Stop Building SaaS. Build an AI-Enabled Agency Instead (Step by Step Guide)

If you're still trying to vi code an app that can make millions, the chances are you're building the wrong thing. Because there's a new way to build with AI tools that for a lot of people is going to give them a 10 times higher chance of success. And I'm not talking about automation or trying to help businesses with AI. I'm talking about something that you can actually build to help you in your work so that you can start or grow your own AI business.

And in this video, I want to break down exactly what apps are a trap to build, what to build instead, and how to start building it today, so that you know exactly how you can go from where you are now to building the right thing with AI that's going to help you to launch or grow your business. If you don't know me, my name is Chris and for the last 15 years, I've been designing apps and advising startups on product and design. And if you are building with AI, I've got a community helping people do that where you can build alongside 250 other people building apps via coding and launching them to real customers.

So if you're interested in that, click the link in the description down below. Otherwise, let's get straight into the video and show you exactly what you should be building with AI. So if you're vioding an app with AI, the wrong thing to build is a featurebased SAS. And this is the thing that a lot of people start to build when you start vibe coding and you start to experiment with these tools, lovable, claude code, a bunch of other tools.

And it's kind of like creating a software as a service business where it's based on a bunch of features that do a bunch of stuff on a dashboard where you can kind of click and create things and do schedules and visualize a bunch of data. And there's nothing wrong with starting out there because that's the place that a lot of people start when they start building applications with AI. But the AI space is moving at a incredible speed. And the amount that you can do using different agents and using different AI coding tools is accelerating much faster than traditional software.

And if you're building an application that is just built on a set of features instead of actually providing someone with an outcome or a value that they're going to get from using that app, then that is ultimately destined to fail. And I want to show you something else that you can be building instead of building a feature-based SAS platform. Even if you build a SAS platform that is focused on getting someone an outcome or a specific value, that is going to be better than building this kind of SAS idea which is like a traditional software as a service which is based on a per seat model where you pay a subscription per user who has access to the platform.

We have moved beyond that in the AI world with agents and with all the capabilities that they can do into a different place. And one of the best business models that you can build instead of SAS is this AI service business model. So the things with traditional SAS startups is that they take a long time to reach first revenue. They have a higher startup cost.

You need a higher number of users based on the subscription price to actually make money. And when AI gets better and agents get better, this threatens your product instead of actually helping it. And so there are two ways to combat this. If you're building anything with AI, the first is to build an agent native piece of software or an agent native application.

And this just means you're building an application where an agent that someone's using like Claude Code or Claude or Codeex or even Co-work in the Claude desktop app. someone uses co-work using that agent in order to access the data and be able to do the things that your application can do. So the first thing to do is to think about how you can build agent native applications instead of just building traditional SAS platforms that are on a per se basis. The second thing that you can do here is actually to build an AI service business and a service business that is enabled to AI tools like AI applications, AI agents, AI workflows, and custom AI skills that you can build to perform a bunch of different tasks.

And the benefit of doing this AI service business approach is that you'll basically have a much lower time to first revenue. you have a much smaller startup cost because chances are you might be doing this already if you're a freelancer or you run your own agency. You might be doing this already inside of a company and you can use this to get a promotion or get a new job that pays more. You need less clients to actually make more money with it and it has a low level of technical expertise required.

And when AI gets better, your service, your workflows, they also get better as well. So all of this means that if you're building an AI enabled service business instead of a feature-based SAS, you especially if you already have these skills and you already have a set of clients, this is going to make it much much easier to be able to actually make more money and even start making money and launch your own business with this model. And in this video, I want to show you exactly how this model works, how to actually start building it, and I'm going to give you a prompt that you can go and use to kind of find some ideas around your own skills and expertise that will enable you to leverage your own skills, build some AI enabled services, and then start going to clients and actually selling that as a service.

And the thing to know about feature-led SAS is this. Investors are actively avoiding these thin workflows, software as a service platforms, generic horizontal tools that just do a specific set of actions and basically anything that an agent can now do. And agents can go and perform multiple steps, take actions for you, create content, all of these sorts of things. So anything that an agent can do, if you're trying to wrap that into a SAS platform into a software, unless you're making it much much easier from a product point of view for the user to actually do, then this is going to have less and less value over time.

Like Salesforce and Adobe are actually losing market share at the moment. These outcomebased applications are going to take the bottom of the market. So think about outcomebased applications as saying like I have created an app that's going to help you double the amount of followers that you get on X for example. So we want to create an outcome that people can specifically get when they use our application rather than just saying like here's a tool to help you manage all of your content across your social platforms.

