The stack nobody audited: dozens of subscriptions, most underused
Six AI tools, under $300 a month, replacing tens of thousands in annual SaaS spending across every department.
Pull up your company's SaaS spending. Not the number your CFO quotes at the quarterly review, the real number. The one that includes the marketing team's analytics dashboard, the sales director's proposal software, the HR manager's onboarding platform, the office manager's visitor check-in system, and the seventeen other tools that auto-renewed last month while nobody was looking.
For a mid-size company with about a hundred employees, that number averages $47,000 per year.
Let that land for a second. Forty-seven thousand dollars. Not on salaries, not on office space, not on the products you sell. On software subscriptions. Tools. Platforms. Dashboards. Half of which your team uses at maybe 20% capacity, and a quarter of which nobody has logged into since the free trial ended.
This book is about killing the ones that deserve to die.
Every department has them. The tools that seemed essential when someone signed up during a busy Tuesday, the ones that solved a real problem for about two weeks before becoming background noise on the credit card statement.
Marketing runs a dashboard builder at $3,600 a year to visualize data that lives in a spreadsheet. Sales pays $4,800 annually for a pipeline analytics add-on that three people actually open. HR subscribes to a performance management platform at $4,800 a year, using it mostly for review templates. The office manager has a visitor check-in system costing $3,600 annually that amounts to a glorified sign-in sheet. IT maintains a knowledge base platform at $6,000 a year that could be a well-organized shared folder.
These are not bad tools. They work. They do what they promised. The problem is simpler and more expensive than that: they are wildly overkill for what your team actually needs from them.
You are paying enterprise prices for features you could now build yourself. Not "yourself" as in you need to become a software engineer. "Yourself" as in you type a description of what you want, and an AI tool builds it for you, often in minutes.
That shift happened. It happened fast. And most companies have not caught up yet.
This book covers six AI tools. Combined, they cost under $300 per month. Together, they can replace tens of thousands in annual SaaS spending across every department in your company.
An autonomous agent that works with your files. You give it access to a folder on your computer, tell it what you need, and walk away. It reads your documents, understands the context, and produces work. Content briefs from your strategy docs. Proposals from your client notes. Onboarding packets from your HR templates. It is the closest thing to handing a task to a capable assistant who already has access to everything they need.
A terminal-based AI coding agent. It reads your project files, writes software, fixes bugs, and builds real applications. This one sits higher on the technical spectrum than the others, but the key is: you do not need to be a developer to use it. With clear instructions and guided prompts, a non-technical person can direct Claude Code to build a working web app, a data processing script, or an internal tool.
Interactive tools built inside a conversation. You chat with Claude, describe what you want, and it creates a working component right there in the conversation window. Calculators, comparison tools, scoring systems, budget trackers, quiz interfaces. No code required on your end. You can share them via a link with anyone.
A desktop command center for parallel AI agents. Think of it as dispatching multiple workers on different tasks at the same time. One thread builds your dashboard layout while another processes your data. You review the results, approve the changes, and end up with real applications. Runs on macOS.
A free browser-based platform that builds apps from descriptions. Its Build mode takes a natural language description of what you want and generates a working React application. Its Annotation mode lets you point at parts of the interface and click to edit them visually. No local setup, no installation, no cost. This is the tool that makes non-technical people genuinely dangerous in the best possible way.
Push-to-deploy cloud hosting. When you build something with the other tools and want your team to access it, Railway is how you put it on the internet. Connect a GitHub repository, and Railway gives you a live URL in minutes.
This is not a book for software engineers. If you write code for a living, you already know most of what these tools can do, and you probably have opinions about it.
This book is for the marketing manager who is tired of paying for six tools that each do one thing. It is for the sales director who knows there has to be a better way to generate proposals than a $3,600-a-year platform. It is for the HR manager who just wants a simple leave calculator without subscribing to an entire HR suite. It is for the office manager who needs a visitor check-in system, not a $3,600 annual contract. And it is for the freelance IT consultant who pays for every subscription out of their own pocket and feels every dollar.
