The Agency-to-SaaS Play: Turn Your Service Into SaaS
Right now, I run two businesses.
The first is Flyletter, the AI newsletter writing platform. The second is a done-for-you service where I write and manage newsletters for other founders.
One is a SaaS company. The other is an agency. Turns out, they're actually the same business, and they make each other better every single week.
Most people who build a SaaS product start by building the product first. They pick an idea (likely under-vetting it), burn a few months building it (and a chunk of savings too), and then go looking for someone who wants it.
Usually, nobody does.
But there's an easier path to building a SaaS business, and it starts with something you already know how to do. Sell your expertise and knowledge as a service. Get paid to learn exactly what your clients need. Then build the software that delivers it, and sell that too.
Before long, your SaaS revenue will eclipse your service revenue, and you can go all in on your product.
I call it the Agency-to-SaaS Play. I'm running it right now with Flyletter, and I've watched several friends run it successfully, too.
I'm going to walk you through the playbook, why it's easier to pull off now than ever, and the exact steps to run it yourself (plus what to do if your agency falls flat on its face).
Let's get after it.
Why an Agency Is the Path of Least Resistance
If you've spent any amount of time in your career, you've already built a skill someone will pay for. Maybe it's bookkeeping, design, paid ads, or writing. A company paid you a salary to do it. Other companies will pay you a fee to do the same thing.
That's all it takes to start an agency or consultancy. Identify a core competency you have that others lack, and offer it as a service to clients. There's no product to build or funding to raise, and you get paid as soon as you deliver.
And today, one person can now deliver what used to take an entire team with the capabilities of AI.
For example, Flyletter's original done-for-you newsletter service only included the writing and publishing of client newsletters. As AI models improved, however, it allowed us to manage the full content funnel and build landing pages, lead magnets (like this one), and offers for our clients that drove better outcomes and made us more valuable.
Of course, agencies are also a grind. Sometimes it feels like I just created a job that works me harder and pays me less. But that's exactly why you don't stay an agency owner forever.
Because every client doubles as paid market research. You're able to see their problems from the inside, learn their workflows, and refine your offer. It's the type of business intelligence a SaaS founder building from scratch would kill for, and you're the one getting paid for it.
How the Agency-to-SaaS Play Works
In the normal SaaS model, you brainstorm ideas, pitch investors or burn savings, build a prototype, and then see if anyone wants it. If no one does (which is likely), you spend more time and money iterating on the product and finding people to give you more feedback.
It reminds me a little of the movie business. You have an idea, write a script, get some feedback from friends and family, hopefully sell it to a studio, get a budget to make it, and then spend most of it before your intended audience even sees it. No wonder most projects lose money, and why most studios stick to yet another superhero movie.
Pretty bad business model if you ask me, and not much different than starting a SaaS company from scratch.
Luckily, there's another path to a product business. I call it the Agency-to-SaaS Play:
- Identify a core competency that someone already pays you for.
- Start by selling your skill or expertise as a service.
- Find an offer clients keep buying and productize it into a repeatable package.
- After you prove it's repeatable, build internal tools to deliver it faster and better.
- Refine the internal tooling, then package it and sell it as a product.
- Use your brand story and client work as social proof for your product.
- Continue offering your service to select clients as market feedback on your product.
That product could be a SaaS app, but it could also be a set of prompt packs if you use AI to deliver the work, or an info product that teaches your process. The key is that you're packaging and selling your unique knowledge as a product rather than a service, which is infinitely more scalable.
Why It's Easier Now Than Ever
A few years ago, the "build internal tools" stage was where most agency owners or consultants stopped. Building software meant hiring developers, and most service businesses don't have the margins to fund a dev team.
That excuse is gone. With AI, a non-technical operator can build working software and automations on their own. I'm a growth guy by background, no technical experience, and even I'm able to build production-ready web apps with Claude Code.
