Skip to main content
Where AI fits

Where AI fits

Don't let AI lock you into one vendor

The real cost of adopting AI isn't the software—it's losing control of your code, data, and accounts to a vendor you can't leave.

· 5 min read

Lock-in is the hidden cost of adopting AI. It happens quietly, over months, and then you discover you can't leave.

Here's how it works. A vendor builds a custom tool for your business—say, a system that reads your invoices and files them automatically. The code lives on their servers. Your data lives on their servers. Your login account is registered under their email address, not yours. Everything works great for two years. Then they raise prices by 40 percent. Or they decide to stop serving your industry. Or they get acquired and the new owner cancels your product line.

Now you're stuck. You can't take your data and move to another vendor because the vendor owns it. You can't rebuild the tool yourself because you don't have the code. You can't even access your own account because the login isn't in your name. You either pay the new price, rebuild from scratch, or lose the tool entirely.

This happens to Canadian businesses every year. A Burnaby accounting firm paid $8,000 to migrate off a vendor's platform after three years. A Calgary distributor had to rebuild a custom intake system—six weeks of work—because the original builder went out of business and kept the code. A property manager in Surrey lost two years of tenant records when a vendor's servers failed and they had no backup.

The good news: lock-in is avoidable. It just requires you to ask the right questions upfront and insist on the right answers.

Own your data from day one

Your data is the most valuable part of any AI tool. It's your invoices, your customer records, your past projects, your filing system. If a vendor owns it, you can't leave.

When you start a project, ask: "Where does my data live, and can I export it anytime in a standard format." Standard format means CSV, JSON, or a format your accountant or next vendor can read—not a proprietary database that only the vendor can open.

The vendor should say yes without hesitation. If they say your data lives in their cloud and they control exports, or if exports cost extra, or if they need 30 days to process an export request, that's a red flag. A good vendor treats your data as yours. They let you download it anytime, in full, in a format you can use elsewhere.

For a dental clinic in Kelowna with 15 years of patient records, this difference is enormous. If you own your data, you can switch to a new system and bring your history with you. If the vendor owns it, you lose 15 years of context.

Keep your code portable

Custom code is the second lock-in point. If a vendor writes code that only runs on their platform, you're trapped.

When you hire someone to build an AI tool, ask: "Will this code run on standard infrastructure." Standard means cloud platforms like AWS, Google Cloud, or Azure—services that dozens of vendors use. If the code runs only on the vendor's custom infrastructure, you can't move it.

A good vendor will write code that's portable. It costs a little more upfront—maybe 10 to 20 percent—because they're building for flexibility, not lock-in. But it means if you decide to move, you take the code with you. You can hire a different team to maintain it. You can run it on your own servers if you want.

For a 12-person software company in Gastown, this matters. They might build a tool with a vendor, then decide to bring it in-house later. If the code is portable, they can. If it's locked to the vendor's platform, they can't.

Register accounts in your name

The third lock-in point is accounts. If your login credentials are registered under the vendor's email address, you don't control access to your own tool.

When you start, insist that every account—cloud hosting, code repositories, login credentials—is registered under your company's email address or your personal email. Not the vendor's. You should be able to log in yourself. You should be able to reset your password without calling the vendor. You should be able to add or remove team members anytime.

This sounds basic. Many vendors will do it without question. But some will register everything under their own domain, telling you it's easier to manage. It's easier for them. It's harder for you.

A plumber in Richmond discovered this when his AI scheduling tool went down. He couldn't access the account to check the status because the login was in the vendor's name. He had to wait for the vendor to respond.

What it costs

Owning your data, code, and accounts costs more upfront. Expect to pay 10 to 20 percent more than a vendor who locks you in. For a $50,000 project, that's $5,000 to $10,000 extra.

It's worth it. Here's why: if you ever need to move, the cost of rebuilding from scratch is $30,000 to $100,000 and takes three to six months. Paying 15 percent more upfront to avoid that is a good trade.

Also: a vendor who agrees to ownership terms is usually a better vendor overall. They're confident enough in their work to let you leave. They're not relying on lock-in to keep you. That's a sign they'll actually solve your problem, not just trap you in a contract.

Where to start

  • This week, make a list of any AI tools or custom software your business already uses. For each one, spend 15 minutes finding out: Can you export your data. Is the code portable. Are your accounts in your name. Write down what you find. You'll probably discover at least one lock-in risk.
  • Before you start any new AI project, write three questions into your brief: "Where does my data live and how do I export it." "Will the code run on standard cloud platforms." "Will all accounts be registered in my company's name." Share these with anyone you're considering hiring. Their answer will tell you a lot.
  • Talk to your accountant or IT person about what data formats they can work with. If they say "We need CSV and JSON," that's your standard. When you hire someone to build an AI tool, tell them: "All exports must be CSV or JSON." It's a small requirement that prevents big problems later.

Not sure where AI fits yet?

In a 30-minute call we’ll find the one or two places AI will actually pay off for your business, and tell you honestly if it isn’t worth doing yet.