For as long as anyone in the industry can remember, shops have relied on industry software to help organize, price, and keep in contact with our customers. We’ve been 100% renting our collective shop brain from someone else for a long time.

And not just one tool.  Many tools.  It’s, in fact, a Frankenstein of tools.  Sometimes they don’t even talk to or are connected to each other.  These days, it is even common to have tools that serve as intermediaries between tools.

Quoting, ordering, invoicing, production scheduling, online stores, marketing, financial tasks, and more all live on rented land.  When these tools can’t do something you desperately need, such as accurate data reporting, you probably built some sort of workaround and move on.

After all, building your own shop system was never on your bingo card.  That has always taken a team of brainiacs, a lot of time, and stacks of cash you don’t have.

With AI, all of that has changed.  Basically, any shop owner, production manager, or other employee can instantly become a software developer.

This past year, I’ve spoken with several dozen shop owners who have created their own suite of tools using AI.  Some are simply for one aspect of understanding something in the shop, like production data.  Others are full-blown replacements for what they have been using to run their entire businesses.  They aren’t waiting for the industry legacy software programs to give them what they need.  They are building it themselves. 

This is happening right now.  Maybe with one down the street from you.

For the first time in forever, you now can own how your shop runs instead of renting it.  Think about that the next time you have to pay a fee to get access to your own data.

What Have You Stopped Noticing?

Let’s start with something that might be a little uncomfortable.  A long time ago, you probably made peace with a few things you maybe shouldn’t have.

  • The inability to obtain accurate data about some crucial aspect of your business because it is not available to you.
  • Maybe the spreadsheet you keep on the side, because your shop software can’t do the one thing you actually need.
  • Some financial question about your shop you can’t answer without three tabs open and a calculator.
  • A task you repeat often, such as manually scheduling orders to a press for a particular day next week, instead of having a defined process handle it for you.
  • Company-wide information sharing, based on real-time data and current order progress across all departments.  If Domino's can track your pizza delivery with an app, why can’t we see where an order is in the production process?
  • Pricing that isn’t built on real costs with a net profit goal in mind.
  • Whatever annoys you the most today.

A long time ago you probably gave up on this stuff.  It is what it is.

Well, what if it isn’t anymore?

One Thing to Consider

Naming a list of frustrations doesn’t mean the answer is to necessarily build your own software.  

One thing to consider is that you currently aren’t using the tools you are paying for as intended.  I’ve been to a crazy number of shops over the years, and a lot of the frustration a shop feels comes from a lack of processes and accountability to handle their tasks in one agreed-upon shop way.  

Frank does it differently than Susie.  Bill does things in his own special way.  Usually, the business owner doesn’t follow the established processes and is their own internal source of frustration, too.

Before you go on an adventure building your own tools, stop and consider that much of your pain could be caused by an internal lack of controls.  Honestly look at this.

And another thing.  Building your own tools sounds like an in-thing to do.  This year and next, you are going to hear a lot about it.  But I can tell you that it isn’t for every shop.  To use an analogy, probably the majority of the people would rather buy a car off of the dealer lot than build one from scratch themselves, regardless of how much easier AI makes it.

This article is aimed at that shop that has the real need, process workflow, and alignment that make something off-the-shelf never feel correct.  Is that you?  If so, then the rest of this article might be something you need to hear.

What Pain Are You Actually Trying to Solve?

Before you dive into the AI Software build pool, let’s review why you want to start.  Let’s get specific on what’s not working the way you want.  “Our software sucks” isn’t a problem you can fix without defining it in detail.  You can’t build a working tool on feelings.

Do any of these feel familiar?  This, by far, isn’t a complete list either.