The internal tool doesn't even have to be an app for it to become a product. Just ask Slack, originally an internal communication tool built by video game developer Tiny Speck. They pivoted the company, and were eventually acquired by Salesforce for $27 billion.
For example, the AI CRM I built with Claude Cowork runs Flyletter's outreach pipeline, and it's mostly a Notion database, a written operating manual, and a scheduled task. Completely plausible that as I work with clients I find that outreach is one of their pain points, and I can package and sell the workflow I built for Flyletter.
If you already use AI to manage your business and deliver your service, you're halfway there. The prompts, workflows, and quality checks you run by hand for each client are the first iteration of your product. You just have to wrap it all into something other people can use.
Step 1: Pick Your Core Competency
Start with a core competency you've built in your career, meaning the skills or expertise people already pay for. The operative word is already. Skip the passion projects and the skills you hope to learn someday. You want the thing companies have already written checks for.
Then narrow it down. "Marketing services" is too broad to sell now and productize later. "Newsletter writing for founders" is narrow enough that every client looks roughly the same, making it easier to acquire customers, productize your service, and turn it into SaaS.
How do you know it's narrow enough? If you can explain who it's for and what they get in one sentence, you're there.
Step 2: Sell It as a Service First
Before you build anything, get someone to pay you to do the work by hand.
This is the step I wanted to skip most, because every bone in my body wanted to sell a product over a service. But the service is how you learn what to build. Every client teaches you what they need, what they'll pay for, and what "done well" actually looks like to them.
It also covers your rent while you figure out the rest on the fly (a luxury most first-time SaaS founders don't get).
In the early days of Flyletter, I created a Claude MCP connector that let you use the Flyletter app directly within a Claude chat. It was fine, but didn't deliver the same writing quality as Flyletter's in-app experience.
But it was way quicker and more versatile, and as I started onboarding clients I found myself defaulting to Flyletter in Claude. I quickly realized that for the Flyletter SaaS app to be truly valuable, it needed to deliver the same fidelity in Claude that it did in-product.
That learning not only reshaped the product direction, but improved what I could deliver to the clients while reducing the time I spent delivering it.
Your first clients are usually a few messages away. Former bosses, old colleagues, and anyone who has already seen your work are the warmest leads you'll ever have.
Step 3: Productize the Service
When you first start selling to clients, you don't know exactly what to sell. Your core competency is "content marketing," sure, but what part of content marketing is specifically valuable to your clients, and how do you effectively communicate that value to them?
The first clients who buy typically go through a phase where you discover exactly what they need as you start working with them. This means that in the beginning, you're often creating custom service agreements for each deal, or vastly underestimating the time it takes to deliver what you promised.
As you refine your offer and prove you can deliver real outcomes to clients, turn the work into a productized service package with a fixed scope, fixed pricing, and the same deliverable every time. That way, you can build repeatable systems around your offer, scale more efficiently, and start to feel like you aren't running around with your hair on fire.
Try writing your service down as a one-page offer with a standardized package and price in words your ideal customer would understand. If you can't, it isn't productized yet. If you can, you're ready for the next step.
Step 4: Build Internal Tools Around Delivery
Now build tools that help you deliver the productized service. The goal is higher quality work in less time and at a lower cost to you. Every hour those tools save goes straight to your bottom line.
Build them for you first, and don't worry about anyone else using them yet. Internal tools are allowed to be ugly. They just have to work.
Start with the task you repeat most for every client, or the one that would result in the most time savings. That's your first feature.
One of my friends runs an outsourced CFO and accounting firm. A big part of what his clients paid for was executive-level insights and recommendations for their business, which he delivered by hand through written reports and client meetings.
So he built accounting software that plugs into QuickBooks and uses AI to deliver those same insights and recommendations. The analysis he used to do manually, client by client, became an internal tool that saved him time, and then a product for other outsourced accountants to use.
Step 5: Treat Every Client as R&D
This step is what makes the internal tooling you built worth paying for.
Every client teaches you something, but only if you're paying attention to it.