You do not need to know how to code. You do not need a computer science degree. You do not even need to be particularly technical. You need to be able to describe what you want in plain language, follow step-by-step instructions, and review output with a critical eye.
If you can write a clear email, you can use these tools.
Not every use case in this book requires the same comfort level. Some you can do on your first day. Others need a bit of technical familiarity, though the AI handles the hard parts. Every use case is tagged with one of four skill levels, so you know what you are getting into before you start.
No technical knowledge required. You type what you want in natural language and the tool handles everything else. If you can have a conversation, you can do these. Most Claude Artifacts and Claude Cowork use cases fall here.
You can follow step-by-step instructions without panic. You are fine navigating folders on your computer, editing text files, and clicking through a browser interface. Google AI Studio's annotation mode lives at this level.
You understand basic web concepts. You know what a URL is, you have seen a dashboard before, you can look at generated code and tell if something looks obviously wrong. You can push a file to GitHub by following instructions. This is where Codex App projects and some Claude Code builds sit.
You need some comfort with technical tools, but the AI does the heavy lifting. You can open a terminal and type a command when told exactly what to type. You are not writing code from scratch, you are directing an AI that writes code for you. Claude Code deployments and Railway setups live here.
Nobody stays at one level. The marketing manager who starts as a Beginner on content briefs builds confidence. Within a few weeks, she is Comfortable. Within a month, she is trying Intermediate use cases.
This book is organized by department, because that is how you actually work. You do not care about "AI tools" in the abstract. You care about the specific problems on your desk this Tuesday morning.
Each department chapter contains ten use cases. Each use case follows the same structure: the problem you are solving, the tool you are using, a step-by-step walkthrough of how to build it, what SaaS subscriptions it replaces, and exactly how much you save.
Follows Maya, a marketing manager at a 60-person B2B company. Her team runs content calendars, social media campaigns, performance dashboards, email sequences, and competitive intelligence. She is paying for at least eight different tools. Use cases range from a content brief factory (Beginner, Cowork) to a campaign performance dashboard (Intermediate, Codex App) to a weekly report automator (Guided Technical, Claude Code).
Follows Priya, a sales director at a 40-person SaaS company. Her world is proposals, pipeline analytics, battle cards, client portals, and forecasting. Sales teams tend to accumulate tools at an alarming rate because every quarter someone finds a new platform that promises to close more deals. Use cases span proposal generators (Beginner, Cowork) through client portals (Guided Technical, Codex App plus Railway).
Follows James, an HR manager at a 50-person company. HR software is notoriously expensive relative to what most teams use it for. James pays for onboarding platforms, performance management suites, benefits tools, and an LMS. His team uses about 15% of those features. Use cases go from onboarding packet generators (Beginner, Cowork) to an employee handbook chatbot (Guided Technical, Claude Code plus Railway).
Follows Raj, an IT manager at a 70-person company. IT readers prefer straight talk, so this chapter leans heavily on direct address. Raj's stack includes knowledge bases, ticketing analytics, status pages, and asset trackers. Use cases range from security awareness quizzes (Beginner, Artifacts) to API documentation portals (Guided Technical, Claude Code plus Railway).
Follows Lisa, an office manager at a 45-person company. This is the most accessible chapter for non-technical readers. Office management involves visitor check-ins, room bookings, supply tracking, and event planning, tasks that feel like they should not require expensive SaaS subscriptions. They should not, and after this chapter, they will not.
Follows Alex, an independent consultant with eight regular clients. This chapter is the most personal. Every subscription comes directly out of Alex's pocket. There is no company credit card absorbing the cost, no IT department negotiating volume discounts. The math hits different when it is your money.
Pulls it all together. Every department's savings tallied, every subscription cancelled or downgraded, every dollar accounted for. The net number, after paying for the AI tools themselves, is the number that matters.