  • Data you can’t access.  Do you have years of order history locked inside a platform, and does pulling your own data feel like a knife fight every time you do it?
  • A schedule you can’t trust.  Nobody in the shop can look at a schedule and see at a glance progress on orders.  Are they ok?  Missing something?  Already shipped?  Look at your schedule right now.  What are you working on two Wednesdays from now, and what is the first order you will run?  If you can book a plane ticket six months in advance and know you need to be at the airport at 9:30 am, why can’t you do the same with the orders you are taking?
  • Profit margins per job.  Which jobs made you money?  Lost money?  Are you making the same net profit % on a hoodie as on a t-shirt or a bandana?  How do you know?
  • Intake chaos.  Orders scheduled for production arrive by email, by text, by phone call, from your online stores, or from an API you’ve set up with a client.  Are they all following the same rules, guardrails, pricing, scheduling, and data collection?  Where is a human involved, and where can an AI agent handle some tasks?
  • Information repetition.  I’ve seen shops copy and paste information by hand from one platform to another as part of their daily grind because the software they use doesn’t talk to other software they use.  Madness!
  • Employee Time Data.  You have an entire string of employees touch every job you run. How much of their time is applied to an order, so you can track expectations and profitability?  If an order was supposed to take 30 minutes to run and it took 90, does anybody even know?  How much easier and more accurate would production scheduling be if you baked into employee time averages for tasks based on the work for that order? 

I’m sure you have your own list.  If you don’t, then spend 5 minutes with a legal pad and jot down the things that would feel incredible to fix.  At this point, you don’t have to know how.  Just list what bugs you.

The next step is to sort this pain into two piles.

The first pile is a “missing-tool problem.”  This is because the software you are using doesn’t do the thing you need.  Whatever that is, add it to the pile.  Even the smallest, or most ridiculous wish.  We’re brainstorming here.

The second pile is the “habit or process problem.”  Be honest with your shop.  Does a good chunk of what frustrates you live in how somebody does something, and not related to the tools you are using as much?  Is this a training issue?  An accountability problem?  Could it be that a good part of the reason why you want something different is habit-related?  News flash: new software won’t fix apathy, a bad hire, or an employee accountability challenge.

Take a hard look at your constraints.  This might be hard to do.  If you keep booking 20,000-stitch embroidery jobs and are surprised your department is backed up, a slicker way to schedule might not be the key to getting more orders out the door.  Fix what’s actually limiting your output, not what seems like the surface problem.

One more thing.  Whatever you build out in the first phase of your app should focus on the top 20% of your jobs that are most profitable.  When you start, know that the messy edge cases, the weird one-offs, the exceptions to the rules, are always going to not fit just right.  Be able to identify those types of potential challenges so they don’t throw you off your game. 

Start with what you see the most.  Refine things after that is working.

Own Your Data First.  

Regardless of whether you build your own app, you should absolutely find a way to secure all of your data.  Not on the day you decide to leave your current platform.  Now.  While everything is still calm.

Start exporting all of your data.  Orders.  Customers.  Financials.  Art information. Production metrics.

If you are not ready to build now, the day you finally decide to make a move, all of your history with your shop software could be exactly what keeps you stuck. That’s not a small inconvenience.  That’s your company’s entire memory, held for ransom.  

Owning your data is the starting line for a lot to follow in the next few years.  You can’t switch, build, or make smart decisions with data you can’t access.

But once you have it, this will be the first real act of owning how your shop runs instead of renting it.

Here’s What Matters Most

Let’s be clear.  Floating around the internet, in Facebook groups, at trade shows, and most certainly from the shop software people, are scary stories about “homemade software” that becomes unfixable, then mysteriously blows up at 2:00 am one day, and the business is doomed.

The real risks are boring, specific, and almost certainly fixable.  Think about these:

Risk 1: You just created a new job

The moment your production manager (or you) builds a tool, they instantly become the owner of it.  On top of their actual job.

Here’s the risk.  If that person gets promoted, quits, burns out, or doesn’t have time to support what you built, do you have anyone else tapped to run the thing?  This is a right-person-in-the-right-seat problem, not a software problem.

To help mitigate this risk, you build out the process correctly, just like you do with any other position in your company.  Write down how it works in plain English that any human can follow.  Make sure more than one person is trained.  Like anything in your business, the correct answers can not live inside one person’s head.  Who owns who trains the backups?  Map that out and implement it.

Risk 2: The unglamorous stuff

Let’s talk about the things that are not obvious.  

Are there data and code backups, and are they generated continuously without human involvement?  Is the math your software is built on actually working as intended?  How do you know?

You should be prepared to understand and budget for a running tab of costs to build out what you need.  Building your app will cost credits.  Keeping it alive and functional will cost a lot more.  You are going to pay for compute power, and costs could increase as your shop grows and the app works harder.  