I built Flyletter for myself. As a busy founder, I didn't have time to keep up with my own newsletter, so I built an AI platform that could help me manage it.
Then I started offering a premium done-for-you service, fully managing and writing newsletters for other founders, and I used Flyletter internally to deliver the work.
Each client showed me more of the patterns every founder's newsletter shares, the edge cases I had to plan for, the quality bar each founder expects, and the workflow I needed to get a founder's expertise out of their head and into a newsletter.
All of it went back into Flyletter. The product got drastically better, and my client work got faster and higher quality at the same time.
Another friend of mine runs a design agency that builds websites and online courses for clients. As he got embedded in their companies, he realized most of them had a lead generation problem. So he started offering lead generation as a service.
That eventually became an AI website builder that builds your site, creates lead magnets and content upgrades for it, and then writes SEO articles to promote them. He never would've found that product from the outside. He found it because clients paid him to be on the inside.
So keep a running doc. For every client, write down what repeats, what commonly breaks, and what you wish your tools had handled for you. Then build the top items on that list.
Step 6: Package Your Tools Into a Product
Once your internal tools reliably deliver results across clients, package them into a product that other people will pay you for.
You have options, and not all of them require a full SaaS build:
- SaaS. Turn your tooling into software people use on their own.
- Prompt packs. If you deliver work with AI, sell the prompts and workflows you use.
- Info products. Teach your process as a course, guide, or set of templates.
Start with whichever option takes the least effort to test. A prompt pack can tell you whether people want a self-serve version of your service before you commit to building the software workflows around it.
A third friend of mine (yes, I have that many friends) ran an agency ghostwriting books for authors. He created some truly impressive AI prompts to help his team deliver books for his clients. With the prompts proven over multiple clients, he built AI software that uses the prompts and helps people write books on their own.
And you don't have to shut the agency down once the product exists. My friend keeps his book writing agency open to high-paying clients, and I still sell the done-for-you service alongside Flyletter for market research.
Step 7: Market It With Your Origin Story
By this point, most of your marketing is already written for you.
Your clients' results prove the process works. The story of building the product inside a working agency proves the product works. Lead with both. Use case studies from your service clients, before-and-after results, and a founder story that explains exactly why the product exists.
Then push that proof out through content and outreach (I broke down the exact system in Three Growth Marketing Pillars That Scale Any Business). Content built on real client results makes your outreach warm before you ever send a message.
What If Your Agency Fails?
Even if your agency or consultancy doesn't work out, you can usually find one thing inside the business that you were great at.
Alex Hormozi tells a story about a client who had a floundering agency. The guy wasn't good at running the business, and he wasn't good at delivering the service. But he was really good at generating leads for his agency.
So he started a lead generation service for agencies, and he built the positioning around his own story. His pitch was basically, "I ran an agency. I was great at lead gen and not so great at the rest, so I'm offering you the part I do best."
It's the same play with a different starting point. Even a failed agency leaves you with paid experience and a story your future customers will relate to. As someone with two failed companies on my resume, I find that quite comforting.
Where This Breaks
Most of the ways this play fails come down to getting stuck in the agency:
- Staying an agency forever. The cash is comfortable, the calendar fills up, and the tools never get built. Block time every week to build, even when the client work is heavy.
- Automating before it's repeatable. If the service isn't productized yet, your tools will just automate chaos.
- Saying yes to every custom request. Each one-off makes the work less repeatable and the product harder to build.
Bottom Line
Most founders guess at what to build, then spend months hoping someone wants it. The Agency-to-SaaS Play flips the order. You get paid to learn exactly what your customers need, you build the tools to deliver it, and then you sell those tools to everyone else.
This week, write your productized service in one sentence, including who it's for, what they get, and what it costs.
Writer, founder, and AI builder. Currently building Flyletter, previously sold Selling Signals (100k+ subscribers). 1 sold, 2 failed, 1 active. This post started as an issue of my weekly newsletter.