Monetizing AI: Workflows, Agents & Business Models [Megathread]

[something_else]

1 hour ago, LordFall said:

I think it's possible to do half decently already and in 6 months it'll be quite good

It really isn't, and the improvements between iterations aren't nearly as significant as they used to be. In fact, some recent iterations have actually gotten worse at writing naturally. For a while, Claude models were practically incapable of speaking in natural English.

Yes, these models can write copy really well with some oversight. I've built landing pages using Codex/ChatGPT before, and while they can produce a great-looking page, their copy choices are often abysmal and they need to be guided through it.

And that's an entirely different ball game from convincing people they're talking to a human when they're actually interacting with an AI agent. Doing that at scale, through custom vibe-coded software, and then trusting your entire business model on that working reliably, is a monumental task with severe risk.

[LordFall]

34 minutes ago, something_else said:

There are tons of solutions out there already. The CRM space is crowded. I guarantee something already exists that can do exactly what you want to do, or is extensible enough to do it with some minor customisations.

Totally I agree with you. Part of my issue with the business I was running was cognitive overload. I had to have the skillset to take good dating pictures, package them into a worthwhile offer both for myself and the customers, shoot the ads to sell the service, run the ads, answer incoming DMs to book calls, take the sales calls, book my calendar, do the photoshoots, edit the pictures, handle customer support, etc. 

I tried to bring on business partners a few times and that worked in some ways but introduced some new issues. One guy was mostly reliable but he wanted 50% of the business to dedicate himself to it and I thought he brought nowhere near the value I did. I was willing to just split the profit 50% with him until we worked successfully together for at least 6 months-1 year before signing anything but that didn't work for him.

Building a successful global enterprise has been a goal of mine for years and I'm always learning new skills to make that goal possible but it's obviously not quite there yet.

Another massive use case for AI for me is a personal assistant that helps me organize all my ideas and plan my workday depending on my priorities and energy levels. I use Gemini for that a bit but it's not as advanced as I would like it. 

34 minutes ago, something_else said:

However, trusting your entire client-base to vibe-coded software is... unwise. And letting AI agents talk to your clients with zero approval required is also a huge risk. People can usually still identify AI-speak, and it's especially frustrating if someone tries to pawn off an AI chatbot as a real human. This is terrible for reputation, especially if you're trying to pitch yourself as a successful entrepreneur who makes a lot of money. The natural question is "well why can't he afford a real human".

I don't think it's really all that complicated. You assign it a few lead conversations and see how it does. This stuff is based off of closing percentage so for example myself I would spend about $6 for leads while running meta ads. I would convert around 20% to sales calls so each call could cost me $30. I would close around 20% of calls so customer acquisition cost $150. That number went up closer to $200-250 when I brought on another guy and at that time I sold $500 photoshoots so the margins are really important. 

An AI appointment setter would allow me to analyze all the conversations and try different formats to see how the % change. These are not in depth conversations. Meta Ads lets you set pre-selected prompts for people to send as the first message, so they ask things like: "How much are the photoshoots?" / "How does the photoshoot work?" / "Can you tell me more about this?" then we do have you ever done a photoshoot before? -> There are a few different packages depending on what you need -> When are you available for a phone call? 

I think in the next 1-2 years the voice agents will good enough to handle the sales calls themselves. I've had some call me and try to sell me and it's definitely not good enough to build a sales team around but very soon it will be. The amount of business models that this unlocks is insane. 

@Lyubov This should already be able to handle my personal needs for setting up sales calls en masse. I will play around with it over the next 1-2 months update you guys on how it went. 

Edited 13 hours ago by LordFall

[something_else]

41 minutes ago, LordFall said:

Building a successful global enterprise has been a goal of mine for years and I'm always learning new skills to make that goal possible but it's obviously not quite there yet.

Out of curiosity, what does your business actually do? Just photoshoots?

Quote

Totally I agree with you. Part of my issue with the business I was running was cognitive overload. I had to have the skillset to take good dating pictures, package them into a worthwhile offer both for myself and the customers, shoot the ads to sell the service, run the ads, answer incoming DMs to book calls, take the sales calls, book my calendar, do the photoshoots, edit the pictures, handle customer support, etc. 

AI can absolutely help with this, just not in the way you're imagining. Get a ChatGPT Plus subscription and use Work mode to help set up CRM automations, manage your administrative tasks and analyse your data. You can connect it to external services through plugins or MCPs. I'd explore that before building custom software.

