations where AI gets thrown in but isn’t directed at the right process or workflow, and delivers lackluster results.

Marcus: A lot of people get caught up in the shiny-tool idea — “look at my new AI tool” — instead of what it actually does, what the outcome is, how you apply it without it being a huge headache. I’m not a techie guy, which is part of why I do what I do — the found-money side has always been digitally primitive, working with email and Google Docs for high-ticket sales for years. So the question became: how can I use AI in a way somebody like me can apply without getting deep into the weeds technically?

What I’ve found is that any business doing lead gen and spending money to bring in customers, that isn’t closing 100% of those people, can work well with this — if they’ve got phone numbers. At its core, we take people’s phone numbers and marry texting with ChatGPT/OpenAI to have conversations that are super empathetic and feel really human. We’re moving people toward whatever the business needs to close more sales — sales calls, in-home appointments, in-office appointments, whatever it is.

People don’t answer their phones anymore. Everybody defaults to texting, even with friends and family — someone calls, you send it to voicemail and text “what’s up” instead of just answering. We’ve seen businesses dial a lead 12-15 times with no response, then send a text and get 20-40% of those people to re-engage immediately. Why not go to them via text if that’s how they actually communicate? We still qualify people, find out what they’re interested in, and then book them with a sales rep.

Darby: Even thinking about my own behavior — I’d rather respond to a few texts than take a call that interrupts my day. You’ve dialed in that AI-to-human interaction so it feels like a real conversation. What roadblocks have you run into, and are there points where the AI flags that a human needs to step in?

Marcus: There are so many variations of how people respond. 20-40% of dead, dialed-to-death leads will respond and start a conversation; some just say “not interested.” One interesting pattern: people have told the AI things like “I’d never work with Steve the owner, but you seem really nice, Sarah” — complimenting the AI without realizing, or not caring, that it’s AI. We’ve trained it for the scenarios that get weird too — sometimes people get inappropriate, even x-rated, with customer-service texting in general, and our AI is trained to immediately end that conversation and put the person on do-not-disturb.

When someone wants to talk to a person, we can send a notification to the company, or do a live transfer with all the conversation notes attached — a “whisper transfer” — so the person picking up the phone already knows what’s been discussed (square footage, specific concerns, whatever it is) instead of the customer having to repeat their whole story. The AI is trained to know its place: it’s not a closer, not an in-home inspection expert. Its job is to qualify and schedule, and if someone pushes for a deeper conversation or asks something it can’t answer, it says it would be best to speak with one of the company’s advisers. It’s good at handling objections and staying focused on the goal of getting an appointment scheduled. And it doesn’t have a bad day, doesn’t get moody, can run 24/7 — some of these businesses were taking 60 hours to respond to leads before we came in, and by then the lead is dead. Speed to lead is dialed in now.

Darby: That’s a great segue — let’s see this in action.

Marcus: At its core it’s a marriage of OpenAI, texting, and a lot of background workflows and automations — Zapier, n8n, that kind of thing. It sounds simple, but there are so many potential scenarios you have to account for, so the back end is pretty complex even though the user experience feels like a normal texting conversation.

I’ve got clients in the walk-in tub space — tubs for people who can’t safely use a regular shower or tub, often elderly or at risk of slipping. This whole approach starts with a “hand-raiser.” There are a few places you can install it: database reactivation (old leads the sales team never got on a call), a “48-hour Android” for leads the sales team hammered with calls for 48-72 hours and then moved on from, and a real-time version that texts someone the moment they submit a form. That last one is powerful because a lot of these leads are shopping three or four companies in the same ten-minute window — hitting them fast, when most competitors won’t respond for hours or days, dramatically improves your odds of closing.

[Live demo of a walk-in tub Android]

This is a demo version, not the fully built-out version with pacing and pauses — so responses come back very fast, faster than a real conversation would feel. In the live build, we add pauses to make it feel human.

The first message usually looks like: “Hey it’s Sarah from Walk-in Warehouse, is this the same Casey that got a walk-in tub quote from us recently?” If this is a database of leads a business has already given up on, getting 20-40% to respond and even 10% to convert is genuinely found money — they’d written off 100% of that list.

