The Volume and Variability Test for Dental Work

Frequency alone does not make work suitable for automation. The other question is how often the work follows the same path.

Take one workflow and rate it on two dimensions:

• Volume: Does it happen many times a day, or only occasionally?
• Variability: Do most cases follow the same rules, or does each one require a different response?

The combination points toward a sensible type of support.

High volume, low variability

Start with automation. The rules repeat often enough to repay setup and maintenance.

Dental examples include appointment reminders, report delivery, routine recall messages, review requests, and routing a standard work item to the right queue.

Keep an exception path. Even a reliable process will produce cases that need a person.

High volume, high variability

Use a person with judgment, supported by tools. This can suit a trained in-house team member or a VA when physical presence is unnecessary.

Patient phone conversations, complex benefits questions, claim follow-up, and treatment coordination often sit here. Automation can prepare the work, supply context, or handle the predictable subset. It should not pretend every case is predictable.

Low volume, low variability

Write an SOP. A task that happens twice a month may be easy to automate and still not be worth automating.

Every automation needs testing, monitoring, and occasional repair. For infrequent work, a short checklist may be cheaper and easier to trust.

Low volume, high variability

Keep the work with a skilled person or the owner. Examples include payer negotiations, unusual appeals, staffing decisions, and sensitive patient disputes.

These tasks do not gain much from a rigid workflow. They need context and authority.

Two traps

The first is automating rare work because it is technically possible. The build becomes more expensive than the saved time.

The second is automating high-volume conversations end to end. Volume makes the opportunity look large, but variability makes mistakes more likely. A better pattern is to automate preparation and routine cases, then send exceptions to a person.

The grid does not choose a vendor. It tells you what kind of work system you need. Use it after identifying the failure mode and before comparing specific solutions.