It needs to be specifically outcomebased. Those tools are going to take the bottom of the market and then AI enabled services are actually going to eat into the service revenues from these company as well. And again these are outcome driven and also enterprises are going to be reducing their headcount as AI agents start to do more and more of the work. And a lot of these companies are based on a per seat pricing model where they charge basically per human that is using their tool.

So when there's less humans they're going to be making less money. And the overall thing to know about this is customers want outcomes, not features the majority of the time. So, so what should you build instead? Well, the answer to that is twofold.

It's basically outcomebased apps and AI enabled services. So, if you're building an application, you want to tie it to a specific outcome that someone's going to get. And if you already run a service business or you provide a service to businesses, we want to build an AI enabled service business. And this is what I'm going to show you how to create or how to start creating in this video.

So before I continue, I want to talk about the sponsor for this video, Zero. Now you can build an entire AI enabled services business in a weekend. As soon as you actually start making money, the business side gets chaotic pretty fast and you end up with invoices in one tab, bills in another, constantly refreshing your bank balance, hoping that it all makes sense. But with Zero, this is completely different.

You get your invoicing, online bill payments, bank feeds, and analytics all pulled into one view. So you can see the money in, the money out, and what is left to spend. And if you're building a software or AI enabled service business, it just means that you don't have to worry about that anymore. On average, they found that small businesses were saving up to 5 hours a week running their finances through zero.

So, if you're worried about what to do when your AI business actually starts making money and how to actually keep track of everything from invoice payments to app payments and more, then I would absolutely recommend checking out Zero. And they're currently giving anyone who wants to try it out using the link in the description below. 90% off for 6 months. I personally use it for my own business and I wouldn't be able to run my business without it. So, I highly recommend checking it out.

And with that said, let's jump straight back into the video. So, what is this model of AI enabled services? Well, basically we want to build AI enabled services around the outcome, not the feature. So, we want to sell the outcome that we're getting our customers.

Now, this can be something that's maybe a little bit difficult for service businesses and service providers to even think about even if you're just used to selling your services because the traditional way to think about selling a service is to say that like I design or build websites, so therefore I'm going to sell you a website. And so in this model, we want to think about what does the website actually bring? And the reason that we want to do this is because if you sell a website because if you sell a website, you're putting an arbitrary price on this service for like say $1,000.

If you sell an outcome, that means you can charge more. So if we sell an outcome, and the outcome is to convert two times as many customers on your website, that might be worth 20, 30, 50,000 to a business. And so suddenly you can charge a lot more money for this service. So we always want to be charging based on outcome rather than the actual thing that we're delivering or based on time.

The second reason to sell an outcome is because if you sell a thing that basically takes you a certain amount of hours and clients will attach value to time, the amount of time that it takes to actually do this thing. And so the less time that it takes you to do this thing, the lower value clients will think that your service is. And this is counterintuitive if you have AI enabled services because AI basically enables you to do that service a lot faster. So if you use AI to empower your services, you can do it faster and therefore clients have a lower perception of what that service is.

Whereas if you tie it to an outcome, basically clients don't care. Clients don't care about how long it takes as long as it gets them the outcome. The better the outcome is, the better the value is for the client and the more you can charge. So, the best way that you can think about optimizing this is in three ways.

Basically, optimize the outcome you're getting for the client. optimize the delivery time um in terms of how fast you can actually do it using AI and then scale the system that you use to do that with either more people or bringing on more clients or basically turning that into kind of like a productized service that you can sell. That is the way to do AI enabled services. So it so when we think about these three things, how do we optimize the outcome? How do we optimize the delivery time?

And how do we scale the system for AI enabled services using this website building example? So say instead of selling a website, I sell a service that helps businesses convert two times as many customers on their website and that is worth $10,000. So maybe currently I would take, you know, 2 weeks to a month to be able to go through that project manually, look at all the information, review all the pages, identify where the weak conversion points are, do designs to improve that conversion, deliver them back to the client, get them to review those, come back.

It's a very long manual process. So if we're optimizing the outcome, we want to think about how can we use AI to improve the outcome that we're getting for the customer to make it better. So, we're going to basically build custom apps and agents to do this delivery process. And this is where this fits in.

So, these custom apps and agents are going to do these three things. They're going to optimize outcome. They're going to optimize the delivery time. And they're going to scale the system.