In the demo: “Yes it is.” — “Awesome, my calendar just pinged me to call but I didn’t want to disturb you — are you still looking for help?” That message pre-frames a call while flying under the radar of people who don’t want to be called. “Possibly, what’s up.” The Android asks whether the tub is for the person or a family member — a common qualifying question, since a lot of walk-in tub buyers are adult children shopping for an elderly parent. It’s not just dropping a calendar link immediately; it wants to actually engage.

Objection: “Mom took a pretty bad fall, broke her hip, we’re all pretty worried about her.” The Android responds empathetically — sorry to hear about her, a walk-in tub can be a great solution for safety and ease of use, we can help find the right model and offer professional installation — then asks about scheduling with a remodeling specialist. It’s acknowledging the emotional context, not just pitching the product.

Objection: “Maybe I don’t want to deal with a high-pressure salesman.” — “Totally understand that concern, our specialists are here to provide information and help you find the best option without any pressure.” It keeps the focus on the person’s actual concern — mom, not the tub — because the reason someone wants to buy is never really about the product.

Objection: “I’m not even sure she can have a walk-in tub, she lives in a mobile home.” — “Got it, that’s something we can look at together — mobile homes can have unique requirements but we have options that might work.” Another objection, on affordability: “Not sure we can afford it, times are tight.” — “Completely get that, we offer a variety of models and financing options, there’s a good chance we can find something that fits your budget.”

Once the person agrees to schedule, the Android can either book directly to a calendar, negotiating times back and forth (“Wednesday at 3?” — “We’ve got Wednesday at 2 or 5, would either work?”), or transfer the call live to the business with a whisper of everything discussed — the mom, the finances, the mobile home — so the human rep doesn’t make the customer repeat the story. That’s roughly how it works for database reactivation, and a very similar approach applies to 48-hour follow-up or real-time lead capture, with some tweaking depending on the situation.

Darby: I love how simple the user interface feels, even though the back end clearly needs the right instruction and training. What other industries could this apply to — mostly service businesses, or others too?

Marcus: It works for anyone with a list of leads with phone numbers who opted in compliantly — not purchased lists. Home services, debt relief, mortgages, wholesale real estate investing (people opting in because they want to sell their home), class-action legal cases where attorneys are trying to sign people up, plastic surgeons — really anything. I prefer working with businesses that have large lead lists with a lot of flow, since I typically partner with businesses rather than just building it and walking away. There’s no shortage of texting bots out there, but most can’t sustain a real conversation like this because they haven’t built out all the contingencies for what can go wrong — that’s the hard part. It works well for local/service businesses, but also online businesses with front-end opt-ins that have phone numbers and existing calling teams that can’t get people booked.

Something else worth showing — the same walk-in tub Android can detect and switch into up to 50 languages, seamlessly starting to respond in Spanish, for example, if that’s what the person writes in. If ChatGPT/OpenAI supports the language, the Android can use it. No human phone team can realistically do that across more than a couple of languages.

Darby: That’s a good highlight of the contingencies you need to plan for — you could build something like this and just not think to add that. That’s why people hire this out as a service; it takes time to learn how to program these agents, because AI doing something isn’t automatic — it has to be programmed to do it.

Marcus: At the end of the day it’s about more sales — while protecting brand reputation. Someone who doesn’t speak English as a first language, being met in their native language, feels far more comfortable, and that’s not something people are used to experiencing. If there are three or four other walk-in tub companies competing for that lead, and yours is the one that can speak their language, that’s a real edge.

Audience question (Chris): Does it work with any CRM?

Marcus: Yes, via API. We typically build it for a business, they pay a build fee plus a recurring arrangement (per appointment set, per sale, or a combination — it varies by business), and we connect to their CRM via API. Everything runs inside a robust automation layer with CRM management, tags, and workflows, and it plays nicely with n8n, Zapier, and OpenAI. One walk-in tub client uses a specialized home-services CRM and we’re tapped directly into it.