Across all six departments, the SaaS subscriptions replaced in this book total between $120,000 and $180,000 per year for a mid-size company. That range depends on which tools your specific company uses, how many seats you are paying for, and whether you are replacing tools entirely or downgrading to cheaper tiers.
| Item | Detail | Annual Cost |
|---|---|---|
| Claude Max | Covers Cowork + Code | $2,400 |
| ChatGPT Plus | Codex App access | $240 |
| Google AI Studio | Browser-based app builder | $0 |
| Railway Hosting | ~$50/mo for all deployed apps | $600 |
| Total AI Toolkit | ~$3,240 |
Even at the conservative end of the savings range, you are looking at net savings above $115,000 annually. That is not theoretical. That is the sum of specific, named SaaS subscriptions that you cancel or downgrade after building replacements with these six tools.
For the freelance consultant, the math is smaller in absolute terms but larger in impact. Replacing $10,000 to $15,000 in annual subscriptions with tools costing under $3,000 a year is the difference between a tight year and a comfortable one.
Chapter 7 shows every line item. No hand-waving, no "estimated" savings without specifics. You will see exactly which tools go, which tools stay, and what the net number looks like.
You have two options. The first is to read cover to cover, which gives you the full picture of what is possible across an entire organization. The second, and the one I recommend, is to jump straight to your department.
If you run marketing, go to Chapter 1. If you manage HR, go to Chapter 3. If you are a freelance consultant, Chapter 6 was written specifically for your situation. Read the use cases that match your current pain points, build the first one that resonates, cancel the subscription it replaces. Then come back for more.
Each chapter is self-contained. The tools are introduced fresh in every chapter where they appear, so you will not miss context by skipping ahead. The only chapter that depends on all the others is Chapter 7, the math chapter, which is most useful after you have seen what is possible in your own department.
Start with Beginner-level use cases regardless of your technical comfort. Not because you can't handle more, but because the quick wins build momentum. When you build a working calculator in Claude Artifacts in ten minutes and realize you are paying $1,200 a year for a tool that does the same thing, the rest of the book becomes very compelling.
This is not an AI hype book. I am not going to tell you that artificial intelligence will transform everything, disrupt every industry, and usher in a new era of unprecedented productivity. You have heard that pitch. It is exhausting.
This is a practical book about specific tools solving specific problems and saving specific money. Some of the tools in this book will build things that are 80% as good as the SaaS product they replace. In some cases, they will be 95% as good. In a few cases, they will be better, because you built them for your exact needs instead of paying for someone else's idea of what you might need.
Your CRM stays. Your core infrastructure stays. Compliance-specific tools, deeply networked platforms, anything with years of irreplaceable data, those stay. This book targets the long tail of SaaS, the dozens of point solutions, reporting tools, template platforms, and management add-ons that you subscribed to because building them yourself was not an option.
It is an option now.
Two years ago, replacing your SaaS dashboard with something you built yourself meant hiring a developer, scoping the project, waiting weeks, and paying thousands. The dashboard you got was probably worse than the one you had.
That calculus changed. The tools in this book let a marketing manager, not a developer, describe a dashboard and get a working version in an afternoon. They let an HR manager build an onboarding system without writing a line of code. They let a freelance consultant create a client portal that used to require a $6,000 annual platform subscription.
This is not about AI replacing jobs. It is about AI tools replacing other software tools. The humans stay. The expensive, underused subscriptions go.
The people in this book, Maya, Priya, James, Raj, Lisa, Alex, they are not technical wizards. They are professionals who learned to describe what they need to an AI tool and let the tool build it. The skill ceiling is surprisingly low. The savings ceiling is surprisingly high.
Before you finish this book, pick one use case from your department chapter and build it. Not read about it. Build it.
The first one takes the longest because everything is new, the tool, the process, the mental model of asking software to build software.
The second one takes half the time. By the third, you will start looking at your SaaS subscriptions differently. Not as necessary tools, but as monthly charges that you now have the ability to question, and in many cases, eliminate.
Forty-seven thousand dollars a year. How much of that is sitting in your stack right now, unaudited, auto-renewing, solving problems you could solve yourself with a $270-a-month toolkit?
Open your billing page. Start counting.