So we might build a custom application that goes and reads through an entire website and then creates a dashboard with a report of all the weakest conversion points. Maps out all the like takes screenshots of the entire user journey for the app. Identifies the key areas that we can change to improve conversion for people coming to that website. Connects to the client's data and brings in data sources to actually feed it information about that as well.

That is an example where we could improve the outcome and also optimize the delivery time for that service using AI. And then when you want to scale that system, you can do things like giving your team members access to that application to be able to run that system themselves so that you can start to serve more customers and scale that service through these AI applications and AI agents that you're building to empower that service. Now, you still need to provide expertise and quality control, but instead of being 100% human, what we're going to you want to go to something more like 80% AI and 20% human.

And not only is this going to make the process of delivering this service way, way easier. It's also going to make your margins way, way higher as well. And it's going to mean you can serve loads more clients. And the other piece of this that you need to know is that your pricing needs to be based around the outcome that you're providing as well.

So if I am claiming to help a business convert two times as many customers through my AI enabled service and the value to that business is going to be $50,000. I can easily charge$10,000 for it because I'm saying if you pay 10,000 you're going to get 50,000. And all businesses, regardless of what service or app they're paying for, want to know that if they put money in, they're going to get more money back out. And so that is the way to do AI enabled services.

And so what are some outcomes of AI enabled service businesses that you could build? So you could build an AI powered content agency. So the client pays 2K a month for 30 blog posts. If you build a tool or an app with clawed code that takes their brand guide and topic list, generates drafts in their voice, and outputs formatted docs ready for review.

You could do one person with 10 clients, and this could easily be 20K a month. I don't see why you couldn't do 20, 30, 50, 100 clients with that. The margins on that are insane. You could do an AI proposal or grant writing service.

Basically, where small businesses and nonprofits pay 500 to 2K per proposal. You build an app with an agent connected that ingests the grant requirements, pulls from a library of winning templates, and generates a tailored first draft, and then you refine and customize it. Now, all of these ideas basically work really well if you already have skills and knowledge and proof in those industries to build on top of. And that is the key to building one of these AI enabled software businesses and to building applications with AI that can empower your service business rather than building an app that you're going to try and launch as a separate startup and make money independently of the thing that you already know and do really really well.

AI bookkeeping or financial reporting, AI recruitment screening, AI market research service. The whole point of this is to build an application, build applications and agents that take your existing knowledge and workflows and turn them into leveraged AI enabled services. And this is how you can use AI to empower your service business today to actually make it incredibly scalable, incredibly profitable, and grow incredibly quickly with all of these AI tools. And so, how do you actually find an idea?

If you have a service business already, you can think about how you can build tools that match your own workflows, do an audit to figure out what you're doing, and how you could focus on optimizing those three things. How do we optimize the outcome with AI? How do we optimize the delivery time with AI? And then, how do we scale that system through our company?

If you're trying to find an idea of how to do this, then h then you can copy this prompt. And I'm going to put this prompt down in the description down below. But basically, this will help you take your existing skills, your existing knowledge, and turn that into an AI enabled service where you can build custom applications, build custom agents to help you do those three things and build an insanely profitable business. So basically the way this works is you say, I want to identify a profitable AI enabled service business I can start based on my existing skills and experience.

This business should use customuilt AI tools built with claw code, cursor or similar to handle 70 to 90% of the delivery with me providing expertise, quality control and client relationships. Here's my background and you have to give it some information about your background here including topics, tools and software that you use, professional experience, etc. Based on all this, suggest 10 AI enabled service business ideas that meet all of these criteria. The service addresses a painful recurring problem.

I could build a simple app or tool to automate this. My existing knowledge is enough to quality check and improve the tools output. Each client could realistically pay 5 to 5 thou 500 to 5,000 per month. I could land my first paying client within 30 days.

The work is repeatable and it has a similar process for each client. And for each idea, tell me the service in one sentence, who the ideal client is, what tool I'd build to power the delivery, what tool the tool handles versus what I handle personally, realistic monthly pricing, how I'd find my first three three clients, and why this is better than the client just using chat GBT themselves, because it absolutely will be. And the way that you wrap this up in a product is going to help you sell this as a productized service rather than just a client going to chatb or claude and thinking, I don't know what to do.

I can do anything that I want. I don't know how to set this up in the right way. I don't have the knowledge to quality check this and I can't do this quickly enough or even set the thing up that I need. So, I'm just going to jump over to Claude.