41 minutes ago, LordFall said:

An AI appointment setter would allow me to analyze all the conversations and try different formats to see how the % change. These are not in depth conversations. Meta Ads lets you set pre-selected prompts for people to send as the first message, so they ask things like: "How much are the photoshoots?" / "How does the photoshoot work?" / "Can you tell me more about this?" then we do have you ever done a photoshoot before? -> There are a few different packages depending on what you need -> When are you available for a phone call? 

There are already off-the-shelf services that do this. My issue isn't with AI handling basic enquiries. It's with passing off an AI chatbot as a human. That's deceptive, difficult to do, and you're going to piss people off.

Quote

I would spend about $6 for leads while running meta ads. I would convert around 20% to sales calls so each call could cost me $30. I would close around 20% of calls so customer acquisition cost $150. That number went up closer to $200-250 when I brought on another guy and at that time I sold $500 photoshoots so the margins are really important. 

I think there's an issue with obsessively optimising metrics like CTR and conversion rates while neglecting the human element. Better numbers don't necessarily mean better customers or a better brand image. You risk losing sight of who you're actually marketing to and how your business is perceived. That's especially important if you're planning to scale significantly. If you provide top notch human service, the chance of a customer sticking around in the long run and using you in the future goes up significantly, for example. The stuff your business does that makes customers go "wow, they didn't need to do that but they did" is really important for building repeat customers instead of a CTR funnel optimised pump-n-dump.

On a tangential note, I've also run Meta ads, and my experience with its AI enhancements has been pretty poor. Meta pushes you to use them claiming they'll increase CTR. Instead, most noticeably reduced my CTR, and if CTR did improve, it sometimes came with a shift towards older demographics, attracting more people over 50 and fewer under 30. What this tells me is that AI enhancements are both not good enough to run autonomously yet, and shift the targeted demographic significantly further up the age brackets.

41 minutes ago, LordFall said:

I think in the next 1-2 years the voice agents will good enough to handle the sales calls themselves. I've had some call me and try to sell me and it's definitely not good enough to build a sales team around but very soon it will be. The amount of business models that this unlocks is insane. 

I think you're assuming progress will continue linearly. Going from sounding 50% human to 70% is considerably easier than going from 70% to 90%. That final stretch towards being genuinely indistinguishable from a human is much harder, particularly when you need the system to perform reliably across thousands of real conversations.

And again, you still seem ambiguous about whether you intend to tell customers they're speaking to an AI. That's important point to establish.

Edited 12 hours ago by something_else

[LordFall]

49 minutes ago, something_else said:

Out of curiosity, what does your business actually do? Just photoshoots?

I've changed my exact services a few times. My main skillset is in online lead acquisition and content production so I've had b2b clients where I managed their social media profiles for a retainer and last year I was more like a personal brand consultant. What I found the most interest in and I've done the longest is be a dating photographer so I think I'm gonna triple down on that.

So that's dating profile photoshoots, dating consulting, help throwing events, help shooting content to build a nice looking instagram ,etc. I think if I focus on that vertical and really nail it I can scale it to 7-8 figures over the next few years. 

My personal branding client paid me $5,000/month so if I'm looking to productize the services I gave him into something like a $1,000/month offer that solves the dating life of single male professionals 28-40 and perhaps even women since the women I consulted on their dating lives seemed to benefit from my advice. 

49 minutes ago, something_else said:

On a tangential note, I've also run Meta ads, and my experience with its AI enhancements has been pretty poor. Meta pushes you to use them claiming they'll increase CTR. Instead, most noticeably reduced my CTR, and if CTR did improve, it sometimes came with a shift towards older demographics, attracting more people over 50 and fewer under 30. What this tells me is that AI enhancements are both not good enough to run autonomously yet, and shift the targeted demographic significantly further up the age brackets.

Yeah I never touch any of meta's optimizations. I don't know what their aim is with it but they always reduce my performance. I always calibrate every setting myself. 

> And again, you still seem ambiguous about whether you intend to tell customers they're speaking to an AI. That's important point to establish.

No I think that part is mostly irrelevant. When they message my brand they talk to the Royals Dating instagram account. It hasn't made much of a difference when I've had a random guy use my templates when booking appointments for me and even other people doing sales calls is standard.

What matters is how happy they are with the results they're getting from purchasing our services. My role as the founder and face of the company is to guarantee a standard of quality across every offer that goes out under my brand. 