Also, don’t forget that you’ll probably pay for a place to store your code.  It is free on some platforms today, but “free” is a promise that can change quickly.  Added to the bill will be long-term storage, backups, and if you wire AI into your tool, you’ll pay for tokens on every task the AI runs.

Something to keep in mind.  Right now, AI companies are subsidizing token costs to get everyone hooked.  Eventually, those subsidies could go away, and the cost could jump in a way that changes your whole math.  Be prepared for that.

What happens if your AI raises its rates, or goes out of business? Could you move to another platform easily?

Problems with items like these don’t announce themselves.  Are you ready for contingencies?  

It’s easy to skip over stuff because you don’t know what you don’t know.  Before you get too far down the road in your build, put some thought into these types of questions.

Risk 3: The Security Gap You Can’t See

When you build software by describing what you want, you get exactly what you describe.  Logically speaking, the flaw here is assuming that the build contains things you haven’t described.

You don’t ask for a lock on a door for a room you didn’t know you had.  

The same AI that built your tool is the same AI you are entrusting to secure it.  If you don’t mention security, don’t be surprised when the AI tool skips that protection part, or even informs you that it is missing.

Here are some security ideas you can implement.  First, early in the build, ask your AI tool “What security best practices for an app like this have we not covered?”  Build and verify those.  Then, ask your AI to point you to a current security scanner made for AI-built apps, then run it.  These are cheap or free, and they catch the common, findable holes.  Finally, for top protection, you might want to reach out to a security professional who does this for a living to review your build before you go live.

Some security perspective here.  Building an internal tool for your team to use on your own network, without customer data or payments, doesn’t need the full treatment.  Anything that stores customers’ personal information and takes money absolutely does.

Look at the Other Side of the Coin

Building your own shop tools will make you feel incredible.  Just think, “YOU MADE THIS!” Now you have something that works just the way you think and how your shop operates.  If that doesn’t put a large grin on your face, I don’t know what will.

Creating your tool is the easy top 20%.  Running it and keeping up with it for years to come will be the other 80% of the equation.

The updates, fixes, changes, training, and process improvements that will be needed have to live past the person who originally built the app.  Owning, and not renting, your shop software means you have signed up for that.  

That has to be part of your process, accountability, and management structure too.  The fun part is building the tool.  The slog is keeping it running.

What Will You Have More Of?

Frankly, the AI tools everyone has access to these days make it easy to build your own shop software.  But the tools don’t know or understand your business.  How you work.  Your expectations for success.  The subtleties that only you understand.  Getting all of that baked in is 100% on you.

But let’s say you do it.  It’s built and running like a racecar.  If you do this right, what will you have more of?

  1. More visibility into your own shop.
  2. More decisions made with your own data, as you know what you are doing three weeks from now and if jobs are profitable.
  3. More control on how work moves through your building.
  4. More money, as you can save on subscription fees and higher credit card charge fees they make you pay because you are hooked into their platform.
  5. More time, as you can build in automations.
  6. More responsibility to keep your software running perfectly.
  7. More autonomy to build your future, your own way.

Are you ready for that?  This means you are setting these tools up, but you need to build them so they don’t collapse the day your main employees walk out the door.  Ownership is both the reward and the responsibility of building your own tools.

By the way, if the build-or-buy question on your shop software is keeping you up at night right now, it might not be a software problem.  It’s more than likely a strategy question about your shop and your goals, and it is the kind of thing worth talking through with someone before you spend months traveling down the wrong road.  That’s what I’m here for.  Grab some time here.

Before You Do Anything, Run This List

As a closer, I’ll leave you with this short list of actions to take that will help you on this topic.

  • Set a repeating monthly reminder to export all of your data into a secure spot.  Make this a habit.  Get it on your calendar.
  • Document the pains you want or need to resolve.  Ask your staff what bugs them the most. Prioritize what helps your shop the most.
  • Mark each of these pains as either a missing-tool problem or a process problem. 
  • Give your list to your AI tool and see how they would sort them.  How does it compare with how you did it?
  • Name your real bottleneck and ask whether a new tool would solve the problem
  • If you have already built some AI tools for your shop, great!  Get more people involved in running, updating, and understanding them.  Decide today who owns them and will be responsible for them going forward.
  • Ask your AI to point you to a current security scanner and run it.

  • Get any customer-facing AI tool you’ve built reviewed by a security professional.  Do what they say to protect your customers and your business.