I'm going to paste this prompt in. I'm going to update the details of this here. So, we're going to change some of this information and I'm going to see what it comes back with for me for building an AI enabled service business to give you an idea of the sorts of things that you could be building using AI to actually enable your service business to enable your service business to scale even further or to launch your own AI enabled service business. All right, so I've just filled out all of the information for me.

So, let's send this off to Claude and see what it actually comes back with here. Okay, so now we're getting some AI enabled service businesses for my own knowledge, my own skills here. And interestingly, we've got like some that aren't necessarily about via coding and AI or startups or products because that's a lot of what I've done in the past, but I added some other stuff in here to make this a little bit more interesting. So, we've got YouTube growth engine for B2B founders.

So, weekly title, thumbnail concept, and hook delivery for founders growing personal brand channels. Ideal client is a B2B SAS founder making 1 to 10 million AR post posting one to four videos per month treating YouTube as top ofunnel for hiring and sales. So we can build a tool here that takes their channel and five competitor channels via the YouTube API runs each video idea through my title psychology framework outputs 20 scored variants per video plus thumbnail concepts. And this could be a service that you provide if you're doing something like so people still do like thumbnail design and YouTube kind of optimization as services.

If you can use AI to enable that service and serve a lot of clients, you could easily price. So it's saying pricing 1,500 to 3,000 per month for four videos packaged. First three clients pitched to five founders from your advisory list. So this is interesting because it's tying into my own kind of background here.

So a my own network which is I know a lot of startup founders. So can I go and reach out to those founders to actually launch this service business or a landing page conversion rate optimization retainer for SAS products. So being able to optimize landing pages to increase conversion. So targeting these SAS startups at 20K to 200K MRR and then basically building a tool here that audits adds copy variance design specs and test brief documents and you handle picking the winning angle design direction and prioritization of which test runs first.

So it's got broken down in here the things that you have to actually do alongside this as well. Now you could do £2,500 to £5,000 a month for one major rebuild and two AB tests. I actually think that's a little bit punchy for that service. I think I'll probably come in a little bit less than that.

Productized MVP plus iteration service. So, a vibe coded MVP service. So, 30-day MVP build for nontechnical founders. And charging 5,000 for a build and then 1,500 to 3,000 a month for a retainer after that.

I think that's a little bit punchy as well. Maybe maybe the 5,000 is pretty good starting point, but you could definitely do that. I could definitely do that as a service. And then I'm automating most of the build process using my my Plaid kind of methodology. and coding agents to kind of do that.

I have a pretty streamlined process where I can build that very very quickly at the moment. AI personal brand content engine or PRD as a service. Not sure about PRD as a service actually works. ND AI personal brand content engine is pretty interesting.

So I basically take a bunch of transcripts or podcasts or internal calls or Loom videos, run voice analysis and generate posts for content for that person. Maybe a SAS onboarding optimization retainer. So being able to optimize the onboarding flow for different SAS or software platforms, YouTube clip factory for founders and podcasters. So taking YouTube videos and then clipping them up for people using AI and building a process into that.

An AI coding implementation for agencies. So quarterly claude code and cursor playbook and monthly optimization retainer for service businesses. App store listing optimization retainer. So, optimizing people's app store listings and pitch deck and investor update retainer as well.

So, these are pretty interesting. I think some of these are a lot stronger than others. But you can see here it's giving me ideas for service businesses that I could do where I could use AI to enable the service. I could charge a monthly retainer to actually do that service and use AI to handle the majority of the delivery there and tie that into my own personal frameworks and knowledge that I have rather than just letting someone use chat GPT or Claude blindly to create their own things.

And especially if you can tie this into proof that what you do can actually generate outcomes for people, then you're going to have a much easier time selling this in. and the ability to sell something like a content strategy or content generation service for a business. If you have proof that it's going to increase their views, get engagement, and get new clients for them, and those clients are worth a specific value, you can easily reverse engineer the outcome back to a price that makes sense for that business.

So, if you're going to get the business five new leads per month and one to two of them is going to convert to a client that spends even 5 grand a month on their service, then that is worth 10K a month to that business. So you can easily charge a very reasonable amount of money on a monthly basis to offer that AI enabled service and you're going to be able to scale to a large number of clients because a lot of that delivery is is actually powered by the AI tools and applications that you are building.

So how so how do you actually build this and launch it? So I want to go through a process to actually start building and launching your own AI enabled service business here. I would say don't overthink it and test and validate different offers, different options. Define exactly what the client gets.