Edited 12 hours ago by LordFall

[something_else]

1 hour ago, LordFall said:

I've changed my exact services a few times. My main skillset is in online lead acquisition and content production so I've had b2b clients where I managed their social media profiles for a retainer and last year I was more like a personal brand consultant. What I found the most interest in and I've done the longest is be a dating photographer so I think I'm gonna triple down on that.

So that's dating profile photoshoots, dating consulting, help throwing events, help shooting content to build a nice looking instagram ,etc. I think if I focus on that vertical and really nail it I can scale it to 7-8 figures over the next few years. 

My personal branding client paid me $5,000/month so if I'm looking to productize the services I gave him into something like a $1,000/month offer that solves the dating life of single male professionals 28-40 and perhaps even women since the women I consulted on their dating lives seemed to benefit from my advice. 

Brand consultant is cool and a clearly marketable skill, I can see you being good at that. If it made you so much money, why not double down on that?

I think the dating app / instagram profile pic overall as a one off product is solid, which is I assume what you offer now.

My main question about your model as someone close to your demographic would be... what exactly am I paying $1000/m for? Instagram and dating app photos? That's a one off purchase. Style overhaul? Also a one off purchase, really. Dating advice? I mean sure, that can be a service, but I'm going to need to see lots of evidence that your advice is better than what ChatGPT gives me for free before I'm parting with that kind of money. Nothing you have said here convinces me that this $1000/m subscription is a good use of my money.

Quote

I focus on that vertical and really nail it I can scale it to 7-8 figures over the next few years. 

I mean that's insanely optimistic assuming you're starting from <6 figures right now. But I wish you luck 

Quote

No I think that part is mostly irrelevant. When they message my brand they talk to the Royals Dating instagram account. It hasn't made much of a difference when I've had a random guy use my templates when booking appointments for me and even other people doing sales calls is standard.

What matters is how happy they are with the results they're getting from purchasing our services. My role as the founder and face of the company is to guarantee a standard of quality across every offer that goes out under my brand. 

I think you missed my point, it's not about it being you that they message. It's about whether they're messaging an AI or a human and how deceptive you are being in that regard. If someone has a suspicion that they're chatting to an AI chatbot but the brand doesn't make it clear, a potential customer often interprets that as an attempt at deception, which is one of the most damaging things a customer can feel your brand is doing. So it is very relevant.

If AI chatbots are really as good and you think it shouldn't matter, so just make it clear to the customer that they're chatting to an AI upfront and offer them an escape route to talk directly to you if they want it.

And again, to return to the original point, there are tons of services that offer this already. It's not something you need to custom build.

Edited 10 hours ago by something_else

[MarkKol]

Everyone in the agency community knows AI should be used in your business; it shouldn't be your service offering.

I've seen AI offers like,
A. 'We can cut down the hours you spend on this task by 80%.' or
B. 'We can implement an AI agent in your business so it handles and books appointments outside of business hours when there is nobody available to answer, instead of just going to voicemail'

And then you have an offer that is completely unrelated to AI, existed before AI, that is still going to be a better offer and an easier sell, like:
1. 'I can get you 10 high-ticket clients a month using Facebook ads or cold mass email/SMS sendouts; you pay me a portion of what I get you, at no upfront cost to you'
2. 'I can take your database of old clients that haven't booked anything in a while, reactivate them via email, and capture your future website traffic with an email capture pop-up; we will increase your store revenue by 40%, or you pay nothing'

TLDR AI offers suck; I haven't seen one that matches offers that already did well before it

And your offer is like 90% of the game, that and volume of outreach/ads, so it does matter, a lot.

1. My uses of AI have been generating a website fast and without money in minutes

2. Designing Image-based email campaigns; it's a big help.

It's more like a way to speed up your workflow instead of creating entirely new business solutions

It's very useful overall. I haven't really posted everything I know. AI is currently getting a lot of hype, and where hype goes, your skepticism should follow

Edited 9 hours ago by MarkKol

[Joseph Maynor]

Everything is going in the direction of managing AI Workflows.  

[Lyubov]

5 hours ago, something_else said:

However, trusting your entire client-base to vibe-coded software is... unwise.

I told him this twice already and he seems completely sold on the hype... Really @LordFall tell us how it goes for you. I'm not discouraging you but I would take pause to contemplate how you will feel when you potentially had one of those $5000 client but your AI bot went off and you scared them away. I know that no one serious is actually using AI to do sales for them or close deals or "feel things out." Technology has done little to changed the fundamental of sales. No one is using AI for their sales except selling you the idea that using their AI method or startup app will earn you money. We have marginally better chat bots for support after they have gotten your money. Keywords AFTER they have gotten your money.