Darby: A lot of our community uses GoHighLevel and builds tools within it too — cool that you’re compatible there as well.

Audience question (Tyrese): Since different languages can be used, is there a way, early in the conversation, to ask if someone is struggling with English and would prefer another language?

Marcus: I hadn’t thought of that specific approach — there’s a sensitivity to it, since asking could come across as an assumption, but framed as “is there a preferred language you’d like to speak in,” it could be a nice way to break the ice, especially since most people don’t realize the system can actually converse in that many languages. Worth testing.

Darby: Especially in areas with a strong non-English-speaking community — a Hmong community, or a large Polish population in a city like Chicago — that consideration matters, and it’s probably case-by-case. The cool thing is you can test anything: try it, see if it lands or falls flat, and adjust.

Marcus: Even the opener as another gateway into the conversation is interesting to test. Normally when someone says “no hablo inglés,” that reads as the end of a conversation. But if the Android responds in Spanish immediately, the whole dynamic shifts.

Darby: Thanks Tyrese. Any other questions, drop them in the chat. Thank you again, Marcus, for the demo — this is a solid example of how almost any business doing marketing and holding customer data likely has leads they could reactivate or follow up with through a different medium like texting, layered on top of what they’re already doing. And Chris commented that he likes how simple it looks on the front end, even with more complexity on the back end in terms of instruction.

Marcus: You don’t have to be some AI-obsessed technical genius to use this — some of you are into that, and some of you are just business owners who recognize AI is coming for every business in some fashion, so the question is how to treat it as a tool or an employee and leverage it, rather than worry about it. This isn’t a pitch, but most of the time we work on a partnership basis — you don’t pay unless we make you sales. That’s been the found-money philosophy from the start: no spending extra money, just tapping people you’d already written off. Once you realize this works for one segment of leads, you start seeing it apply to 48-hour leads, real-time opt-ins, even post-service follow-up — the Android can also go back after a completed service and encourage customers to leave a Google review, and it’s outperforming people making manual calls to ask for the same thing.

It’s a really interesting time — not necessarily the best time to be a human “setter,” but we’re seeing a redistribution of tasks. I’m not a super techy person, but I like the simplicity of it: at the end of the day it comes down to having a conversation with leads you couldn’t previously reach, without needing a paycheck, running on credits and automation costs instead.

Darby: For setters or people in that role today, this seems like a tool to add to the belt and evolve with, rather than something to fear. Any thoughts on repositioning?

Marcus: Honestly I don’t know how long these sales Androids will exist in this exact form before something like a single ChatGPT operation or GoHighLevel just does the whole thing. But it’s still ground-floor, pioneer-days stuff — most businesses aren’t using any real level of AI yet. I took a skill set I already had — selling high-ticket offers through email and a Google Doc — and, as the economy made that harder, married it with AI and chat sales Androids to build a new offer. Every demo I’ve done, the business wants to move forward, because it solves a real problem and every business needs it. Find where the need is, apply the skill set you already have, and put AI to work on it.

Darby: We were just talking before this call about ChatGPT’s impact on therapy, and it’s a similar idea — what are you going to do to pivot and apply what you’re doing now to what’s coming next. For anyone who wants to reach out and talk about partnering or running a campaign like this, how can they find you?

Marcus: My email goes straight to me — Marcus at We Market Humans dot com. Reach out and tell me a bit about your business and industry; I’m not a high-pressure sales person — I live on Kauai, where high pressure isn’t really in the vocabulary. I can put together a demo for your specific business, showing what it would actually look like — solar, mortgage, debt relief, whatever it is. Once we get into building it together it takes a bit more time and gets more robust, but I’m happy to chat.

Darby: Thank you again, Marcus, for coming on and sharing your process — it’s been great to see how you’ve leaned into found money over the past few years and gotten strong sales results for the clients running these campaigns. If you have a business you think could plug into Marcus’s system, reach out — his contact info is right there. Appreciate you and your time, everyone here live and catching the replay. Until next time, keep coming — everybody, cheers, y’all, thanks.