So, you need to know what the outcome is. Set your price based on the outcome, based on the value, not time. And I'm actually going to change this one here cuz it's not a great example, I don't think. So, think about it like this.

I'll help you land five more clients a month instead of I'll help you redesign your website. Nobody wants to pay for a website redesign, but people will absolutely pay to land five more clients a month. And the way that you're doing that is you're optimizing their website. And the way that you optimize their website is through an AI enabled service that allows you to optimize at a very fast pace to a very high quality.

Create a onepage service description and AI can help you draft this onepage service description. Keep it very very straightforward, very simple. Don't over complicate it. Make it very very easy to to understand.

Build a simple landing page. You can do this with claw code in an hour. Just use Nex.js to build a simple website. deploy it on Versell, buy a domain, get it up there, and you can use AI to research and find the best landing page conversion principles for copywriting to actually write the copy for that as well. And the most important line on your landing page is basically in this format.

It's here's what you get, here's what it costs, here's how we deliver it, and here's what happens next. So, I'll get you five more clients per month for your business. It's going to cost you $2,000 per month. I'm going to deliver this by our custom website optimization engine.

And all you have to do next is book a call in my calendar to get started. And so a business can easily understand exactly what they're going to get, exactly how much it's going to cost them, exactly how you're going to do it for them and also what they need to do as their next step. So what happens when you actually get clients and you need to actually deliver this service and you've gone and we've already built our landing page here. So you can build a landing page in 1 hour, get it live with claude code and versel and now you need to actually start automating and enabling your services with AI.

So we want to build our app and our agent as an as a quick MVP based on how we do that process for existing clients if you run an existing service business or how you would do that process based on the definition that you've put on your landing page. I would then also create content to test the angle and the offer in your existing network with people who might actually want to buy that service. So go and look at other posts on social platforms that already work well and restructure that for your own offer, for your own idea here and ask people to reach out to you if they're interested.

And then I would also start messaging real people about this service that you're offering. Message real people to see if this product, if this new offer is going to land with them or not. And some MVP examples here would be like if you're a content agency and you want to build a tool that ingests a brand guide and a brief and outputs formatted drafts for businesses, you could absolutely build an app that does that using AI. Or you could build a research tool that scrapes competitors, pulls their reviews, and generates structured reports and even better pain points that you can that that company can address in their own product. or you could build a screening tool that passes job descriptions, scores resumes, and drafts outreach emails if you're a recruitment agency.

So, we're building these custom tools to leverage and enable our own services to scale even further so that we can grow our businesses and launch new AI enabled service products. And I would say to start off with outreach and sales is key. You have to tell people about what you're doing. Your tool is your competitive advantage, but it's not the outcome that your customer is going to get.

Anyone can paste a prompt into chat GBT, but nobody else has your customuilt delivery system tuned to this exact problem. And that's the difference between a freelancer and an AI enabled business. And this is basically your unique business operating system. And you can do it all with these two things of custom applications and agent workflows that you can build into your own business.

So you can see how building a different type of application with Vive coding tools, instead of thinking about how do you build a startup that can launch and get loads of customers and become its own business, you can think about how can you build custom applications for your own business or to launch a new business based on the knowledge and the expertise that you already have. And if you combine these customuilt applications with agent workflows in these tools, you can feasibly deliver 70 to 90% of the service that you're offering to clients all through AI tools in a fraction of the time and at a fraction of the cost.

And if you tie this to an outcome that is super valuable for the businesses or the clients that you're servicing, then it is very easy to charge a very high price for these as well. especially because so many businesses out there don't understand how to set these things up for themselves or would even have the willingness to start thinking about how to set these up. Ultimately, if you can deliver an outcome for someone and the amount that you're charging is less than the value of that outcome, you are always going to get people paying because it's an easy win and it's a no-brainer decision.

They put in a certain amount of money and they get out five times more. And if you can wrap that in a service empowered by these custom AI tools and sell that to the right clients, then you are going to build a business that is going to be successful for a very long time. And if you're interested in learning how to build applications with AI and build these custom business applications as well, then I have a community helping people do that. And you can find out more information over at school.com/iapps.

There's over 250 people building applications inside of this community, launching them to real customers and businesses who are building their own custom tools as well. And joining the community will help you understand how you can build these custom tools for your own business and for your own skills as well. If you enjoyed the video, don't forget to like and subscribe. Thank you for watching and I will see you next