Edited 8 hours ago by Lyubov

[Lyubov]

5 hours ago, LordFall said:

 

@Lyubov This should already be able to handle my personal needs for setting up sales calls en masse. I will play around with it over the next 1-2 months update you guys on how it went. 

You can test it but I can tell you now I do not trust that AI influencer guy. His hustle is making YouTube videos and selling you (the masses) these idea that this stuff works. This is not something that anyone is seriously implementing. Not because they don't know but because it isn't reliable. We can split hairs, of course you may have clients fill out an intake form from a bot (no different than that of a web form) or whatever that's transparent that they aren't talking to a person, you have may have AI ping out a message to see if they reply and qualify themselves for the service (not too different from scheduled emails and messages we have had for years), but having an AI agent step in to replace a salesman (human job) and schedule them up, answer questions, converse and collect payment? nah it's not really a thing right now except an idea amongst AI influencers.

Edited 7 hours ago by Lyubov

[yetineti]

@Lyubov

I’ve talked to voice bots and email bots from multi million and billion dollar companies, including Walmart and was not entirely disappointed.

I feel like they’re pretty serious.

I know someone who works for a multi billion dollar company and runs the customer service department and actively implements these things.

The owner of my company built his own voice bot for the company calls.

Perhaps your skepticism is being too broadly applied.

Sales is hard, yes. Many companies are not even considering it at the moment. But scheduling, talking, problem solving, follow-ups, leads and acquisition are just becoming increasingly accessible.

Edited 6 hours ago by yetineti

[yetineti]

@LordFall

It sounds like you’ve already done a lot of the AI work, and the rest is really about building the business around it.

I think you’re clearly a good operator, but being a good operator and a good businessman aren’t necessarily the same thing. Sometimes they need each other.

Right now you’re the photographer, scheduler, salesman, customer service, etc. If you got hit by a bus, most of the business goes with you.

Why not hire a part-time assistant at around CA$19–22/hour to start taking some of that over? Maybe consider 2–5% equity later with aggressive vesting protections if you find someone genuinely capable of helping manage the business. But I definitely wouldn’t give away 50% for a role you haven’t clearly defined.

It seems like you already have a decent SOP in your head and through your sales templates. It just needs to be something another person can follow.

Start with 3–5 lead qualifiers, a basic CRM, standardized responses, and a process for tracking and retaining customer information. Let the assistant handle initial emails, DMs, scheduling, and follow-ups while you keep sales and photography.

Every time they encounter something unfamiliar, you establish a rule: “For scenario X, respond with Y, unless Z.” Now that’s documented, and they don’t need to ask you again.

Over time you’re building a system of repeatable decisions while collecting actual data on response times, conversions, lead quality, and what works.

Then you start introducing AI, but one responsibility at a time.

Take the most common, lowest-risk scenario, like someone asking about availability. Have AI identify the inquiry and draft the response, with your assistant approving it. Once it’s consistently reliable, let it handle that specific interaction independently and escalate anything unusual.

Then move to the next scenario. Eventually you can expand from email and chat into scheduling, follow-ups, and voice calls using the same approach.

The important thing is you’re not expecting an AI agent to suddenly replace customer service or close sales. You’re removing one validated responsibility at a time, with a human still overseeing the process.

It’s almost like building a safety net. Every scenario either has a tested procedure or gets caught by a person who can handle it. As the procedures improve, fewer things need your attention.

Eventually your assistant isn’t doing all the repetitive work anymore. They’re supervising the workflows, handling exceptions, and helping identify what to automate next.

And even if AI never gets good enough to handle everything, you’ve still built a business that can operate without you personally managing every interaction.

I think that’s the bigger opportunity here. You’re already good at doing the work. Now it’s about building something that can reliably do the work without depending entirely on you.

AI seems like the obvious solution to that, but I think there’s an order of operations you’re potentially skipping. The company my friend works at already has a consistent business model generating revenue without AI. That gives them established processes, real customer interactions, and measurable outcomes to build automation around.

You can consolidate the common denominators across emails, chats, and customer interactions, automate those first, and eventually build voice capabilities on top of the same underlying system.

