Every practice owner has heard the same argument: you need AI or you will be left behind. The pressure is real. Your competitors are talking about it. Your software vendor is promoting it. Your team is curious.
But “everyone else is doing it” is not a business case. It is a feeling. A real business case starts with measurement.
## The problem with AI adoption without a framework
Practices that succeed with AI adoption have one thing in common: they know what they are measuring before they buy. Practices that fail have a different thing in common: they bought first and hoped the benefits would become obvious later.
Hope is not a strategy. Most AI tools cost money to implement and time to learn. If you have not defined what “success” looks like before you start, you will not recognise it when it happens. Worse, you may implement the tool, spend the money, train the staff, and then realise it did not actually solve the problem you thought it solved.
## Three questions to ask before committing
**Question 1: What specific part of your practice workflow takes time or creates errors?**
Not “we need to be more efficient”. Specific. Does it take your receptionist 20 minutes per day to manually enter patient data from referral letters? Does it take your clinician 10 minutes per appointment to write clinical notes? Does your staff make frequent errors in appointment scheduling?
Name the problem in measurable terms. Time. Error rate. Staff stress. Something that currently exists and you can track.
**Question 2: Will this AI tool actually fix that specific problem?**
Read case studies from practices like yours. Ask the vendor for examples where the tool solved exactly your problem, in your software, in Australian practices. Do not accept generic examples. Do not accept promises. Ask for practices you can call.
If the vendor cannot show you evidence that the tool solves your specific problem, do not buy it. The promise that it might is not evidence.
**Question 3: How will you measure whether it actually worked?**
Before you implement, decide what you will measure. If you are trying to save time, decide how much time saved counts as success. If you are trying to reduce errors, decide what error rate counts as success.
Do not measure feelings. Measure time. Measure error rates. Measure patient satisfaction if that is your goal. Measure something concrete. Then measure it after 30 days, 60 days, and 90 days.
## What realistic ROI looks like
In the first 6 months after implementing an AI tool, the ROI comes from three sources:
**Time savings.** An AI tool that eliminates 5 minutes from your receptionist’s day saves roughly 20 hours per month. At even a modest hourly rate, that is measurable value.
**Error reduction.** An AI tool that catches appointment scheduling errors, or flags missing information before it becomes a clinical problem, reduces costs from having to fix those errors.
**Patient experience.** An AI tool that makes appointment reminders more useful, or gives patients faster responses, can improve retention and referrals.
Real practices see 10 to 30 per cent improvement in the area they targeted. That is not transformative. It is incremental. But incremental improvement is real improvement.
Some practices see no improvement. That usually means the tool did not fit their workflow, or the staff did not adopt it fully. Those are valuable discoveries to make in 90 days, not six months in.
## When to walk away
If after 90 days the tool has not delivered on the specific outcome you measured, you have two choices: adapt how you are using it, or stop using it.
Do not keep paying for a tool because you already bought it. That is a sunk cost fallacy. The money you spent is gone. The decision now is whether the future benefit justifies the future cost.
Plenty of AI tools work brilliantly for some practices and not at all for others. That does not mean the tool is bad. It means it was not the right tool for your practice.
## Start here
The worst time to evaluate an AI tool is when the vendor is selling it to you. At that moment, the incentives are aligned toward you buying. A better time is when you have already named the problem, measured the baseline, and written down what success looks like.
Then when a vendor shows you their tool, you can ask the right question: does this solve my specific problem, in my workflow, at a cost that makes sense. Not: is this the future of dentistry. Not: is everyone else using it. The right question is: does this work for my practice.
That question is one only you can answer. But you can only answer it well if you have done the measurement first.