Something like:

Proven business model > Document workflows > Delegate responsibilities > Collect data > Standardize decisions > Automate email/chat > Validate with human oversight > Expand into voice > Scale automation

Otherwise, you’re trying to build the automation before you’ve fully defined the business it’s supposed to operate.

 

[Lyubov]

1 hour ago, yetineti said:

@Lyubov

I’ve talked to voice bots and email bots from multi million and billion dollar companies, including Walmart and was not entirely disappointed.

I feel like they’re pretty serious.

I know someone who works for a multi billion dollar company and runs the customer service department and actively implements these things.

The owner of my company built his own voice bot for the company calls.

Perhaps your skepticism is being too broadly applied.

Sales is hard, yes. Many companies are not even considering it at the moment. But scheduling, talking, problem solving, follow-ups, leads and acquisition are just becoming increasingly accessible.

I'm applying my skepticism to sales. I know that at the top of the sales funnel there is some occasional application of AI overseen by a person. I know that having an AI bot send out a message to qualify a warm lead makes sense especially if the lead asked to be contacted, maybe you could push that to having it scheduling an intake call. That is very different from selling an expensive product, answering questions, taking payment and scheduling. Very dynamic stuff. A human has to step in after 1-3 messages. 

I previously mentioned how customer service has implemented it more widely because there is less risk after they get your money and to have you pass some information to an AI labeled chatbot to get some customer support. The dynamic changes entirely when you're contacting a company for support. We have had support chat and automated call lines for ages now and they have gotten somewhat better with LLMs but not enough to justify the hype. 

Edited 5 hours ago by Lyubov

[yetineti]

@Lyubov

I think a good 50% of traditional sales could disappear in the next 5–10 years. People are focusing on how bad AI is at sales right now, but I think they’re missing how much the buying process itself will change.

Gartner’s 2026 research found that 67% of B2B buyers prefer a sales-rep-free experience, while 69% still want human validation of AI-generated insights.

Eventually, consumers will have personal AI agents that understand their needs, finances, and preferences, collate data across providers, compare actual value, and negotiate directly with seller-side agents.

Businesses will have their own agents doing the same thing, with encrypted, privacy-preserving discussions and analytics that both parties can review before finalizing. For B2B, this could mean sharing relevant data securely to calculate mutually beneficial agreements without exposing sensitive information.

The other half is that businesses will increasingly have to cater to these buying agents. Instead of spending so much on persuasion, they’ll need to compete on actual value, transparent pricing, better products, and how well they meet each customer’s specific needs. Their own AI will continuously analyze demand, customer preferences, and competitors to adjust offerings accordingly.

So you end up with a full feedback loop: buyer agents identifying needs and comparing offers, seller agents tailoring products and terms, both sides negotiating, humans approving, and the resulting data improving future transactions.

I think this automates huge portions of retail, insurance, healthcare, and B2B sales. Human sales shifts toward consulting, complex negotiations, finalization, and purchases where people actually enjoy the experience, like cars and luxury goods.

The real disruption isn’t AI becoming better at persuading people. It’s a system where persuasion becomes much less necessary in the first place.

A good example is roofing. Right now, you have someone coming into your house, sitting in your living room with a pamphlet, and spending an hour trying to sell you a roof.

Why would any of that be necessary when your AI could gather quotes, compare materials, warranties, contractor reputations, and pricing, negotiate with their agents, and give you the best options to approve?

You’d still have someone inspect and install the roof, but the whole sales pitch in your living room becomes pretty unnecessary.

 

Edited 2 hours ago by yetineti

[yetineti]

It’s going to be a pain but the economy will eventually turn into hieroglyphics.

The responsibility on the consumer to track market analytics and price statistics will only increase.

Prices are going to go up and down so fast you’ll have to use AI to purchase things or you’ll be robbing yourself.

We already see the shifts over the past decade or so:

- Utility prices can vary by time of day

- Gas prices can vary by $1-2 depending on location or membership

- Movie prices are based on time since release or what platform you purchase it through

- Medical prices vary by provider

- Food prices vary independent of convenience fees, depending on if it is delivery, pickup or in-store

- Airline tickets vary by demand, timing, placement

- Concert tickets have dynamic pricing

- Living complexes do not have standardized fees for admin., garbage services, parking, etc.

- Fast food prices vary drastically depending on where the order is placed

Ultimately, AI could push companies toward consolidation through either efficiency or trust. The real question is whether consumers benefit long-term or just become dependent on systems they can’t verify.

 

 

 

 

Edited 1 hour ago by yetineti